[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-microsoft--responsible-ai-toolbox":3,"tool-microsoft--responsible-ai-toolbox":64},[4,17,27,35,43,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,3,"2026-04-05T11:01:52",[13,14,15],"开发框架","图像","Agent","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":23,"last_commit_at":24,"category_tags":25,"status":16},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",140436,2,"2026-04-05T23:32:43",[13,15,26],"语言模型",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":23,"last_commit_at":33,"category_tags":34,"status":16},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107662,"2026-04-03T11:11:01",[13,14,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":23,"last_commit_at":41,"category_tags":42,"status":16},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[13,26],{"id":44,"name":45,"github_repo":46,"description_zh":47,"stars":48,"difficulty_score":23,"last_commit_at":49,"category_tags":50,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[14,51,52,53,15,54,26,13,55],"数据工具","视频","插件","其他","音频",{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":10,"last_commit_at":62,"category_tags":63,"status":16},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,"2026-04-04T04:44:48",[15,14,13,26,54],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":69,"readme_en":70,"readme_zh":71,"quickstart_zh":72,"use_case_zh":73,"hero_image_url":74,"owner_login":75,"owner_name":76,"owner_avatar_url":77,"owner_bio":78,"owner_company":79,"owner_location":79,"owner_email":80,"owner_twitter":81,"owner_website":82,"owner_url":83,"languages":84,"stars":107,"forks":108,"last_commit_at":109,"license":110,"difficulty_score":23,"env_os":111,"env_gpu":112,"env_ram":112,"env_deps":113,"category_tags":126,"github_topics":127,"view_count":23,"oss_zip_url":79,"oss_zip_packed_at":79,"status":16,"created_at":148,"updated_at":149,"faqs":150,"releases":179},2030,"microsoft\u002Fresponsible-ai-toolbox","responsible-ai-toolbox","Responsible AI Toolbox is a suite of tools providing model and data exploration and assessment user interfaces and libraries that enable a better understanding of AI systems. These interfaces and libraries empower developers and stakeholders of AI systems to develop and monitor AI more responsibly, and take better data-driven actions.","Responsible AI Toolbox 是一套帮助开发者和团队更负责任地构建与评估人工智能系统的开源工具集。它提供多个可视化仪表盘，涵盖模型错误分析、公平性检测、可解释性解读和整体决策支持，让使用者能直观发现模型在不同人群或数据子集中的表现偏差、误判原因，并采取针对性改进措施。通过整合微软开源的 InterpretML 和 Fairlearn 等库，它能自动识别模型对性别、年龄等敏感特征的潜在歧视，同时支持从数据不平衡到模型修正的全流程诊断。这套工具特别适合机器学习工程师、数据科学家和AI产品负责人使用，帮助他们在模型开发与部署阶段提前发现伦理与性能风险，做出更透明、更公正的数据驱动决策。界面友好，支持 Python 和 Web 交互，无需深度算法背景也能上手，是推动负责任AI实践的实用助手。","![MIT license](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-blue.svg)\n\n![Responsible AI Widgets Python Build](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002FCI-raiwidgets-pytest.yml?branch=main&label=Responsible%20AI%20Widgets%20Python%20Build)\n![UI deployment to test environment](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002FCD.yml?branch=main&label=UI%20deployment%20to%20test%20environment)\n\n![PyPI raiwidgets](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fraiwidgets?label=PyPI%20raiwidgets)\n![PyPI responsibleai](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fresponsibleai?label=PyPI%20responsibleai)\n![PyPI erroranalysis](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Ferroranalysis?label=PyPI%20erroranalysis)\n![PyPI raiutils](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fraiutils?label=PyPI%20raiutils)\n![PyPI rai_test_utils](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Frai_test_utils?label=PyPI%20rai_test_utils)\n\n![npm model-assessment](https:\u002F\u002Fimg.shields.io\u002Fnpm\u002Fv\u002F@responsible-ai\u002Fmodel-assessment?label=npm%20%40responsible-ai%2Fmodel-assessment)\n\n# Responsible AI Toolbox\nResponsible AI is an approach to assessing, developing, and deploying AI systems in a safe, trustworthy, and ethical manner, and take responsible decisions and actions.\n\nResponsible AI Toolbox is a suite of tools providing a collection of model and data exploration and assessment user interfaces and libraries that enable a better understanding of AI systems. These interfaces and libraries empower developers and stakeholders of AI systems to develop and monitor AI more responsibly, and take better data-driven actions.\n\n\n\u003Cp align=\"center\">\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmicrosoft_responsible-ai-toolbox_readme_e96c7fc09bc7.png\" alt=\"ResponsibleAIToolboxOverview\" width=\"750\"\u002F>\n\n\nThe Toolbox consists of three repositories: \n\n \n| Repository| Tools Covered  |\n|--|--|\n| [Responsible-AI-Toolbox Repository](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox) (Here) |This repository contains four visualization widgets for model assessment and decision making:\u003Cbr>1. [Responsible AI dashboard](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox#introducing-responsible-ai-dashboard), a single pane of glass bringing together several mature Responsible AI tools from the toolbox for a holistic responsible assessment and debugging of models and making informed business decisions. With this dashboard, you can identify model errors, diagnose why those errors are happening, and mitigate them. Moreover, the causal decision-making capabilities provide actionable insights to your stakeholders and customers.\u003Cbr>2. [Error Analysis dashboard](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fdocs\u002Ferroranalysis-dashboard-README.md), for identifying model errors and discovering cohorts of data for which the model underperforms. \t\u003Cbr>3. [Interpretability dashboard](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fdocs\u002Fexplanation-dashboard-README.md), for understanding model predictions. This dashboard is powered by InterpretML.\u003Cbr>4. [Fairness dashboard](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fdocs\u002Ffairness-dashboard-README.md), for understanding model’s fairness issues using various group-fairness metrics across sensitive features and cohorts. This dashboard is powered by Fairlearn. \n| [Responsible-AI-Toolbox-Mitigations Repository](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox-mitigations) | The Responsible AI Mitigations Library helps AI practitioners explore different measurements and mitigation steps that may be most appropriate when the model underperforms for a given data cohort. The library currently has two modules: \u003Cbr>1. DataProcessing, which offers mitigation techniques for improving model performance for specific cohorts. \u003Cbr>2. DataBalanceAnalysis, which provides metrics for diagnosing errors that originate from data imbalance either on class labels or feature values. \u003Cbr> 3. Cohort: provides classes for handling and managing cohorts, which allows the creation of custom pipelines for each cohort in an easy and intuitive interface. The module also provides techniques for learning different decoupled estimators (models) for different cohorts and combining them in a way that optimizes different definitions of group fairness.|  \n[Responsible-AI-Tracker Repository](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox-tracker) |Responsible AI Toolbox Tracker is a JupyterLab extension for managing, tracking, and comparing results of machine learning experiments for model improvement. Using this extension, users can view models, code, and visualization artifacts within the same framework enabling therefore fast model iteration and evaluation processes. Main functionalities include: \u003Cbr>1. Managing and linking model improvement artifacts\u003Cbr> 2. Disaggregated model evaluation and comparisons\u003Cbr>3. Integration with the Responsible AI Mitigations library\u003Cbr>4. Integration with mlflow|\n [Responsible-AI-Toolbox-GenBit Repository](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox-genbit) | The Responsible AI Gender Bias (GenBit) Library helps AI practitioners measure gender bias in Natural Language Processing (NLP) datasets. The main goal of GenBit is to analyze your text corpora and compute metrics that give insights into the gender bias present in a corpus.|\n\n  \n\n\n## Introducing Responsible AI dashboard\n\n[Responsible AI dashboard](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Ftour.ipynb) is a single pane of glass, enabling you to easily flow through different stages of model debugging and decision-making. This customizable experience can be taken in a multitude of directions, from analyzing the model or data holistically, to conducting a deep dive or comparison on cohorts of interest, to explaining and perturbing model predictions for individual instances, and to informing users on business decisions and actions.\n\n\n\u003Cp align=\"center\">\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmicrosoft_responsible-ai-toolbox_readme_70bf6a968f19.png\" alt=\"ResponsibleAIDashboard\" width=\"750\"\u002F>\n\n\n\n\nIn order to achieve these capabilities, the dashboard integrates together ideas and technologies from several open-source toolkits in the areas of\n\n\n\n- \u003Cb>Error Analysis\u003C\u002Fb> powered by [Error Analysis](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-widgets\u002Fblob\u002Fmain\u002Fdocs\u002Ferroranalysis-dashboard-README.md), which identifies cohorts of data with higher error rate than the overall benchmark. These discrepancies might occur when the system or model underperforms for specific demographic groups or infrequently observed input conditions in the training data.\n- \u003Cb>Fairness Assessment\u003C\u002Fb> powered by [Fairlearn](https:\u002F\u002Fgithub.com\u002Ffairlearn\u002Ffairlearn), which identifies which groups of people may be disproportionately negatively impacted by an AI system and in what ways.\n\n- \u003Cb>Model Interpretability\u003C\u002Fb> powered by [InterpretML](https:\u002F\u002Fgithub.com\u002Finterpretml\u002Finterpret-community), which explains blackbox models, helping users understand their model's global behavior, or the reasons behind individual predictions.\n\n- \u003Cb>Counterfactual Analysis\u003C\u002Fb> powered by [DiCE](https:\u002F\u002Fgithub.com\u002Finterpretml\u002FDiCE), which shows feature-perturbed versions of the same datapoint who would have received a different prediction outcome, e.g., Taylor's loan has been rejected by the model. But they would have received the loan if their income was higher by $10,000.\n\n- \u003Cb>Causal Analysis\u003C\u002Fb> powered by [EconML](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FEconML), which focuses on answering What If-style questions to apply data-driven decision-making – how would revenue be affected if a corporation pursues a new pricing strategy? Would a new medication improve a patient’s condition, all else equal?\n\n- \u003Cb>Data Balance\u003C\u002Fb> powered by [Responsible AI](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fdocs\u002Fdatabalance-README.md), which helps users gain an overall understanding of their data, identify features receiving the positive outcome more than others, and visualize feature distributions.\n\nResponsible AI dashboard is designed to achieve the following goals:\n\n- To help further accelerate engineering processes in machine learning by enabling practitioners to design customizable workflows and tailor Responsible AI dashboards that best fit with their model assessment and data-driven decision making scenarios.\n- To help model developers create end to end and fluid debugging experiences and navigate seamlessly through error identification and diagnosis by using interactive visualizations that identify errors, inspect the data, generate global and local explanations models, and potentially inspect problematic examples.\n- To help business stakeholders explore causal relationships in the data and take informed decisions in the real world.\n\nThis repository contains the Jupyter notebooks with examples to showcase how to use this widget. Get started [here](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fgetting-started.ipynb).\n\n\n### Installation\n\nUse the following pip command to install the Responsible AI Toolbox.\n\nIf running in jupyter, please make sure to restart the jupyter kernel after installing.\n\n```\npip install raiwidgets\n```\n\n\n### Responsible AI dashboard Customization\n\nThe Responsible AI Toolbox’s strength lies in its customizability. It empowers users to design tailored, end-to-end model debugging and decision-making workflows that address their particular needs. Need some inspiration? Here are some examples of how Toolbox components can be put together to analyze scenarios in different ways:\n\nPlease note that model overview (including fairness analysis) and data explorer components are activated by default!