[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-hurshd0--must-read-papers-for-ml":3,"tool-hurshd0--must-read-papers-for-ml":61},[4,18,26,36,44,53],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":17},4358,"openclaw","openclaw\u002Fopenclaw","OpenClaw 是一款专为个人打造的本地化 AI 助手，旨在让你在自己的设备上拥有完全可控的智能伙伴。它打破了传统 AI 助手局限于特定网页或应用的束缚，能够直接接入你日常使用的各类通讯渠道，包括微信、WhatsApp、Telegram、Discord、iMessage 等数十种平台。无论你在哪个聊天软件中发送消息，OpenClaw 都能即时响应，甚至支持在 macOS、iOS 和 Android 设备上进行语音交互，并提供实时的画布渲染功能供你操控。\n\n这款工具主要解决了用户对数据隐私、响应速度以及“始终在线”体验的需求。通过将 AI 部署在本地，用户无需依赖云端服务即可享受快速、私密的智能辅助，真正实现了“你的数据，你做主”。其独特的技术亮点在于强大的网关架构，将控制平面与核心助手分离，确保跨平台通信的流畅性与扩展性。\n\nOpenClaw 非常适合希望构建个性化工作流的技术爱好者、开发者，以及注重隐私保护且不愿被单一生态绑定的普通用户。只要具备基础的终端操作能力（支持 macOS、Linux 及 Windows WSL2），即可通过简单的命令行引导完成部署。如果你渴望拥有一个懂你",349277,3,"2026-04-06T06:32:30",[13,14,15,16],"Agent","开发框架","图像","数据工具","ready",{"id":19,"name":20,"github_repo":21,"description_zh":22,"stars":23,"difficulty_score":10,"last_commit_at":24,"category_tags":25,"status":17},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,"2026-04-05T11:01:52",[14,15,13],{"id":27,"name":28,"github_repo":29,"description_zh":30,"stars":31,"difficulty_score":32,"last_commit_at":33,"category_tags":34,"status":17},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",144730,2,"2026-04-07T23:26:32",[14,13,35],"语言模型",{"id":37,"name":38,"github_repo":39,"description_zh":40,"stars":41,"difficulty_score":32,"last_commit_at":42,"category_tags":43,"status":17},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。",107888,"2026-04-06T11:32:50",[14,15,13],{"id":45,"name":46,"github_repo":47,"description_zh":48,"stars":49,"difficulty_score":32,"last_commit_at":50,"category_tags":51,"status":17},4721,"markitdown","microsoft\u002Fmarkitdown","MarkItDown 是一款由微软 AutoGen 团队打造的轻量级 Python 工具，专为将各类文件高效转换为 Markdown 格式而设计。它支持 PDF、Word、Excel、PPT、图片（含 OCR）、音频（含语音转录）、HTML 乃至 YouTube 链接等多种格式的解析，能够精准提取文档中的标题、列表、表格和链接等关键结构信息。\n\n在人工智能应用日益普及的今天，大语言模型（LLM）虽擅长处理文本，却难以直接读取复杂的二进制办公文档。MarkItDown 恰好解决了这一痛点，它将非结构化或半结构化的文件转化为模型“原生理解”且 Token 效率极高的 Markdown 格式，成为连接本地文件与 AI 分析 pipeline 的理想桥梁。此外，它还提供了 MCP（模型上下文协议）服务器，可无缝集成到 Claude Desktop 等 LLM 应用中。\n\n这款工具特别适合开发者、数据科学家及 AI 研究人员使用，尤其是那些需要构建文档检索增强生成（RAG）系统、进行批量文本分析或希望让 AI 助手直接“阅读”本地文件的用户。虽然生成的内容也具备一定可读性，但其核心优势在于为机器",93400,"2026-04-06T19:52:38",[52,14],"插件",{"id":54,"name":55,"github_repo":56,"description_zh":57,"stars":58,"difficulty_score":10,"last_commit_at":59,"category_tags":60,"status":17},4487,"LLMs-from-scratch","rasbt\u002FLLMs-from-scratch","LLMs-from-scratch 是一个基于 PyTorch 的开源教育项目，旨在引导用户从零开始一步步构建一个类似 ChatGPT 的大型语言模型（LLM）。它不仅是同名技术著作的官方代码库，更提供了一套完整的实践方案，涵盖模型开发、预训练及微调的全过程。\n\n该项目主要解决了大模型领域“黑盒化”的学习痛点。许多开发者虽能调用现成模型，却难以深入理解其内部架构与训练机制。通过亲手编写每一行核心代码，用户能够透彻掌握 Transformer 架构、注意力机制等关键原理，从而真正理解大模型是如何“思考”的。此外，项目还包含了加载大型预训练权重进行微调的代码，帮助用户将理论知识延伸至实际应用。\n\nLLMs-from-scratch 特别适合希望深入底层原理的 AI 开发者、研究人员以及计算机专业的学生。对于不满足于仅使用 API，而是渴望探究模型构建细节的技术人员而言，这是极佳的学习资源。其独特的技术亮点在于“循序渐进”的教学设计：将复杂的系统工程拆解为清晰的步骤，配合详细的图表与示例，让构建一个虽小但功能完备的大模型变得触手可及。