\n \n| Responsible AI Dashboard Flow| Use Case  |\n|--|--|\n| Model Overview -> Error Analysis -> Data Explorer | To identify model errors and diagnose them by understanding the underlying data distribution\n| Model Overview -> Fairness Assessment -> Data Explorer | To identify model fairness issues and diagnose them by understanding the underlying data distribution\n| Model Overview -> Error Analysis -> Counterfactuals Analysis and What-If | To diagnose errors in individual instances with counterfactual analysis (minimum change to lead to a different model prediction)\n| Model Overview -> Data Explorer -> Data Balance | To understand the root cause of errors and fairness issues introduced via data imbalances or lack of representation of a particular data cohort\n | Model Overview -> Interpretability | To diagnose model errors through understanding how the model has made its predictions\n | Data Explorer -> Causal Inference | To distinguish between correlations and causations in the data or decide the best treatments to apply to see a positive outcome\n  | Interpretability -> Causal Inference | To learn whether the factors that model has used for decision making has any causal effect on the real-world outcome.\n | Data Explorer -> Counterfactuals Analysis and What-If | To address customer questions about what they can do next time to get a different outcome from an AI.\n  | Data Explorer -> Data Balance | To gain an overall understanding of the data, identify features receiving the positive outcome more than others, and visualize feature distributions\n\n### Useful Links\n\n- [Take a tour of Responsible AI Dashboard](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Ftour.ipynb)\n- [Get started](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fgetting-started.ipynb)\n\nTabular Examples:\n- [Try the tool: make decisions for house improvements (decision making)](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fresponsibleaidashboard-housing-decision-making.ipynb)\n- [Try the tool: provide recommendations to patients using diabetes data (decision making)](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fresponsibleaidashboard-diabetes-decision-making.ipynb)\n- [Try the tool: model debugging of a census income prediction model (classification)](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fresponsibleaidashboard-census-classification-model-debugging.ipynb)\n- [Try the tool: model debugging of a housing price prediction model (classification)](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fresponsibleaidashboard-housing-classification-model-debugging.ipynb)\n- [Try the tool: model debugging of a diabetes progression prediction model (regression)](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fresponsibleaidashboard-diabetes-regression-model-debugging.ipynb)\n\nText Examples:\n- [Try the tool: model debugging of an OpenAI Question Answering model on SQuAD](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftext\u002Fresponsibleaidashboard-openai-model-debugging.ipynb)\n- [Try the tool: model debugging of a HuggingFace Question Answering model on SQuAD](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftext\u002Fresponsibleaidashboard-question-answering-model-debugging.ipynb)\n- [Try the tool: model debugging of a DBPedia text classification model](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftext\u002Fresponsibleaidashboard-DBPedia-text-classification-model-debugging.ipynb)\n- [Try the tool: model debugging of a binary text classification model](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftext\u002Fresponsibleaidashboard-blbooksgenre-binary-text-classification-model-debugging.ipynb)\n- [Try the tool: model debugging of a COVID-19 multilabel text classification model](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftext\u002Fresponsibleaidashboard-covid-event-multilabel-text-classification-model-debugging.ipynb)\n\nVision Examples:\n- [Try the tool: model debugging of a fridge image classification model](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Fvision\u002Fresponsibleaidashboard-fridge-image-classification-model-debugging.ipynb)\n- [Try the tool: model debugging of a fridge multilabel image classification model](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Fvision\u002Fresponsibleaidashboard-fridge-multilabel-image-classification-model-debugging.ipynb)\n- [Try the tool: model debugging of a fridge object detection model](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Fvision\u002Fresponsibleaidashboard-fridge-object-detection-model-debugging.ipynb)\n\n## Supported Models\n\nThis Responsible AI Toolbox API supports models that are trained on datasets in Python `numpy.ndarray`, `pandas.DataFrame`, `iml.datatypes.DenseData`, or `scipy.sparse.csr_matrix` format.\n\nThe explanation functions of [Interpret-Community](https:\u002F\u002Fgithub.com\u002Finterpretml\u002Finterpret-community) accept both models and pipelines as input as long as the model or pipeline implements a `predict` or `predict_proba` function that conforms to the Scikit convention. If not compatible, you can wrap your model's prediction function into a wrapper function that transforms the output into the format that is supported (predict or predict_proba of Scikit), and pass that wrapper function to your selected interpretability techniques.\n\nIf a pipeline script is provided, the explanation function assumes that the running pipeline script returns a prediction. The repository also supports models trained via **PyTorch**, **TensorFlow**, and **Keras** deep learning frameworks.\n\n## Other Use Cases\n\nTools within the Responsible AI Toolbox can also be used with AI models offered as APIs by providers such as [Azure Cognitive Services](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fcognitive-services\u002F). To see example use cases, see the folders below:\n\n- [Cognitive Services Speech to Text Fairness testing](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Ftree\u002Fmain\u002Fnotebooks\u002Fcognitive-services-examples\u002Fspeech-to-text)\n- [Cognitive Services Face Verification Fairness testing](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Ftree\u002Fmain\u002Fnotebooks\u002Fcognitive-services-examples\u002Fface-verification)\n\n## Maintainers\n\n- [Ke Xu](https:\u002F\u002Fgithub.com\u002FKeXu444)\n- [Roman Lutz](https:\u002F\u002Fgithub.com\u002Fromanlutz)\n- [Ilya Matiach](https:\u002F\u002Fgithub.com\u002Fimatiach-msft)\n- [Gaurav Gupta](https:\u002F\u002Fgithub.com\u002Fgaugup)\n- [Vinutha Karanth](https:\u002F\u002Fgithub.com\u002Fvinuthakaranth)\n- [Tong Yu](https:\u002F\u002Fgithub.com\u002Ftongyu-microsoft)\n- [Ruby Zhu](https:\u002F\u002Fgithub.com\u002FRubyZ10)\n- [Mehrnoosh Sameki](https:\u002F\u002Fgithub.com\u002Fmesameki)\n- [Hannah Westra](https:\u002F\u002Fgithub.com\u002Fhawestra)\n- [Ziqi Ma](https:\u002F\u002Fgithub.com\u002Fziqi-ma)\n- [Kin Chan](https:\u002F\u002Fgithub.com\u002Fkicha0)\n","![MIT 许可证](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-blue.svg)\n\n![负责任的 AI 小部件 Python 构建](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002FCI-raiwidgets-pytest.yml?branch=main&label=负责任的 AI 小部件 Python 构建)\n![UI 部署到测试环境](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002FCD.yml?branch=main&label=UI%20部署%20到%20测试%20环境)\n\n![PyPI raiwidgets](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fraiwidgets?label=PyPI%20raiwidgets)\n![PyPI responsibleai](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fresponsibleai?label=PyPI%20responsibleai)\n![PyPI erroranalysis](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Ferroranalysis?label=PyPI%20erroranalysis)\n![PyPI raiutils](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fraiutils?label=PyPI%20raiutils)\n![PyPI rai_test_utils](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Frai_test_utils?label=PyPI%20rai_test_utils)\n\n![npm model-assessment](https:\u002F\u002Fimg.shields.io\u002Fnpm\u002Fv\u002F@responsible-ai\u002Fmodel-assessment?label=npm%20%40responsible-ai%2Fmodel-assessment)\n\n# 负责任的 AI 工具箱\n负责任的 AI 是一种以安全、可信和合乎道德的方式评估、开发和部署 AI 系统，并做出负责任决策与行动的方法。\n\n负责任的 AI 工具箱是一套工具，提供了一系列模型与数据探索及评估的用户界面和库，帮助更好地理解 AI 系统。这些界面和库赋能 AI 系统的开发者与利益相关方更负责任地开发与监控 AI，并采取更好的数据驱动行动。\n\n\n\u003Cp align=\"center\">\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmicrosoft_responsible-ai-toolbox_readme_e96c7fc09bc7.png\" alt=\"ResponsibleAIToolboxOverview\" width=\"750\"\u002F>\n\n\n该工具箱由三个仓库组成： \n\n \n| 仓库| 涵盖的工具  |\n|--|--|\n| [负责任的 AI 工具箱仓库](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox)（此处） | 此仓库包含四个用于模型评估与决策的可视化小部件：\u003Cbr>1. [负责任的 AI 仪表板](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox#introducing-responsible-ai-dashboard)，一个集成了工具箱中多个成熟的负责任 AI 工具的单一窗口，用于对模型进行全面的负责任评估与调试，并作出明智的业务决策。借助此仪表板，您可以识别模型错误，诊断错误产生的原因并加以缓解。此外，因果决策功能还能为您的利益相关方与客户提供可操作的洞察。\u003Cbr>2. [错误分析仪表板](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fdocs\u002Ferroranalysis-dashboard-README.md)，用于识别模型错误并发现模型表现不佳的数据群体。\u003Cbr>3. [可解释性仪表板](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fdocs\u002Fexplanation-dashboard-README.md)，用于理解模型预测。此仪表板由 InterpretML 提供支持。\u003Cbr>4. [公平性仪表板](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fdocs\u002Ffairness-dashboard-README.md)，用于通过多种群体公平性指标，理解模型在敏感特征与群体中的公平性问题。此仪表板由 Fairlearn 提供支持。 \n| [负责任的 AI 工具箱缓解措施仓库](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox-mitigations) | 负责任的 AI 缓解措施库帮助 AI 从业者探索不同的度量方法与缓解步骤，这些方法与步骤可能最适合于特定数据群体中模型表现不佳的情况。该库目前有两个模块：\u003Cbr>1. DataProcessing，提供针对特定群体改进模型性能的缓解技术。\u003Cbr>2. DataBalanceAnalysis，提供用于诊断源于数据不平衡的错误的指标，包括类别标签或特征值的不平衡。\u003Cbr>3. Cohort：提供用于处理与管理群体的类，允许您以简单直观的界面为每个群体创建自定义管道。该模块还提供学习不同解耦估计器（模型）的技术，并以优化不同群体公平性定义的方式进行组合。|  \n[负责任的 AI 工具箱追踪器仓库](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox-tracker) | 负责任的 AI 工具箱追踪器是 JupyterLab 的扩展程序，用于管理、追踪与比较机器学习实验的结果，以改进模型。使用此扩展程序，用户可在同一框架内查看模型、代码与可视化产物，从而实现快速的模型迭代与评估流程。主要功能包括：\u003Cbr>1. 管理与链接模型改进产物\u003Cbr> 2. 分解式模型评估与比较\u003Cbr>3. 与负责任的 AI 缓解措施库集成\u003Cbr>4. 与 mlflow 集成|\n [负责任的 AI 工具箱 GenBit 仓库](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox-genbit) | 负责任的 AI 性别偏见（GenBit）库帮助 AI 从业者衡量自然语言处理（NLP）数据集中的性别偏见。GenBit 的主要目标是分析您的文本语料库，并计算能揭示语料库中性别偏见的指标。\n\n## 介绍负责任的人工智能仪表板\n\n[负责任的人工智能仪表板](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Ftour.ipynb) 是一个单一的可视化界面，可让您轻松地完成模型调试和决策制定的不同阶段。这种可定制的体验可朝多个方向发展：从整体分析模型或数据，到深入研究或比较感兴趣的群体，再到解释并扰动单个实例的模型预测，以及为业务决策和行动提供建议。\n\n\n\u003Cp align=\"center\">\n\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmicrosoft_responsible-ai-toolbox_readme_70bf6a968f19.png\" alt=\"ResponsibleAIDashboard\" width=\"750\"\u002F>\n\n\n\n\n为了实现这些功能，该仪表板整合了来自多个开源工具包的思想和技术，涉及以下领域：\n\n\n- **误差分析**，由 [Error Analysis](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-widgets\u002Fblob\u002Fmain\u002Fdocs\u002Ferroranalysis-dashboard-README.md) 提供支持，可识别出误差率高于总体基准的数据群体。当系统或模型在特定人口群体或训练数据中不常出现的输入条件下表现不佳时，就可能出现这些差异。\n- **公平性评估**，由 [Fairlearn](https:\u002F\u002Fgithub.com\u002Ffairlearn\u002Ffairlearn) 提供支持，可识别哪些人群可能受到人工智能系统的不成比例的负面影响，以及具体的影响方式。\n\n- **模型可解释性**，由 [InterpretML](https:\u002F\u002Fgithub.com\u002Finterpretml\u002Finterpret-community) 提供支持，可解释黑盒模型，帮助用户理解模型的整体行为，或单个预测背后的原因。\n\n- **反事实分析**，由 [DiCE](https:\u002F\u002Fgithub.com\u002Finterpretml\u002FDiCE) 提供支持，可展示同一数据点经过特征扰动后的版本，这些版本原本会得到不同的预测结果。例如，泰勒的贷款被模型拒绝了，但如果他的收入增加1万美元，他本可以成功获得贷款。\n\n- **因果分析**，由 [EconML](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002FEconML) 提供支持，专注于回答“如果……会怎样”的问题，以实现数据驱动的决策——如果一家公司实施新的定价策略，收入会如何变化？在其他条件不变的情况下，一种新药能否改善患者的病情？\n\n- **数据平衡**，由 [Responsible AI](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fdocs\u002Fdatabalance-README.md) 提供支持，可帮助用户全面了解其数据，识别哪些特征比其他特征更易获得积极结果，并可视化特征分布。\n\n负责任的人工智能仪表板旨在实现以下目标：\n\n- 帮助进一步加速机器学习中的工程流程，使从业者能够设计可定制的工作流，并量身打造最适合其模型评估和数据驱动决策场景的负责任的人工智能仪表板。\n- 帮助模型开发者创建端到端、流畅的调试体验，通过交互式可视化无缝地进行错误识别与诊断：识别错误、检查数据、生成全局和局部模型解释，甚至检查有问题的示例。\n- 帮助业务利益相关者探索数据中的因果关系，并在现实世界中做出明智的决策。\n\n此仓库包含 Jupyter 笔记本，其中提供了使用该小部件的示例。请从[这里](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fgetting-started.ipynb)开始使用。\n\n\n### 安装\n\n使用以下 pip 命令安装负责任的人工智能工具箱。\n\n如果在 Jupyter 中运行，请确保安装完成后重启 Jupyter 内核。\n\n```\npip install raiwidgets\n```\n\n\n### 负责任的人工智能仪表板的定制\n\n负责任的人工智能工具箱的优势在于其高度可定制性。它赋予用户能力，设计量身定制的端到端模型调试与决策工作流，以满足他们的特殊需求。需要一些灵感吗？以下是几种将工具箱组件组合起来以不同方式分析场景的示例：\n\n请注意，模型概览（包括公平性分析）和数据探索器组件默认已启用！\n\n\n| 负责任的人工智能仪表板流程 | 使用场景 |\n|--|--|\n| 模型概览 -> 误差分析 -> 数据探索器 | 识别模型误差并通过理解底层数据分布进行诊断\n| 模型概览 -> 公平性评估 -> 数据探索器 | 识别模型公平性问题并通过理解底层数据分布进行诊断\n| 模型概览 -> 误差分析 -> 反事实分析与“如果……会怎样” | 通过反事实分析诊断单个实例的误差（最小改动即可导致不同的模型预测）\n| 模型概览 -> 数据探索器 -> 数据平衡 | 理解因数据不平衡或特定数据群体代表性不足而引入的误差和公平性问题的根本原因\n| 模型概览 -> 可解释性 | 通过理解模型如何做出预测来诊断模型误差\n| 数据探索器 -> 因果推断 | 区分数据中的相关性和因果关系，或决定最佳治疗方案以获得积极效果\n| 可解释性 -> 因果推断 | 了解模型用于决策的因素是否对现实世界的最终结果有因果影响\n| 数据探索器 -> 反事实分析与“如果……会怎样” | 解答客户关于下次如何采取行动以获得不同人工智能结果的问题\n| 数据探索器 -> 数据平衡 | 全面了解数据，识别哪些特征比其他特征更易获得积极结果，并可视化特征分布\n\n### 有用链接\n\n- [体验负责任的人工智能仪表板](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Ftour.ipynb)\n- [开始使用](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fgetting-started.ipynb)\n\n表格示例：\n- [试用该工具：为房屋改善决策提供支持（决策制定）](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fresponsibleaidashboard-housing-decision-making.ipynb)\n- [试用该工具：利用糖尿病数据为患者提供建议（决策制定）](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fresponsibleaidashboard-diabetes-decision-making.ipynb)\n- [试用该工具：对人口普查收入预测模型进行模型调试（分类）](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fresponsibleaidashboard-census-classification-model-debugging.ipynb)\n- [试用该工具：对房屋价格预测模型进行模型调试（分类）](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fresponsibleaidashboard-housing-classification-model-debugging.ipynb)\n- [试用该工具：对糖尿病进展预测模型进行模型调试（回归）](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fresponsibleaidashboard-diabetes-regression-model-debugging.ipynb)\n\n文本示例：\n- [试用该工具：对OpenAI问答模型在SQuAD上的模型调试](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftext\u002Fresponsibleaidashboard-openai-model-debugging.ipynb)\n- [试用该工具：对HuggingFace问答模型在SQuAD上的模型调试](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftext\u002Fresponsibleaidashboard-question-answering-model-debugging.ipynb)\n- [试用该工具：对DBPedia文本分类模型进行模型调试](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftext\u002Fresponsibleaidashboard-DBPedia-text-classification-model-debugging.ipynb)\n- [试用该工具：对二元文本分类模型进行模型调试](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftext\u002Fresponsibleaidashboard-blbooksgenre-binary-text-classification-model-debugging.ipynb)\n- [试用该工具：对COVID-19多标签文本分类模型进行模型调试](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftext\u002Fresponsibleaidashboard-covid-event-multilabel-text-classification-model-debugging.ipynb)\n\n视觉示例：\n- [试用该工具：对冰箱图像分类模型进行模型调试](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Fvision\u002Fresponsibleaidashboard-fridge-image-classification-model-debugging.ipynb)\n- [试用该工具：对冰箱多标签图像分类模型进行模型调试](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Fvision\u002Fresponsibleaidashboard-fridge-multilabel-image-classification-model-debugging.ipynb)\n- [试用该工具：对冰箱目标检测模型进行模型调试](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Fvision\u002Fresponsibleaidashboard-fridge-object-detection-model-debugging.ipynb)\n\n## 支持的模型\n\n本负责任的人工智能工具箱API支持以下格式的数据集上训练的模型：Python中的`numpy.ndarray`、`pandas.DataFrame`、`iml.datatypes.DenseData`或`scipy.sparse.csr_matrix`。