无论你是想夯实理论基础，还是为未来研发更大规模的模型做准备",90106,"2026-04-06T11:19:32",[35,15,13,14],{"id":62,"github_repo":63,"name":64,"description_en":65,"description_zh":66,"ai_summary_zh":66,"readme_en":67,"readme_zh":68,"quickstart_zh":69,"use_case_zh":70,"hero_image_url":71,"owner_login":72,"owner_name":73,"owner_avatar_url":74,"owner_bio":75,"owner_company":76,"owner_location":77,"owner_email":76,"owner_twitter":76,"owner_website":76,"owner_url":78,"languages":76,"stars":79,"forks":80,"last_commit_at":81,"license":82,"difficulty_score":83,"env_os":84,"env_gpu":85,"env_ram":85,"env_deps":86,"category_tags":89,"github_topics":91,"view_count":32,"oss_zip_url":76,"oss_zip_packed_at":76,"status":17,"created_at":104,"updated_at":105,"faqs":106,"releases":107},5324,"hurshd0\u002Fmust-read-papers-for-ml","must-read-papers-for-ml","Collection of must read papers for Data Science, or Machine Learning \u002F Deep Learning Engineer","must-read-papers-for-ml 是一个专为数据科学、机器学习及深度学习领域打造的精选论文库。面对海量且复杂的学术文献，从业者往往难以筛选出真正具有奠基意义或前沿价值的核心文章，而该资源库正是为了解决这一痛点而生。它系统地整理了从数据预处理、统计建模基础，到模型评估、异常检测、梯度提升算法（如 XGBoost、LightGBM）以及模型可解释性等关键主题的必读论文、综述与技术博客。\n\n这份清单不仅提供了高质量的文献链接，还独具匠心地通过金牌、银牌、铜牌标识对阅读优先级进行了排序，帮助读者规划学习路径。项目维护者特别强调，阅读高难度数学推导的论文需要耐心与坚持，鼓励用户反复研读直至融会贯通。此外，该项目保持开放协作，欢迎社区共同更新与维护，确保内容的时效性。\n\nmust-read-papers-for-ml 非常适合机器学习工程师、数据科学家、算法研究人员以及相关专业的学生使用。无论是希望夯实理论基础的新手，还是寻求技术突破的资深开发者，都能从中快速定位到高价值资料，高效构建自己的知识体系，避免在文献海洋中迷失方向。","# Must Read Papers for Data Science, ML, and DL\n### Curated collection of Data Science, Machine Learning and Deep Learning papers, reviews and articles that are on must read list.\n\n---\n\n> NOTE: :construction: in process of updating, let me know what additional papers, articles, blogs to add I will add them here. \n\n### How to use  \n> :point_right: :star: this repo\n\n\n\n## Contributing\n- :point_right: :arrows_clockwise: Please feel free to [Submit Pull Request](https:\u002F\u002Fgithub.com\u002Fhurshd0\u002Fmust-read-papers-for-ml\u002Fpulls), if links are broken, or I am missing any important papers, blogs or articles.\n\n[![Maintenance](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FMaintained%3F-yes-green.svg)](https:\u002F\u002Fgithub.com\u002Fhurshd0\u002Fmust-read-papers-for-ml\u002Fgraphs\u002Fcommit-activity)\n\n### :point_down: READ THIS :point_down:\n\n- :point_right: Reading paper with heavy math is hard, it takes time and effort to understand, most of it is dedication and motivation to not quit, don't be discouraged, read once, read twice, read thrice,... until it clicks and blows you away.\n\n\n:1st_place_medal: - Read it first\n\n:2nd_place_medal: - Read it second \n\n:3rd_place_medal: - Read it third\n\n---\n\n## Data Science\n\n### :bar_chart: Pre-processing & EDA\n\n:1st_place_medal: :page_facing_up:[Data preprocessing - Tidy data - by Hadley Wickham](https:\u002F\u002Fvita.had.co.nz\u002Fpapers\u002Ftidy-data.pdf)\n\n### :notebook: General DS\n\n:1st_place_medal: :page_facing_up: [Statistical Modeling: The Two Cultures - by Leo Breiman](https:\u002F\u002Fprojecteuclid.org\u002Fdownload\u002Fpdf_1\u002Feuclid.ss\u002F1009213726)\n\n:2nd_place_medal: :page_facing_up: [A study in Rashomon curves and volumes: A new perspective on\ngeneralization and model simplicity in machine learning](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1908.01755.pdf)\n\n- :video_camera: [KDD 2019 Cynthia Rudin's Keynote](https:\u002F\u002Fyoutu.be\u002FwL4X4lG20sM)\n\n:1st_place_medal: :page_facing_up: [Frequentism and Bayesianism: A Python-driven Primer by Jake VanderPlas](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1411.5018.pdf)\n\n---\n\n## Machine