\n\n[Interpret-Community](https:\u002F\u002Fgithub.com\u002Finterpretml\u002Finterpret-community)的解释功能接受模型和管道作为输入，只要模型或管道实现了符合Scikit规范的`predict`或`predict_proba`函数即可。如果不兼容，您可以将模型的预测函数包装成一个适配器函数，将其输出转换为Scikit支持的格式（即Scikit的`predict`或`predict_proba`），然后将此适配器函数传递给您选择的可解释性技术。\n\n如果提供了管道脚本，解释功能会假定运行的管道脚本返回预测结果。该仓库还支持通过**PyTorch**、**TensorFlow**和**Keras**深度学习框架训练的模型。\n\n## 其他应用场景\n\n负责任的人工智能工具箱中的工具也可用于由诸如[Azure认知服务](https:\u002F\u002Fazure.microsoft.com\u002Fen-us\u002Fservices\u002Fcognitive-services\u002F)等提供商提供的AI模型作为API的情况。如需查看示例应用场景，请参阅以下文件夹：\n\n- [认知服务语音转文本公平性测试](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Ftree\u002Fmain\u002Fnotebooks\u002Fcognitive-services-examples\u002Fspeech-to-text)\n- [认知服务人脸验证公平性测试](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Ftree\u002Fmain\u002Fnotebooks\u002Fcognitive-services-examples\u002Fface-verification)\n\n## 维护者\n\n- [Ke Xu](https:\u002F\u002Fgithub.com\u002FKeXu444)\n- [Roman Lutz](https:\u002F\u002Fgithub.com\u002Fromanlutz)\n- [Ilya Matiach](https:\u002F\u002Fgithub.com\u002Fimatiach-msft)\n- [Gaurav Gupta](https:\u002F\u002Fgithub.com\u002Fgaugup)\n- [Vinutha Karanth](https:\u002F\u002Fgithub.com\u002Fvinuthakaranth)\n- [Tong Yu](https:\u002F\u002Fgithub.com\u002Ftongyu-microsoft)\n- [Ruby Zhu](https:\u002F\u002Fgithub.com\u002FRubyZ10)\n- [Mehrnoosh Sameki](https:\u002F\u002Fgithub.com\u002Fmesameki)\n- [Hannah Westra](https:\u002F\u002Fgithub.com\u002Fhawestra)\n- [Ziqi Ma](https:\u002F\u002Fgithub.com\u002Fziqi-ma)\n- [Kin Chan](https:\u002F\u002Fgithub.com\u002Fkicha0)","# Responsible AI Toolbox 快速上手指南\n\n## 环境准备\n\n- **系统要求**：Python 3.8+，支持 Windows、macOS、Linux\n- **前置依赖**：Jupyter Notebook 或 JupyterLab（推荐使用）\n- **推荐环境**：使用 Anaconda 或 Miniconda 管理 Python 环境，提升兼容性\n\n> 建议使用国内镜像源加速安装：`-i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple`\n\n## 安装步骤\n\n```bash\npip install raiwidgets -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n```\n\n> 安装完成后，请**重启 Jupyter 内核**以确保组件正确加载。\n\n## 基本使用\n\n1. 启动 Jupyter Notebook：\n\n```bash\njupyter notebook\n```\n\n2. 新建 Notebook，运行以下代码加载并启动 Responsible AI Dashboard：\n\n```python\nfrom raiwidgets import ResponsibleAIDashboard\nfrom sklearn.datasets import load_iris\nfrom sklearn.ensemble import RandomForestClassifier\n\n# 加载示例数据\ndata = load_iris()\nX, y = data.data, data.target\n\n# 训练简单模型\nmodel = RandomForestClassifier()\nmodel.fit(X, y)\n\n# 启动仪表盘\nResponsibleAIDashboard(model, X, y)\n```\n\n3. 浏览器中将自动打开交互式仪表盘，支持模型偏差、公平性、可解释性、反事实分析等多维度探索。\n\n> 更多示例请参考：[Getting Started Notebook](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fblob\u002Fmain\u002Fnotebooks\u002Fresponsibleaidashboard\u002Ftabular\u002Fgetting-started.ipynb)","某大型银行的信贷审批模型团队正在上线一款基于机器学习的贷款风险评估系统，目标是提升审批效率，同时避免对特定人群（如低收入群体、少数民族）产生系统性歧视。团队由数据科学家、风控工程师和合规负责人组成，需在模型上线前完成公平性审计与错误诊断。\n\n### 没有 responsible-ai-toolbox 时\n- 团队只能依赖静态的准确率报表，无法定位模型在哪些客户群体（如年龄\u003C25岁或月收入\u003C5000元）上错误率显著偏高。\n- 公平性检查依赖手动编写多个自定义脚本，耗时近两周，且不同成员使用的指标不一致，导致合规部门无法信任结论。\n- 模型解释仅靠SHAP值图表，无法直观对比不同群体的特征重要性差异，难以向业务部门解释“为何拒绝了某类客户”。\n- 错误分析与公平性分析分散在不同平台，团队需在Excel、Python Notebook和BI工具间反复切换，沟通成本高，决策延迟。\n- 一旦模型上线后出现投诉，团队无法快速回溯是数据偏差、特征工程问题，还是模型本身歧视，修复周期长达一个月。\n\n### 使用 responsible-ai-toolbox 后\n- 通过公平性仪表盘，团队在30分钟内识别出模型对“无固定职业者”的拒绝率高出基准37%，并自动标注出关键驱动特征（如“社保缴纳月数”）。\n- 错误分析仪表盘清晰展示出模型在“首次申请贷款+低学历”群体中误判率高达22%，并推荐了可操作的重采样策略。\n- 解释性仪表盘联动展示不同群体的特征影响图，帮助风控团队向合规部门直观证明“拒绝逻辑并非基于种族，而是基于还款稳定性指标”。\n- 所有分析集成在一个统一仪表盘中，数据科学家、风控与法务人员可实时协同查看、标注、导出报告，会议决策效率提升70%。\n- 模型上线前，团队基于工具内置的Cohort分析模块，提前修正了训练数据中的样本偏差，上线后相关群体投诉率下降61%。\n\nresponsible-ai-toolbox 让信贷模型从“黑箱决策”转变为可解释、可审计、可修正的负责任系统，真正实现了技术合规与商业信任的双赢。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fmicrosoft_responsible-ai-toolbox_e96c7fc0.png","microsoft","Microsoft","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fmicrosoft_4900709c.png","Open source projects and samples from Microsoft",null,"opensource@microsoft.com","OpenAtMicrosoft","https:\u002F\u002Fopensource.microsoft.com","https:\u002F\u002Fgithub.com\u002Fmicrosoft",[85,89,93,96,100,104],{"name":86,"color":87,"percentage":88},"TypeScript","#3178c6",82.4,{"name":90,"color":91,"percentage":92},"Python","#3572A5",14.3,{"name":94,"color":95,"percentage":10},"Jupyter Notebook","#DA5B0B",{"name":97,"color":98,"percentage":99},"JavaScript","#f1e05a",0.3,{"name":101,"color":102,"percentage":103},"HTML","#e34c26",0,{"name":105,"color":106,"percentage":103},"Shell","#89e051",1748,471,"2026-04-04T04:14:23","MIT","Linux, macOS, Windows","未说明",{"notes":114,"python":115,"dependencies":116},"建议在 Jupyter Notebook 环境中使用，安装后需重启内核；部分功能依赖外部库（如 Fairlearn、InterpretML、EconML、DiCE）需确保环境兼容；推荐使用虚拟环境管理依赖。","3.8+",[117,118,119,120,121,122,123,124,125],"raiwidgets","responsibleai","erroranalysis","raiutils","rai_test_utils","fairlearn","interpretml","dice-ml","econml",[51,14,13,54],[128,129,130,131,132,133,134,135,136,137,138,139,140,141,142,143,144,145,146,147],"ui","responsible-ai","data-science","fairness","fairness-ml","fairness-ai","explainable-ai","explainable-ml","explainability","machinelearning","machine-learning","ml","visualization","widgets","widget","jupyter","error-analysis","data-analysis","data-visualization","interpretability","2026-03-27T02:49:30.150509","2026-04-06T11:31:10.117271",[151,156,161,166,171,175],{"id":152,"question_zh":153,"answer_zh":154,"source_url":155},9225,"为什么 ErrorAnalysisDashboard 在 Jupyter Notebook 中无法显示，但 ExplanationDashboard 可以？","尝试升级 raiwidgets 和 interpret-community 包，并确保使用 Python 3.7 环境。执行以下命令：conda create -y --name py37_errorML01 python=3.7，激活环境后运行 %pip install --upgrade interpret-community、%pip install --upgrade raiwidgets、%pip install --upgrade lime 和 %pip install --upgrade rai-core-flask。同时确保安装了 psutil 和升级了 ipykernel。","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fissues\u002F375",{"id":157,"question_zh":158,"answer_zh":159,"source_url":160},9226,"运行 ExplanationDashboard 或 ErrorAnalysisDashboard 时内核崩溃怎么办？","该问题通常由 gevent 和 greenlet 包的版本冲突引起。解决方案是卸载并重新安装这两个包：pip uninstall gevent greenlet，然后重新安装 pip install gevent==20.6.2 和 pip install greenlet==1.0.0。确保环境干净，避免与其他包冲突。","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fissues\u002F356",{"id":162,"question_zh":163,"answer_zh":164,"source_url":165},9227,"Causal Analysis 中使用分类变量作为处理特征时报错怎么办？","在 EconML 的 _causal_analysis.py 文件中，需在 individualized_policy 函数中添加对 pandas Categorical 类型的处理：if isinstance(current_treatment, pd.core.arrays.categorical.Categorical): current_treatment = current_treatment.to_numpy()。该修复已提交至 EconML 仓库（PR #602），后续需升级 EconML 版本并同步到 responsibleai。","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fissues\u002F1311",{"id":167,"question_zh":168,"answer_zh":169,"source_url":170},9228,"ErrorAnalysisDashboard 报错 'can only convert an array of size 1 to a Python scalar' 如何解决？","此错误已在 PR #815 中修复。请确保使用最新版本的 raiwidgets（>=0.2.2）和 interpret-community。若仍出现该问题，检查传入的 dataset 是否为 NumPy 数组或 Pandas DataFrame，避免混合类型或非标准格式，推荐使用标准 DataFrame 格式输入数据。","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fissues\u002F710",{"id":172,"question_zh":173,"answer_zh":174,"source_url":155},9229,"如何在远程 Azure DSVM 上成功运行 RAI Widgets？","在 Azure DSVM 上，请使用 Python 3.7 环境，避免使用 Python 3.6。执行：\u002Fanaconda\u002Fbin\u002Fconda create -y --name py37_errorML01 python=3.7，激活后安装依赖：pip install psutil、pip install --upgrade ipykernel、git clone https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-widgets.git，再使用 %pip install --upgrade raiwidgets interpret-community lime rai-core-flask。确保在 Jupyter 中切换至该内核运行 Notebook。",{"id":176,"question_zh":177,"answer_zh":178,"source_url":155},9230,"为什么 Dashboard 在 Notebook 中无法内联显示，只能在新标签页打开？","这是由于前端资源加载或内联渲染配置问题。尝试在 Notebook 开头添加 %matplotlib inline 或 %config InlineBackend.figure_format = 'retina'，并确保安装了最新版本的 jupyter-server-proxy 和 rai-core-flask。若仍无效，可使用 .show_in_new_tab() 方法强制在新标签页打开，或检查浏览器控制台是否有 CORS 或 JS 加载错误。",[180,185,190,195,200,205,210,215,220,225,230,235,240,245,250,255,260,265,270,275],{"id":181,"version":182,"summary_zh":183,"released_at":184},116310,"v0.36.0","## v0.36.0\r\n\r\n- educational materials\r\n  - Update getting-started.ipynb by @oohtmeel1 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2572\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - Fix robot page exception - Update StatsTableUtils.ts by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2568\r\n    - Fix robot page exception on TypeError: Cannot read properties of undefined (reading 'cohort') by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2540\r\n    - Fix responsibleai toolbox build failures due to dependency updates by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2571\r\n    - Fix build failures due to missing fridge datasets by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2573\r\n  - ## RAIInsights\r\n    - Update responsibleai to scikit-learn 1.5.1 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2570\r\n\r\n## New Contributors\r\n* @oohtmeel1 made their first contribution in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2572\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fcompare\u002Fv0.35.1...v0.36.0","2024-07-08T18:03:34",{"id":186,"version":187,"summary_zh":188,"released_at":189},116311,"v0.35.1","## v0.35.1\r\n* upgrade nx to 12.10.1 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2561\r\n* fix wrapper constructor in ErrorAnalysisManager and RAIVisionInsights load call for RAI Vision Dashboard by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2560\r\n* update responsibleai-text and responsibleai-vision packages to responsibleai 0.35.0 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2563\r\n* release responsibleai-vision 0.3.8 and responsibleai-text 0.2.7 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2564\r\n* remove pins on werkzeug, itsdangerous and markupsafe since they are currently not needed by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2565\r\n* release rai-core-flask 0.7.6 by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2566\r\n* release raiwidgets and responsibleai 0.35.1 by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2567\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fcompare\u002Fv0.35.0...v0.35.1","2024-05-20T22:12:33",{"id":191,"version":192,"summary_zh":193,"released_at":194},116312,"v0.35.0","## v0.35.0\r\n\r\n- new features\r\n  - ## RAI Vision\r\n    - Exif Nested Metadata support for Image Dashboards by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2542\r\n    - Nested Exif Metadata support + Gate fixes for py 3.7 and keras version by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2545\r\n    - Release responsibleai_vision 0.3.7 with Nested Exif Metadata support by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2549\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - fix bad whitespace in url format for azure compute instance base url by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2537\r\n    - release rai-core-flask 0.7.4 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2538\r\n    - Gates fix to remove py 3.7 for EOL & AutoML 2.0.6 by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2543\r\n    - update ipython and werkzeug dependencies for responsibleai and rai-core-flask packages by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2551\r\n    - release rai-core-flask 0.7.5 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2553\r\n    - upgrade packages and CI to rai-core-flask==0.7.5, rai-utils==0.4.2 and rai-test-utils==0.4.2 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2556\r\n  - ## RAIInsights\r\n    - fix builds due to OneHotEncoder sparse parameter breaking change in scikit-learn by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2539\r\n    - update responsibleai and raiwidgets to erroranalysis 0.5.4 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2558\r\n  - ## RAI Vision\r\n    - fix label column incorrectly added by feature extractors in RAI Vision dashboard for automl models by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2532\r\n    - fix index into image_dimensions for drawing bounding boxes in object detection flyout by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2550\r\n  - ## RAI Vision and Text\r\n    - update responsibleai-text and responsibleai-vision packages to responsibleai 0.34.1 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2534\r\n    - release responsibleai-vision 0.3.6 and responsibleai-text 0.2.6 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2535\r\n    - fix release github actions for RAI vision and text packages by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2536\r\n  - ## Error analysis\r\n    - fix erroranalysis test failures due to new shap release having inconsistent dimensions for single valued target by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002F\r\n    - release erroranalysis 0.5.4 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2557pull\u002F2552\r\n  - ## RAIUtils\r\n    - release rai-utils 0.4.2 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2555\r\n  - ## RAI Test Utils\r\n    - release rai-test-utils 0.4.2 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2554\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fcompare\u002Fv0.34.1...v0.35.0","2024-05-01T20:26:32",{"id":196,"version":197,"summary_zh":198,"released_at":199},116313,"v0.34.1","## v0.34.1\r\n\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - Fix forecast flattening prediction list error by @jamesbchao in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2529\r\n  - ## RAI Text\r\n    - Add methods and constants for genai metrics by @kartik727 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2524\r\n    - Genai integration by @kartik727 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2522\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - fix CD.yml permissions errors for dashboard links on PRs by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2525\r\n  - ## RAI Vision and Text\r\n    - update responsibleai-text and responsibleai-vision packages to responsibleai 0.34.0 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2530","2024-02-09T18:33:49",{"id":201,"version":202,"summary_zh":203,"released_at":204},116314,"v0.34.0","## v0.34.0\r\n\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - Private Link URL support by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2493\r\n    - PL Support on raiwidgets by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2500\r\n    - Flyout Bounding Box legend & accessibility fixes by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2497\r\n    - update UI to support genai task for RAI text dashboard by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2508\r\n    - add genai metrics endpoint in UI for model overview metrics (#2517) by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2520\r\n  - ## RAI Text\r\n    - Added metrics for genai text by @kartik727 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2514\r\n    - added generative text explainer by @mohsinposts in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2516\r\n    - Add helper method for genai metrics by @kartik727 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2519\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - upgrade plotly to latest version 2.27.1 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2484\r\n    - LEGO: Pull request from JUNO\u002Fhb_a12a4630-4852-4e7d-9cbc-c0e1117da1ab_20231212124106972 to main by @csigs in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2457\r\n    - Localized file check-in by OneLocBuild Task: Build definition ID 21368: Build ID 113458687 by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2460\r\n  - ## RAIInsights\r\n    - update responsibleai to interpret-community 0.31.0 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2488\r\n    - Release rai_core_flask 0.7.3 by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2496\r\n    - fix builds by supporting sparse_output renamed parameter in OneHotEncoder in new scikit-learn version update by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2507\r\n  - ## RAI Vision\r\n    - fix vision data explorer elements truncated in smaller screen by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2478\r\n    - fix duplicate categorical features added from image exif metadata in RAI Vision Dashboard by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2483\r\n    - Accessibility Fix for Metric Checkboxes by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2485\r\n    - fix wrong metrics count missing in vision explorer of RAI Vision dashboard for object detection by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2495\r\n    - Class View Search Case Fix by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2498\r\n    - Dropdown option fix by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2505\r\n  - ## RAI Vision and Text\r\n    - update responsibleai-text and responsibleai-vision packages to responsibleai 0.33.0 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2477\r\n    - release responsibleai-vision 0.3.5 and responsibleai-text 0.2.5 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2481\r\n    - release tensorflow pin in RAI vision and text package tests which was added last year to fix huggingface transformer models by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2472\r\n  - ## Error analysis\r\n  - ## RAIUtils\r\n    - Fix retry function not stopping for long time after sending an interrupt by @kartik727 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2486\r\n- other\r\n  - Update MANIFEST.in for erroranalysis to include requirements-object-detection.txt by @rxm7706 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2476\r\n  - fix release script due to dependency test errors by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2482\r\n  - ignore error when uploading test dashboard from build for external forks by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2487\r\n  - ci: Use GITHUB_OUTPUT envvar instead of set-output command by @arunsathiya in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2489\r\n\r\n\r\n## New Contributors\r\n* @rxm7706 made their first contribution in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2476\r\n* @arunsathiya made their first contribution in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2489\r\n* @mohsinposts made their first contribution in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2516\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fcompare\u002Fv0.33.0...v0.34.0","2024-02-02T15:38:36",{"id":206,"version":207,"summary_zh":208,"released_at":209},116315,"v0.33.0","## v0.33.0\r\n\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - add AccessibleDetailsList with resizable columns for accessibility by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2470\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - fix aria label for screen reader on scatter plot color axis picker by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2440\r\n    - fix RAI dashboard treat as categorical accessibility on new cohort create by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2446\r\n    - fix line-height override on add tab button for accessibility by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2449\r\n    - fix pivot focus indicator for accessibility by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2450\r\n    - fix confusion matrix highcharts heatmap keyboard focus order accessibility by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2452\r\n    - fix class importance weights callout truncated in small view for accessibility by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2462\r\n    - fix tooltip callout keyboard focus order for accessibility by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2464\r\n    - announce highcharts confusion matrix heatmap tooltip aria label for accessibility by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2467\r\n    - announce search results in RAI Vision dashboard toolbar for accessibility by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2466\r\n  - ## RAIInsights\r\n    - remove python 3.6 from builds, tests and packages by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2442\r\n    - Upgrade setup python from v4 to v5 by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2454\r\n    - release rai-core-flask 0.7.2 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2458\r\n    - update raiwidgets to rai-core-flask v0.7.2 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2459\r\n    - update numpy and numba dependencies by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2473\r\n  - ## RAI Vision\r\n    - fix luminosity ratio of RAI Vision dashboard success and failure instances text to background for accessibility by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2444\r\n    - fix error when extracting exif metadata features from images in RAI Vision dashboard by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2461\r\n    - fix images missing alternate text in RAI Vision dashboard for accessibility by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2463\r\n  - ## RAI Vision and Text\r\n    - update responsibleai-text and responsibleai-vision packages to responsibleai 0.32.1 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2445\r\n    - release responsibleai-vision 0.3.4 and responsibleai-text 0.2.4 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2447\r\n    - fix socketio cors authentication for object detection and qa metrics requests by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2455\r\n  - ## Error analysis\r\n    - Data Imputation Fix in `erroranalysis` by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2436\r\n    - release erroranalysis 0.5.3 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2468\r\n    - update responsibleai and raiwidgets to erroranalysis 0.5.3 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2471\r\n- other\r\n  - update vscode settings for codeActionsOnSave to use new field instead of boolean by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2465\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fcompare\u002Fv0.32.1...v0.33.0","2023-12-28T21:53:33",{"id":211,"version":212,"summary_zh":213,"released_at":214},116316,"v0.32.1","## Description\r\nPatch update to use latest rai_core_flask 0.7.1 containing latest dashboard port range for private link AML workspace support.\r\n\r\n## v0.32.1\r\n\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - UI Improvement for Aggregate Object Detection Labels by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2431\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - Update raiwidgets to rai-core-flask v0.7.1 by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2437\r\n  - ## RAI Vision\r\n    - update responsibleai-text and responsibleai-vision packages to responsibleai 0.32.0 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2420\r\n    - release responsibleai-vision 0.3.3 and responsibleai-text 0.2.3 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2421\r\n    - Added aria-labels to dropdowns by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2417\r\n    - make elements under image explorer view tab keyboard accessible by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2422\r\n    - Added aria-labels for sliders and checkboxes\u002FDetailsList by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2419\r\n    - Added aria-labels to buttons by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2418\r\n    - fix switch cohort window getting clipped when zoomed in for accessibility by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2430\r\n    - fix vision dashboard flyout accessibility on smaller screen size by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2432\r\n    - fix image downloader failing with automl format on deserialize due to missing label transformations by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2435\r\n    - fix save and switch button text overflowing for accessibility by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2438\r\n    - Object Detection Model Overview missing-class fix for Cohorts by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2429\r\n- other\r\n  - Localized file check-in by OneLocBuild Task: Build definition ID 21368: Build ID 111947035 by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2433\r\n  - LEGO: Pull request from JUNO\u002Fhb_a12a4630-4852-4e7d-9cbc-c0e1117da1ab_20231125124253772 to main by @csigs in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2428\r\n  - Removed python 3.6 from CI-notebook gate by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2441\r\n  - Localized file check-in by OneLocBuild Task: Build definition ID 21368: Build ID 112224195 by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2434\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fcompare\u002Fv0.32.0...v0.32.1","2023-12-06T14:56:49",{"id":216,"version":217,"summary_zh":218,"released_at":219},116317,"v0.32.0","## Description\r\n\r\n## v0.32.0\r\n\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - add socket implementation to RAI dashboard object detection and question answering