Learning\n\n### :dart: General ML\n\n:1st_place_medal: :page_facing_up: [Model Evaluation, Model Selection, and Algorithm Selection in Machine Learning - by Sebastian Raschka](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1811.12808.pdf)\n\n:1st_place_medal: :page_facing_up: [A Brief Introduction into Machine Learning - by Gunnar Ratsch](https:\u002F\u002Fevents.ccc.de\u002Fcongress\u002F2004\u002Ffahrplan\u002Ffiles\u002F105-machine-learning-paper.pdf)\n\n:3rd_place_medal: :page_facing_up: [An Introduction to the Conjugate Gradient Method Without the Agonizing Pain - by Jonathan Richard Shewchuk](http:\u002F\u002Fwww.cs.cmu.edu\u002F~quake-papers\u002Fpainless-conjugate-gradient.pdf)\n\n:3rd_place_medal: :page_facing_up: [On Model Stability as a Function of Random Seed](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1909.10447)\n\n### :mag: Outlier\u002FAnomaly detection\n\n:1st_place_medal: :newspaper: [Outlier Detection : A Survey](https:\u002F\u002Fpdfs.semanticscholar.org\u002F912b\u002F0b7879ca99bf654a26bbb0d50d4b8e0ed6c0.pdf)\n\n### :rocket: Boosting\n\n:2nd_place_medal: :page_facing_up: [XGBoost: A Scalable Tree Boosting System](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1603.02754.pdf)\n\n:2nd_place_medal: :page_facing_up: [LightGBM: A Highly Efficient Gradient BoostingDecision Tree](https:\u002F\u002Fpapers.nips.cc\u002Fpaper\u002F6907-lightgbm-a-highly-efficient-gradient-boosting-decision-tree.pdf)\n\n:2nd_place_medal: :page_facing_up: [AdaBoost and the Super Bowl of Classifiers - A Tutorial Introduction to Adaptive Boosting](http:\u002F\u002Fwww.inf.fu-berlin.de\u002Finst\u002Fag-ki\u002Fadaboost4.pdf)\n\n:3rd_place_medal: :page_facing_up: [Greedy Function Approximation: A Gradient Boosting Machine](https:\u002F\u002Fprojecteuclid.org\u002Fdownload\u002Fpdf_1\u002Feuclid.aos\u002F1013203451)\n\n\n### :book: Unraveling Blackbox ML\n\n:3rd_place_medal: :page_facing_up: [Peeking Inside the Black Box: Visualizing Statistical Learning with Plots of Individual Conditional Expectation](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1309.6392.pdf)\n\n:3rd_place_medal: :page_facing_up: [Data Shapley: Equitable Valuation of Data for Machine Learning](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1904.02868.pdf)\n\n### :scissors: Dimensionality Reduction \n\n:1st_place_medal: :page_facing_up: [A Tutorial on Principal Component Analysis](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1404.1100.pdf)\n\n:2nd_place_medal: :page_facing_up: [How to Use t-SNE Effectively](https:\u002F\u002Fdistill.pub\u002F2016\u002Fmisread-tsne\u002F)\n\n:3rd_place_medal: :page_facing_up: [Visualizing Data using t-SNE](https:\u002F\u002Flvdmaaten.github.io\u002Fpublications\u002Fpapers\u002FJMLR_2008.pdf)\n\n\n### :chart_with_upwards_trend: Optimization\n\n:1st_place_medal: :page_facing_up: [A Tutorial on Bayesian Optimization](https:\u002F\u002Farxiv.org\u002Fabs\u002F1807.02811)\n\n:2nd_place_medal: :page_facing_up: [Taking the Human Out of the Loop: A review of Bayesian Optimization](https:\u002F\u002Fwww.cs.ox.ac.uk\u002Fpeople\u002Fnando.defreitas\u002Fpublications\u002FBayesOptLoop.pdf)\n\n---\n\n### Famous Blogs\n\n[Sebastian Raschka](https:\u002F\u002Fsebastianraschka.com\u002Fblog\u002Findex.html)\n[Chip Huyen](https:\u002F\u002Fhuyenchip.com\u002Fblog\u002F)\n\n---\n\n### :8ball: :crystal_ball: Recommenders\n\n#### Surveys\n\n:1st_place_medal: :page_facing_up: [A Survey of Collaborative Filtering Techniques](http:\u002F\u002Fdownloads.hindawi.com\u002Farchive\u002F2009\u002F421425.pdf)\n\n:1st_place_medal: :page_facing_up: [Collaborative Filtering Recommender Systems](http:\u002F\u002Fciteseerx.ist.psu.edu\u002Fviewdoc\u002Fdownload?doi=10.1.1.130.4520&rep=rep1&type=pdf)\n\n:1st_place_medal: :page_facing_up: [Deep Learning Based Recommender System: A Survey and New Perspectives](https:\u002F\u002Fsci-hub.tw\u002F10.1145\u002F3285029)\n\n:1st_place_medal: :page_facing_up: :thinking: :star: [Explainable Recommendation: A Survey and New Perspectives](https:\u002F\u002Farxiv.org\u002Fabs\u002F1804.11192) :star:\n\n#### Case Studies\n\n:2nd_place_medal: :page_facing_up: [The Netflix Recommender System: Algorithms, Business Value,and Innovation](http:\u002F\u002Fdelivery.acm.org\u002F10.1145\u002F2850000\u002F2843948\u002Fa13-gomez-uribe.pdf)\n\n- :globe_with_meridians: Netflix Medium Blog\n  - [Netflix Recommendations: Beyond the 5 stars Part 1](https:\u002F\u002Fmedium.com\u002Fnetflix-techblog\u002Fnetflix-recommendations-beyond-the-5-stars-part-2-d9b96aa399f5)\n  - [Netflix Recommendations: Beyond the 5 stars Part 2](https:\u002F\u002Fmedium.com\u002Fnetflix-techblog\u002Fnetflix-recommendations-beyond-the-5-stars-part-2-d9b96aa399f5)\n\n:2nd_place_medal: :page_facing_up: [Two Decades of Recommender Systems at Amazon.com](https:\u002F\u002Fpdfs.semanticscholar.org\u002F0f06\u002Fd328f6deb44e5e67408e0c16a8c7356330d1.pdf)\n\n:2nd_place_medal: :globe_with_meridians: [How Does Spotify Know You So Well?](https:\u002F\u002Fmedium.com\u002Fs\u002Fstory\u002Fspotifys-discover-weekly-how-machine-learning-finds-your-new-music-19a41ab76efe)\n\n:point_right: More In-Depth study, :closed_book: [Recommender Systems Handbook](https:\u002F\u002Fwww.amazon.com\u002FRecommender-Systems-Handbook-Francesco-Ricci\u002Fdp\u002F1489976361) \n\n---\n\n### Famous Deep Learning Blogs :cowboy_hat_face:\n\n:globe_with_meridians: [Stanford UFLDL Deep Learning Tutorial](http:\u002F\u002Fufldl.stanford.edu\u002Ftutorial\u002F)\n\n:globe_with_meridians: [Distill.pub](https:\u002F\u002Fdistill.pub\u002F)\n\n:globe_with_meridians: [Colah's Blog](http:\u002F\u002Fcolah.github.io\u002F)\n\n:globe_with_meridians: [Andrej Karpathy](https:\u002F\u002Fkarpathy.github.io\u002F)\n\n:globe_with_meridians: [Zack Lipton](http:\u002F\u002Fzacklipton.com\u002Farticles\u002F)\n\n:globe_with_meridians: [Sebastian Ruder](https:\u002F\u002Fruder.io\u002F)\n\n:globe_with_meridians: [Jay Alammar](http:\u002F\u002Fjalammar.github.io\u002F)\n\n---\n\n## :books: Neural Networks and Deep Learning Neural Networks\n\n:star: :1st_place_medal: :newspaper: [The Matrix Calculus You Need For Deep Learning - Terence Parr and Jeremy Howard](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1802.01528.pdf) :star:\n\n:1st_place_medal: :newspaper: [Deep learning -Yann LeCun, Yoshua Bengio & Geoffrey Hinton](https:\u002F\u002Fwww.cs.toronto.edu\u002F~hinton\u002Fabsps\u002FNatureDeepReview.pdf)\n\n:1st_place_medal: :page_facing_up: [Generalization in Deep Learning](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1710.05468.pdf)\n\n:1st_place_medal: :page_facing_up: [Topology of Learning in Artificial Neural Networks](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1902.08160v1.pdf)\n\n:1st_place_medal: :page_facing_up: [Dropout: A Simple Way to Prevent Neural Networks from Overfitting](https:\u002F\u002Fwww.cs.toronto.edu\u002F~hinton\u002Fabsps\u002FJMLRdropout.pdf)\n\n:2nd_place_medal: :page_facing_up: [Polynomial Regression As an Alternative to Neural Nets](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1806.06850)\n\n:2nd_place_medal: :globe_with_meridians: [The Neural Network Zoo](https:\u002F\u002Fwww.asimovinstitute.org\u002Fneural-network-zoo\u002F?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more)\n\n:2nd_place_medal: :globe_with_meridians: [Image Completion with Deep Learning in TensorFlow](http:\u002F\u002Fbamos.github.io\u002F2016\u002F08\u002F09\u002Fdeep-completion\u002F?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more)\n\n:2nd_place_medal: :page_facing_up: [Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1502.03167v3)\n\n:3rd_place_medal: :page_facing_up: [A systematic study of the class imbalance problem in convolutional neural networks](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1710.05381)\n\n:3rd_place_medal: :page_facing_up: [All Neural Networks are Created Equal](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1905.10854)\n\n:3rd_place_medal: :page_facing_up: [Adam: A Method for Stochastic Optimization](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1412.6980)\n\n:3rd_place_medal: :page_facing_up: [AutoML: A Survey of the State-of-the-Art](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1908.00709v1)\n\n### :framed_picture: CNNs\n\n:1st_place_medal: :page_facing_up: [Visualizing and Understanding Convolutional Networks -by Andrej Karpathy Justin Johnson Li Fei-Fei](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1311.2901.pdf)\n\n:2nd_place_medal: :page_facing_up: [Deep Residual Learning for Image Recognition](https:\u002F\u002Fwww.cv-foundation.org\u002Fopenaccess\u002Fcontent_cvpr_2016\u002Fpapers\u002FHe_Deep_Residual_Learning_CVPR_2016_paper.pdf)\n\n:2nd_place_medal: :page_facing_up:[AlexNet-ImageNet Classification with Deep Convolutional\nNeural Networks](https:\u002F\u002Fpapers.nips.cc\u002Fpaper\u002F4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf)\n\n:2nd_place_medal: :page_facing_up:[VGG Net-VERY DEEP CONVOLUTIONAL NETWORKS FOR LARGE-SCALE IMAGE RECOGNITION](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1409.1556v6.pdf)\n\n:3rd_place_medal: :page_facing_up: [A Mathematical Theory of Deep Convolutional Neural Networks for Feature Extraction](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1512.06293)\n\n:3rd_place_medal: :page_facing_up: [Large-scale Video Classification with Convolutional Neural Networks](https:\u002F\u002Fwww.cv-foundation.org\u002Fopenaccess\u002Fcontent_cvpr_2014\u002Fpapers\u002FKarpathy_Large-scale_Video_Classification_2014_CVPR_paper.pdf)\n\n:3rd_place_medal: :page_facing_up: [Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answering](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1707.07998.pdf)\n\n### :black_circle: CapsNet :trident:\t\n\n:1st_place_medal: :page_facing_up: [Dynamic Routing Between Capsules](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1710.09829.pdf)\n\n- Blog explaning, [\"What are CapsNet, or Capsule Networks?