metrics calls by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2394\r\n    - Dashboard Port Updates by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2406\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - release rai-core-flask 0.7.0 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2398\r\n    - update raiwidgets to rai-core-flask v0.7.0 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2401\r\n    - Release new version of rai_core_flask by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2407\r\n    - fallback to request if socket connection fails for object detection and question answering metrics for backcompat by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2409\r\n    - Side Panel Bug Fix by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2414\r\n  - ## RAI Vision\r\n    - OD Explanation Loader Fix by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2381\r\n    - Optional support for IS_CROWD for OD COCO label format by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2395\r\n    - Import & data loading fixes for IC & multilabel notebooks by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2410\r\n    - fix RangeIndex error in error analysis for object detection models by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2411\r\n    - add average parameter to MeanAveragePrecision to specify micro or macro calculation by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2412\r\n  - ## Counterfactual\r\n    - Update requirements.txt to upgrade dice-ml to 0.11.0 by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2397\r\n    - Refer to predicted value from Counterfactual data by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2402\r\n  - ## Error analysis\r\n    - Accessibility fix for 'Error Analysis' header by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2415\r\n- other\r\n  - Localized file check-in by OneLocBuild Task: Build definition ID 21368: Build ID 107831016 by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2393\r\n  - LEGO: Pull request from JUNO\u002Fhb_a12a4630-4852-4e7d-9cbc-c0e1117da1ab_20231003114057604 to main by @csigs in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2384\r\n  - LEGO: Pull request from JUNO\u002Fhb_a12a4630-4852-4e7d-9cbc-c0e1117da1ab_20231030114314322 to main by @csigs in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2400\r\n  - Localized file check-in by OneLocBuild Task: Build definition ID 21368: Build ID 109426168 by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2408\r\n  - fix build errors due to changed exception messages and UCI repository expired certificate by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2413","2023-11-15T15:27:53",{"id":221,"version":222,"summary_zh":223,"released_at":224},116318,"v0.31.0","## v0.31.0\r\n\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - Heatmap Support for Object Detection by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2303\r\n    - Vision Explanation Dashboard render refactor by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2306\r\n    - Single Container in Image Data Explorer for Object Detection by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2330\r\n    - Error Labeling integration into Image Data Explorer for Agg Labels by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2334\r\n    - Detection info inside Object Detection Flyout by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2339\r\n    - Categorical Exif Metadata support for Vision Dashboard by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2348\r\n    - Error Labeling Integration in Table View by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2350\r\n    - Numerical Exif Metadata support for Image Dashboards by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2358\r\n    - add local feature importance info button and include shap values in description for RAI text dashboard by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2373\r\n    - Error Labeling Integration in Class View by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2354\r\n    - Sorting support for Object Detection labels by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2376\r\n    - Forecasting: generate time series as cohorts by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2355\r\n  - ## RAI utils\r\n    - Add pandas timestamp support to json serialization utility function in raiutils by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2320\r\n  - ## RAI tabular\r\n    - Forecasting: consume locally served model by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2319\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - accessibility: add aria label to RAI text dashboard slider by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2293\r\n    - Added ariaLabel to cohort dropdown by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2296\r\n    - UI Tests for Object Detection Widgets in Model Overview by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2295\r\n    - move text UI individual feature importance tests by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2299\r\n    - RAI Text Dashboard a11y: make scrollable pane focusable and use label attribute on slider by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2305\r\n    - Heatmap e2e test by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2239\r\n    - fix true y and pred y not appearing for QA text dashboard in table view by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2336\r\n    - fix failing object detection e2e jupyter notebook UI tests on linux by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2341\r\n    - Expanded Object Detection Flyout & Image Sizes by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2335\r\n    - add model explanation data to test app for question answering example on squad dataset by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2326\r\n    - fix alignment of true and predicted labels in vision data explorer by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2369\r\n    - fix choose cohorts close or cancel updating feature cohorts selection in flyout by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2366\r\n    - remove duplicate text in feature importance graph of the RAI Text dashboard by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2371\r\n    - remove non-functional page size selector from RAI Vision dashboard table view by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2370\r\n    - Backend and OD Notebook Dataset fixes by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2360\r\n    - Data Analysis Label Fix for OD by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2365\r\n    - Table View Search Fix for Multilabel & OD by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2367\r\n    - Enabling Heatmap test for Feature Cohort + MO test fixes by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2340\r\n  - ## RAI Text\r\n    - fix failing vision tests due to new RAI validation logic which doesn't allow non string dropped column names by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2302\r\n    - add text feature","2023-10-30T19:02:36",{"id":226,"version":227,"summary_zh":228,"released_at":229},116319,"v0.30.0","## v0.30.0\r\n\r\n- educational materials\r\n  - ## RAI Text\r\n    - Doc update for `responsibleai_text` by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2169\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - [Question Answering] Cache support for Model Overview Metrics by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2166\r\n    - Individual feature importance interpret QA - part1 by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2167\r\n    - Individual feature importance interpret QA by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2186\r\n    - Error Analysis UI & Template support for Object Detection by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2203\r\n    - add multiclass statistics to text and tabular RAI dashboards by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2222\r\n    - Add forecasting to raiwidgets and add a forecasting notebook by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2252\r\n  - ## RAI Text\r\n    - [Question Answering] Cache support for Model Overview Metrics by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2166\r\n    - update responsibleai-text and responsibleai-vision packages to responsibleai 0.29.0 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2172\r\n    - add base values and importances for ending tokens to QA explanations by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2178\r\n  - ## RAI Vision\r\n    - [Question Answering] Cache support for Model Overview Metrics by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2166\r\n    - update responsibleai-text and responsibleai-vision packages to responsibleai 0.29.0 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2172\r\n    - [Object Detection] Model Overview Cache Support Extension for torchmetrics by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2170\r\n    - OD model type support in raiutils for EA by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2227\r\n    - GPU support for OD metrics by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2266\r\n  - ## RAI Tabular\r\n    - OD model type support in responsibleai for EA by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2230\r\n  - ## Error analysis\r\n    - Error Analysis backend support for Object Detection by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2225\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - ModelAssessmentVision UI Test Template by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2174\r\n    - Multiclass Image Classification Notebook for UI testing by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2175\r\n    - Image Classification Notebook for UI testing #2175 by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2176\r\n    - Vision Data Analysis \u002F Data Balance E2E Tests + Refactor by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2183\r\n    - Vision Model Overview E2E Tests - Refactor by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2185\r\n    - OD ID Plot Hotfix by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2190\r\n    - Fix heatmap by removing erroneous color configuration by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2201\r\n    - Vision Data Explorer Table & Class View tests by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2220\r\n    - Vision Data Explorer Image Explorer Tests by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2188\r\n    - Enabled EA UI testing for OD by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2251\r\n    - CanvasTools OD Bounding Box Fix by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2267\r\n    - Add forecasting to raiwidgets and add a forecasting notebook by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2252\r\n    - add e2e UI notebook tests to DBPedia text classification notebook by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2265\r\n    - Endpoint Interruption Logic for Model Overview with OD by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2268\r\n    - Endpoint Interruption Logic for Model Overview with QA by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2274\r\n    - add e2e UI notebook tests to blbooksgenre text classification notebook by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2273\r\n    - add RAITextInsights question answering notebook example and tests by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-to","2023-09-05T14:05:47",{"id":231,"version":232,"summary_zh":233,"released_at":234},116320,"v0.29.0","## v0.29.0\r\n\r\n- educational materials\r\n  - Doc update for `responsibleai_vision` by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2164\r\n  - [Object Detection] User Flows \u002F Architectures by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2134\r\n  - [Object Detection] DFD Documentation by @natalie-isak in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2138\r\n  - [Object Detection] Explanation Documentation by @natalie-isak in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2136\r\n  - [Object Detection] CanvasTools documentation by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2133\r\n  - [Object Detection] Model Overview Object Detection Documentation by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2132\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - [Object Detection] Cache support for Model Overview Metrics by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2158\r\n    - [Image Explorer] CanvasTools Image Loading support for Object Detection by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2097\r\n    - [Object Detection] [Image Explorer] Bounding Box Support with CanvasTools on Flyout by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2116\r\n- bug fixes and tests\r\n\r\n  - ## Responsible AI Dashboard\r\n    - update responsibleai-text and responsibleai-vision packages to latest responsibleai 0.28.0 release by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2159\r\n    - Robot page exception in DatasetUtils where jointDataset.metaDict[property] is undefined by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2148\r\n    - Remove console log by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2125\r\n    - Refactor- Add datasetCohorts to ModelAssessementContext && create datasetCohort && add compositeFilter in DatasetCohort by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2119\r\n  - ## RAIInsights\r\n    - raise error instead of warning when a user has missing data and add c… by @hawestra in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2143\r\n    - Add columns names to `create_multiclass_classification_dataset` by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2114\r\n    - add validation target column type in the classification scenario by @hawestra in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2127\r\n    - update econml to latest 0.14.1 release by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2112\r\n    - Update dice-ml to 0.10.0 by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2145\r\n  - ## Causal\r\n    - Fix individual causal plot not rendering for 5k points by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2113\r\n    - Add loading spinner for causal policy tab by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2121\r\n  - ## RAI Text\r\n    - merge RAI Text repository to RAI toolbox by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2118\r\n    - add responsibleai-text package tests to build gate by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2139\r\n    - release responsibleai-text 0.1.4 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2163\r\n    - add release scripts for responsibleai-text and responsibleai-vision pypi packages by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2156\r\n  - ## RAI Vision\r\n    - add responsibleai-vision package to responsible-ai-toolbox by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2135\r\n    - add responsibleai-vision package to gated CI tests by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2152\r\n    - add retry logic to common vision utils to make tests more robust by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2154\r\n    - add device parameter to responsibleai_vision package and pass to model wrappers by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2157\r\n    - release responsibleai-vision 0.2.3 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2162\r\n  - ## RAI Test Utils\r\n    - Update rai_test_utils in erroranalysis and responsibleai to 0.3.0 by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2111\r\n    - `rai_test_utils`: Convert all namespaces in `__all__` list as strings by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2128\r\n\r\n- ## other\r\n  - fix build failures due to lint errors for onelocbuild yml by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2137\r\n  - Skipping python 3.7 tests on MACOS by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible","2023-07-07T00:29:03",{"id":236,"version":237,"summary_zh":238,"released_at":239},116321,"v0.28.0","## v0.28.0\r\n\r\n- educational materials\r\n  - Correct test command and add clarification on GPG Key to CONTRIBUTING.md by @ziqi-ma in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2082\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - QA Scenario metrics template by @mehektulsyan in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2034\r\n    - Reset ComboBox Option for Object Detection by @natalie-isak in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2049\r\n    - Integrate Question answering metrics from python backend by @mehektulsyan in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2054\r\n    - Display details of points and bubble for highcharts in full-screen mode by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2060\r\n    - Show message to select bubble after bubble plot renders by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2070\r\n    - Add dataset metadata to `IDataset` by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2071\r\n    - enable confusion matrix for binary and multiclass text and vision RAI dashboards by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2108\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - remove QA statistics which were mistakenly added to multilabel metrics by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2048\r\n    - fix multilabel RAI Vision Dashboard table view labels squished by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2057\r\n    - [Big data] Few bug fixes and changes by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2056\r\n    - fix outline bug and add contributors to readme by @hawestra in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2081\r\n    - Fix background color not changing with theme switch by @hawestra in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2084\r\n    - Revert \"fix outline bug\" by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2086\r\n    - Fix table view multi-selection by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2087\r\n    - [Big Data] Fix create cohort, filter range for index and dataset by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2089\r\n    - Fix background color and outline by @hawestra in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2092\r\n    - [Big data] Add telemetry by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2093\r\n    - Fix bug -\"All data\" cohort detection is based on localized strings by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2094\r\n    - Retain filter selections when editing filter by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2096\r\n    - Adding Class Name To OD Dropdown by @natalie-isak in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2098\r\n    - Refactor cohortList to be less than 300 lines & add datasetCohorts to context by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2103\r\n  - ## RAIInsights\r\n    - Use cohort filter constatns from erroranalysis\u002Fresponsibleai in tests by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2047\r\n    - fix build failures by pinning statsmodels to older version due to numpy incompatibility by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2053\r\n    - Raise warning in case the test data has missing values by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2065\r\n    - Add user feedback to compute() function by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F697\r\n    - Remove un-needed `requests` dependency by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2083\r\n    - Raise value error for min max computation of feature range by @kicha0 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2072\r\n    - Move `common_utils.py` into `responsibleai_dashboard` directory by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2095\r\n    - Port over logic of `load_adult_income_dataset` from `dice-ml` to unblock gates by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2106\r\n    - Fix dimension mismatch error when dropped_features contains cat_features by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2099\r\n  - ## Error Analysis\r\n    - Prevent overwriting of dataset in `erroranalysis` by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2050\r\n    - fix multilabel erroranalysis metrics dropdown appearing blank by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2063\r\n    - release erroranalysis 0.4.4 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2105\r\n    - update responsibleai and raiwidgets ","2023-06-10T03:55:48",{"id":241,"version":242,"summary_zh":243,"released_at":244},116322,"v0.27.0","## v0.27.0\r\n\r\n- educational materials\r\n  - Update explanation-dashboard-employee-attrition.ipynb to remove deprecated `iteritems()` and use `items()` instead by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2030\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - Drop Down For Object Detection by @natalie-isak in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2010\r\n    - Renaming Data Characteristics Tab by @natalie-isak in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2032\r\n    - Object Detection UX & Documentation updates by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2037\r\n    - Ensuring backwards compat b\u002Fw OD scenario and other vision scenarios by @natalie-isak in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2039\r\n  - ## Model Overview\r\n    - [Model Overview] Object Detection Metrics support by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2025\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - [Refactor] Add preBuiltCohort to datasetCohorts in modelAssessmentContext by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2007\r\n    - Fix bugs 1.Save and switch doesn't do anything in cohort edit panel 2.Unable to edit cohort filters in certain scenario by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2040\r\n  - ## RAIInsights\r\n    - make train data optional on RAIBaseInsights class by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2029\r\n    - responsibleai: Add test case for dataset having boolean features by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2042\r\n  - ## Error Analysis\r\n    - fix indexing error when creating a filtered error analysis tree view with a dataset that contains categoricals by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2026\r\n    - add EBM and LightGBM with TreeShap methods to calculate error correlation in error analysis by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2018\r\n    - release erroranalysis v0.4.3 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2043\r\n    - update responsibleai and raiwidgets to erroranalysis 0.4.3 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2045\r\n    - fix transient test failures by increasing timeout for gbm shap method by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2044\r\n  - ## RAI Utils\r\n    - Move import of `cohort` module from `raiwidgets` to `raituils` by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2027\r\n  - ## Interpret\r\n    - update responsibleai to interpret-community 0.29.0 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2031\r\n- ## other\r\n  - Add python 3.10 tests by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2033\r\n  - fix failing builds by upgrading pip-tools after latest pip release by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2038\r\n  - Add 3.10 classifier for python packages by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2035\r\n  - Upgrade python from 3.7 to 3.10 in few workflows by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2041\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fcompare\u002Fv0.26.0...v0.27.0","2023-04-27T13:14:59",{"id":246,"version":247,"summary_zh":248,"released_at":249},116323,"v0.26.0","## v0.26.0\r\n\r\n- educational materials\r\n  - Update notebooks for categorical features to pass through FeatureMetadata by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2011\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - Adding Object Detection Fridge Data by @natalie-isak in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1998\r\n  - ## RAIInsights\r\n    - Support forecasting in RAIInsights by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1948\r\n  - ## Model Overview\r\n    - [Model Overview] Metrics Template support with OD-specific optional args by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2002\r\n    - Object Detection Model Overview Flask template by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2004\r\n    - Model Overview: Object Detection Widgets for Aggregate Methods, Class selection, & IoU Threshold by @Advitya17 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1997\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - Forecasting: UI code adjustments to renamed datetime_features and time_series_id_features by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1993\r\n    - [Refactor]1.Add DatasetCohort 2. Move FeatureFlights to core by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2003\r\n    - [Refactor]Add columnRanges and modelType to ModelAssessmentContext by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2006\r\n  - ## RAIInsights\r\n    - Add validations for model predictions by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2008\r\n    - Add validations if target column is included in `FeatureMetadata` by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2013\r\n    - try pickling model instead of failing based on state functions and change to UserConfigValidationException by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2014\r\n    - update dependencies for rai-core-flask, raiwidgets and responsibleai packages by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2016\r\n    - Fix deprecated DataFrame indexer type in Error Analysis package by @JarvisG495 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2019\r\n  - ## Error Analysis\r\n    - fix bug in tree view where slider params are not used by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1990\r\n    - fix error analysis showing float values with many zeros for some classification task nodes by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1991\r\n    - release erroranalysis v0.4.2 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2000\r\n    - update raiwidgets and responsibleai to erroranalysis 0.4.2 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2001\r\n\r\n## New Contributors\r\n* @natalie-isak made their first contribution in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1998\r\n* @Advitya17 made their first contribution in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F2002\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fcompare\u002Fv0.25.0...v0.26.0","2023-04-04T21:16:32",{"id":251,"version":252,"summary_zh":253,"released_at":254},116324,"v0.25.0","## v0.25.0\r\n\r\n- educational materials\r\n  - Update README badges by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1933\r\n  - DOC improving several error strings by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1943\r\n  - Add `rai_test_utils` badge in README.md by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1953\r\n  - improve installation instructions for onboarding to RAI toolbox repository by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1962\r\n  - fix outdated example link in Error Analysis README by @JarvisG495 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1965\r\n  - Update example notebook links in README docs by @JarvisG495 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1970\r\n  - Create README.md for python tests by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1984\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - Forecasting RAI dashboard: expand test data, add categorical what if by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1921\r\n  - ## Dataset Explorer\r\n    - [Large data Data explorer] Add individual data points - data explorer for large data. by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1919\r\n    - add missing true and predicted labels to axis config for dataset explorer in multilabel text and vision scenarios by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1960\r\n  - ## Counterfactual\r\n    - [Large data counterfactuals] Use data from local counterfactuals to display in counterfactuals panel by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1913\r\n  - ## Interpret\r\n    - [Large data feature importance] Add individual feature importance by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1929\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - Fix typo in variable name `disaggregatedAnalysisBaseCohortDislaimer` by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1855\r\n    - add real model for dev env to have better test data by @xuke444 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1893\r\n    - Remove `setstate()` out of render function in `RegressionDistributionChart.tsx` by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1922\r\n    - Unify responsibleai constants by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1925\r\n    - Fix bubble chart render issue due to exception in `SidePanel` by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1941\r\n    - Fix table view feature value && scroll bar by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1954\r\n    - [Large counterfactuals & explanations] Do not call local counterfactuals\u002Fexplanations on unselecting the point by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1957\r\n    - [Large data] Map classification outcome to string values, dither, count changes by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1955\r\n    - add (none) category for empty predicted or true labels in multilabel vision dashboard for data characteristics by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1961\r\n    - fix CI notebook build via updated yarn command flags by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F\r\n    - upgrade tslib to 2.5.0 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1973\r\n    - Forecasting: improve description in no-features case by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1978\r\n    - Forecasting: fix dropdown scroll issue by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1979\r\n    - Forecasting: restrict transformation names to 50 chars by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1981\r\n    - Forecasting: create a better UX for value inputs by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1985\r\n    - Forecasting: turn off Error Analysis by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1987\r\n    - Forecasting: fix button text to say \"switch time series\" by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1988\r\n    - Forecasting: validate characters for name field in transformation creation dialog by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1986\r\n  - ## RAIInsights\r\n    - raiwidgets tests: Change `post_data` from tuple to list by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1920\r\n    - Remove duplicate load() and save() for large data by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1930\r\n    - Deprecate categorical_features argument in favor of feature_metadata.categoric","2023-03-02T16:52:45",{"id":256,"version":257,"summary_zh":258,"released_at":259},116325,"v0.24.1","## v0.24.1\r\n\r\n- educational materials\r\n  - Update some sample notebooks to include dropped_features by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1911\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - Add forecasting dashboard - UI - time series selection & forecast comparison by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1894\r\n    - Confusion Matrix by @t-wangjohn in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1794\r\n    - Update release workflows to include forecasting npm package by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1901\r\n    - Add forecasting dashboard - UI - transformation creation, table, and comparison by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1874\r\n  - ## Counterfactual\r\n    - [Big data] Counterfactuals - Part2 by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1832\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - Remove \u003Ch3> tags and replace with \u003CText> by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1900\r\n    - raiwidgets: Fix url in setup.py by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1906\r\n    - Enable logarithmic scaling, reflect cohort change and add add Revert to bubble chart button by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1905\r\n  - ## RAIInsights\r\n    - Move import to top level from within the function by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1898\r\n    - Raise `UserConfigValidationException` when `treatment_feature` is empty list by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1904\r\n  - ## Data balance\r\n    - Make `DataBalanceManager.compute()` throw warning instead of exception by @ms-kashyap in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1902\r\n  - ## Error Analysis\r\n    - fix small data erroring out on mutual info score for error analysis guidance by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1907\r\n    - release erroranalysis v0.4.0 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1909\r\n    - update raiwidgets and responsibleai to erroranalysis 0.4.0 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1910\r\n    - Fix model wrapper to identify model type, remove task type from error analysis by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1912\r\n  - ## Interpret\r\n    - Fix unwrap error in individual feature importance by @hawestra in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1897\r\n- ## other\r\n  - Exempt only files from max-lines rule that are already in violation, but not new files (part 2\u002F2) by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1899\r\n\r\n## New Contributors\r\n* @t-wangjohn made their first contribution in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1794\r\n* @hawestra made their first contribution in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1897\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fcompare\u002Fv0.24.0...v0.24.1","2023-01-18T23:35:01",{"id":261,"version":262,"summary_zh":263,"released_at":264},116326,"v0.24.0","## v0.24.0\r\n\r\n\r\n- educational materials\r\n  - Add more linting instructions to CONTRIBUTING.md by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1810\r\n  - Update some sample notebooks to include dropped_features by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1836\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - add binary classification support for RAI text and vision scenarios by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1778\r\n    - add multilabel support to RAI dashboard and multilabel text classification covid events dataset by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1811\r\n    - Enable dropped_features for error analysis, data explorer and explainer, counterfactual and causal managers by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1824\r\n  - ## Error Analysis\r\n    - Always display tree branching decisions by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1870\r\n    - Show branching decisions on hover over on unselected tree nodes by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1881\r\n  - ## NLP Feature Extractors\r\n  - ## RAI test utils\r\n    - Add rai_test_utils package by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1777\r\n    - Add release-rai-test-utils.yml to release rai_test_utils package by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1797\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - Update e2e visit() function to not refresh the page if provided name doesn't exist in RAINotebookNames by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1792\r\n    - Fix TypeError: Cannot read properties of undefined (reading 'sort') by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1798\r\n    - Fix multiple Scroll bars appear on Model demo page after setting the page to 320\\*256px viewport by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1796\r\n    - Fix ‘Close button’ and ‘Spin button’ present in new cohort dialog does not have any title or aria-label by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1806\r\n    - Fix fast pass accessibility bugs by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1817\r\n    - Add min-width for highchart and fix padding by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1818\r\n    - Add user feature name to box plot hover over by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1819\r\n    - update to latest 16.x node version for tests by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1809\r\n    - Simplify route generation logic for testing with flights by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1530\r\n    - Use \"lower whisker\" and \"upper whisker\" instead of misleading \"minimum\" and \"maximum\" on boxplot tooltips by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1825\r\n    - Remove unused flight \"newSdkEndpoints\" by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1831\r\n    - Gate call to probability distribution end point by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1839\r\n    - Disable partial dependence plots for large data scenarios by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1843\r\n    - add multilabel text example and fix correct and incorrect groups in data analysis by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1848\r\n    - Add number of cohort samples inside generateMetrics() by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1849\r\n    - add multilabel RAI vision example for fridge multilabel dataset by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1844\r\n    - update raiwidgets to rai-core-flask v0.5.0 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1851\r\n    - Remove warning for 'dropped_features are not in use currently.' by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1871\r\n    - add multilabel model overview metrics by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1882\r\n    - Populate probability box chart data on mount by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1883\r\n    - Improve spacing in RAI dashboard by @romanlutz in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1875\r\n  - ## RAIInsights\r\n    - Update requirements.txt to pin econml to 0.14.0 by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1813\r\n    - Remove pip pinning to unblock RAI gates by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1826\r\n    - Update","2023-01-11T19:12:58",{"id":266,"version":267,"summary_zh":268,"released_at":269},116327,"v0.23.0","## v0.23.0\r\n\r\n- educational materials\r\n  - fixed the link to developer blog for housing example by @aminadibi in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1713\r\n  - Fix developer blog link in notebook by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1715\r\n  - readme-enhancements by @mesameki in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1704\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - Update UT - use new ErrorCohort() to mock data by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1711\r\n    - [A11y] 1.Error analysis tree node text alignment 2. Add ariaLabel for new cohort panel 3. Model overview chart fix by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1719\r\n    - reduce file size for e2e by @xuke444 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1723\r\n    - fix max line line error for vision by @xuke444 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1725\r\n    - clean up fairness by @xuke444 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1724\r\n    - rollback local importance selection by @xuke444 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1726\r\n    - add fridge dataset to test app for RAI vision dashboard and fix bug with large images in flyout by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1730\r\n    - fix images in data characteristics panel hiding the success\u002Ffailure instances by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1731\r\n    - add test watch mode by @xuke444 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1728\r\n    - refactor vision dashboard to remove getDerivedStateFromProps by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1732\r\n    - [A11y]Show Error analysis tree focus && Add Screen reader for 1. New cohort panel, 2. Add component callout 3. Callout tooltip by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1727\r\n    - [A11y] Fix bug - Screen reader is not announcing the associated information with axis selection radio buttons in data analysis chart view by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1736\r\n    - Fix cohort deletion not working by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1738\r\n    - fix e2e watch mode by @xuke444 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1740\r\n    - fix an reference issue for chart lib by @xuke444 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1741\r\n    - fix chart that generates duplicate plotly data by @xuke444 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1744\r\n    - [A11y] Fix table view keyboard access by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1739\r\n    - add data characteristics search and fix switching between tabs in RAI vision dashboard by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1745\r\n    - fix RAI dashboard for vision models crashing when selecting individual datapoints in data analysis by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1733\r\n    - remove duplicate settings button in RAI dashboard for images by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1747\r\n    - in data characteristics of RAI vision dashboard, fix bug where labels to display is reset when changing number of rows by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1748\r\n    - [A11y][screen reader] Announce component added after adding a component by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1751\r\n    - [A11y] 1. Hide chart heading & 2. Fix Add component callout high contrast focus indicator by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1758\r\n    - Fix full name not visible for features in highcharts by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1761\r\n    - Filter out undefined filters to resolve a robot page exception by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1764\r\n  - ## RAIInsights\r\n    - Correct predict call to predict_proba by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1717\r\n    - Port over cohort filtering tests for 'Predicted Y' cohort to responsibleai by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1709\r\n    - Raise exception in case categorical feature list doesn't have all string features by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1720\r\n    - Pass classes into the sklearn confusion_matrix by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1674\r\n    - fix flaky test test_rai_insights_add_save_load_save on dataset fetch by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1734\r\n ","2022-10-25T17:58:30",{"id":271,"version":272,"summary_zh":273,"released_at":274},116328,"v0.22.0","## v0.22.0\r\n\r\n- educational materials\r\n  - Notebook change showing the usage of identity feature and feature metadata by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1627\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - Vision Dashboard Table View and initial explanation flyout by @jamesbchao in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1621\r\n    - Vision dashboard backend without data characteristics tab by @jamesbchao in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1668\r\n    - Vision dashboard data characteristics tab by @jamesbchao in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1675\r\n    - Vision dashboard cohorts by @jamesbchao in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1686\r\n    - Vision model overview metrics by @jamesbchao in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1688\r\n  - ## Error Analysis\r\n  - ## NLP Feature Extractors\r\n    - Add the code for the nlp_feature_extractors to the RAI Toolbox by @ilmarinen in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1580\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - [Accessibility] Data analysis scatter chart - use symbol and color to differentiate data series by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1667\r\n    - Bump moment-timezone from 0.5.34 to 0.5.35 by @dependabot in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1677\r\n    - Add validations for dashboard endpoints by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1672\r\n    - [Accessibility][reflow] Error analysis tree map & Data explorer & Feature importance Reflow by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1678\r\n    - [Accessibility][reflow] Error Analysis heat map & Model overview & Counterfactual & Causal analysis by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1680\r\n    - [localization] Localize high chart context menu items by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1691\r\n    - for RAI text dashboard, remove radio buttons and legend text when avg of abs value local importances selected by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1683\r\n    - Vision dashboard UI Fixups by @jamesbchao in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1687\r\n    - touch ups from designer feedback to RAI dashboard for text data by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1695\r\n  - ## RAIInsights\r\n    - Update econml to 0.13.1 by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1660\r\n    - [responsibleai]: Add tests for classification outcome (multiclass) and regression error by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1662\r\n    - JSON serialize the endpoint output and add tests by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1679\r\n    - Cache predict() and predict_proba() output by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1702\r\n    - Reduce the train and test dataset sizes in responsibleai to reduce test time by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1630\r\n  - ## Counterfactual\r\n  - ## Causal\r\n    - Add validations for heterogeneity_features in causal manager by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1659\r\n    - Fix localization for causal tooltips by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1690\r\n  - ## Error Analysis\r\n    - [Accessibility] Error analysis screen reader by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1663\r\n    - Add metric computation of mean prediction by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1693\r\n    - Add support for 'True Y' filter for cohort filtering by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1685\r\n    - Avoid additional predict call in case of ModelAnalyzer by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1701\r\n    - release erroranalysis v0.3.10 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1706\r\n  - ## Dataset Explorer\r\n    - refactor table view into data analysis component by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1673\r\n    - add mini-table to feature importances view by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1682\r\n  - ## Model Overview\r\n    - [Model Overview] Box plot data for classification probabilities from SDK endpoint by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1689\r\n    - Change 'Accuracy' to 'Accuracy score' in ModelOverview by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1700\r\n  - ## Interpret\r\n    - [Accessibility][reflow] Individual conditional expectation plot by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicro","2022-09-13T04:31:53",{"id":276,"version":277,"summary_zh":278,"released_at":279},116329,"v0.21.0","## v0.21.0\r\n\r\n- educational materials\r\n  - [docs] Add Data Balance README by @ms-kashyap in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1567\r\n  - Replace SynapseML links with RAIToolbox links by @ms-kashyap in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1592\r\n- new features\r\n  - ## Responsible AI Dashboard\r\n    - Data Balance UX Part 2: Add visualizations as a flight by @ms-kashyap in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1512\r\n    - add new interpret-vision library skeleton by @jamesbchao in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1586\r\n    - add support for multiclass local importance values for text by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1601\r\n    - Vision explanation dashboard -- Image explorer view by @jamesbchao in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1587\r\n  - ## Error Analysis\r\n    - Make generic class interface for cohort based filtering by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1597\r\n- bug fixes and tests\r\n  - ## Responsible AI Dashboard\r\n    - update release workflows to include interpret-text npm package by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1582\r\n    - Add RAI prefix for OSS UI telemetry events by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1584\r\n    - Bump moment from 2.29.2 to 2.29.4 by @dependabot in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1538\r\n    - Bump eventsource from 1.0.7 to 1.1.1 by @dependabot in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1456\r\n    - [Bugfix] Fix Dropdown bug and Update styling for feature balance measures by @ms-kashyap in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1583\r\n    - Refactor tests in test_responsibleai_dashboard_input.py to have common success and failure criteria by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1591\r\n    - update raiwidgets and responsibleai to erroranalysis 0.3.6 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1588\r\n    - Fix tooltip location showing up incorrectly due to conflicting ids by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1599\r\n    - Update e2e test to expand correct predictions caret symbol by @vinuthakaranth in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1603\r\n    - Updates to RAI text dashboard - update test app values to latest shap release, fix predicted label, limit selection to one row, fix random crash by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1605\r\n    - [300 line limit] Refactor AxisConfigDialog: create AxisConfigBinOptions and AxisConfigChoiceGroup component by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1609\r\n    - [300 line limit] Refactor CohortEditor, OverallMetricChart, set max-lines rule for core-ui package by @tongyu-microsoft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1612\r\n    - design updates for the RAI text dashboard local importances view by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1611\r\n    - [Accessibility] Fix all existing accessibility bugs by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1613\r\n    - Move CalculateBoxPlot.test.tsx to core-ui by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1620\r\n    - Add feature metadata to IDataset and JointDataset by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1614\r\n    - Remove console.log() statement from CohortEditor.tsx by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1628\r\n    - update raiwidgets and responsibleai to erroranalysis 0.3.8 by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1629\r\n    - add new interpret-vision package to npm publish workflows in build and release gates by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1615\r\n    - [Accessibility] fix several accessibility bugs by @RubyZ10 in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1632\r\n    - add UI tests for RAI text dashboard by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1636\r\n    - Fix typo for indices by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1640\r\n    - Add extra space between Identity feature and the actual feature name by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1641\r\n    - Use JointDataset API in ProcessPreBuiltCohort.ts by @gaugup in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1635\r\n    - add wrapping for long text documents to text highlighting component in RAI dashboard for text by @imatiach-msft in https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fresponsible-ai-toolbox\u002Fpull\u002F1643\r\n    - add tests to RAI text dashboard importances part 2 by @imatiach-msft in https:\u002F\u002Fg","2022-08-22T13:57:19"]