\"](https:\u002F\u002Fmedium.com\u002Fai%C2%B3-theory-practice-business\u002Funderstanding-hintons-capsule-networks-part-i-intuition-b4b559d1159b)\n\n- [Capsule Networks Tutorial by Aureline Geron](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=pPN8d0E3900&t=1199s)\n\n###  :national_park: :speech_balloon: Image Captioning\n\n:1st_place_medal: :page_facing_up: [Show and Tell: A Neural Image Caption Generator](https:\u002F\u002Farxiv.org\u002Fabs\u002F1411.4555)\n\n:2nd_place_medal: :page_facing_up: [Neural Machine Translation by Jointly Learning to Align and Translate](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1409.0473v7)\n\n:2nd_place_medal: :page_facing_up: [StyleNet: Generating Attractive Visual Captions with Styles](https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fresearch\u002Fuploads\u002Fprod\u002F2017\u002F06\u002FGenerating-Attractive-Visual-Captions-with-Styles)\n\n:2nd_place_medal: :page_facing_up: [Show, Attend and Tell: Neural Image Caption Generation with Visual Attention](https:\u002F\u002Farxiv.org\u002Fabs\u002F1502.03044)\n\n:2nd_place_medal: :page_facing_up: [Where to put the Image in an Image Caption Generator](https:\u002F\u002Farxiv.org\u002Fabs\u002F1703.09137)\n\n:2nd_place_medal: :page_facing_up: [Dank Learning: Generating Memes Using Deep Neural Networks](https:\u002F\u002Farxiv.org\u002Fabs\u002F1806.04510)\n\n### :car: :walking_man: Object Detection :eagle: :football:\n\n:2nd_place_medal: :page_facing_up:[ResNet-Deep Residual Learning for Image Recognition](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1512.03385)\n\n:2nd_place_medal: :page_facing_up: [YOLO-You Only Look Once: Unified, Real-Time Object Detection](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1506.02640)\n\n:2nd_place_medal: :page_facing_up: [Microsoft COCO: Common Objects in Context](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1405.0312)\n\n- [COCO dataset](http:\u002F\u002Fcocodataset.org\u002F#home)\n\n:2nd_place_medal: :page_facing_up:  [(R-CNN) Rich feature hierarchies for accurate object detection and semantic segmentation](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1311.2524.pdf)\n\n:2nd_place_medal: :page_facing_up: [Fast R-CNN](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1504.08083.pdf)\n\n- :computer: [Papers with Code](https:\u002F\u002Fwww.paperswithcode.com\u002Fpaper\u002Ffast-r-cnn)\n\n:2nd_place_medal: :page_facing_up: [Faster R-CNN](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1506.01497v3.pdf)\n\n- :computer: [Papers with Code](https:\u002F\u002Fwww.paperswithcode.com\u002Fpaper\u002Fmask-r-cnn)\n\n:2nd_place_medal: :page_facing_up: [Mask R-CNN](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1703.06870.pdf)\n\n- :computer: [Papers with Code](https:\u002F\u002Fwww.paperswithcode.com\u002Fpaper\u002Fmask-r-cnn)\n\n### :car: :walking_man: :couple: Pose Detection :runner: :dancer:\n\n:2nd_place_medal: :page_facing_up: [DensePose: Dense Human Pose Estimation In The Wild](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1802.00434v1.pdf)\n\n- :computer: [Papers with Code](https:\u002F\u002Fwww.paperswithcode.com\u002Fpaper\u002Fdensepose-dense-human-pose-estimation-in-the)\n\n:2nd_place_medal: :page_facing_up:  [Parsing R-CNN for Instance-Level Human Analysis](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1811.12596v1.pdf)\n\n- :computer: [Papers with Code](https:\u002F\u002Fwww.paperswithcode.com\u002Fpaper\u002Fparsing-r-cnn-for-instance-level-human)\n\n### :abcd: :symbols: Deep NLP :currency_exchange: :1234:\n\n:1st_place_medal: :page_facing_up: [A Primer on Neural Network Models for Natural Language Processing](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1510.00726.pdf)\n\n:1st_place_medal: :page_facing_up: [Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1412.3555.pdf)\n\n:1st_place_medal: :page_facing_up: [On the Properties of Neural Machine Translation: Encoder–Decoder Approaches](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1409.1259.pdf)\n\n:1st_place_medal: :page_facing_up: [LSTM: A Search Space Odyssey - by Klaus Greff et al.](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1503.04069.pdf)\n\n:1st_place_medal: :page_facing_up: [A Critical Review of Recurrent Neural Networksfor Sequence Learning](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1506.00019.pdf)\n\n:1st_place_medal: :page_facing_up: [Visualizing and Understanding Recurrent Networks](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1506.02078.pdf)\n\n:star: :1st_place_medal: :page_facing_up: [Attention Is All You Need](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1706.03762) :star:\n\n:1st_place_medal: :page_facing_up: [An Empirical Exploration of Recurrent Network Architectures](http:\u002F\u002Fproceedings.mlr.press\u002Fv37\u002Fjozefowicz15.pdf)\n\n:1st_place_medal: :page_facing_up: [Open AI (GPT-2) Language Models are Unsupervised Multitask Learners](https:\u002F\u002Fpaperswithcode.com\u002Fpaper\u002Flanguage-models-are-unsupervised-multitask)\n\n:1st_place_medal: :page_facing_up: [BERT: Pre-training of Deep Bidirectional Transformers forLanguage Understanding](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1810.04805)\n\n- [Google BERT Annoucement](https:\u002F\u002Fai.googleblog.com\u002F2018\u002F11\u002Fopen-sourcing-bert-state-of-art-pre.html)\n\n:3rd_place_medal: :page_facing_up: [Parameter-Efficient Transfer Learning for NLP](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1902.00751)\n\n:3rd_place_medal: :page_facing_up: [A Sensitivity Analysis of (and Practitioners’ Guide to) ConvolutionalNeural Networks for Sentence Classification](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1510.03820v4.pdf)\n\n:3rd_place_medal: :page_facing_up: [A Survey on Recent Advances in Named Entity Recognition from Deep Learning models](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1910.11470v1)\n\n:3rd_place_medal: :page_facing_up: [Convolutional Neural Networks for Sentence Classification](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1408.5882v2)\n\n:3rd_place_medal: :page_facing_up: [Pervasive Attention: 2D Convolutional Neural Networks for Sequence-to-Sequence Prediction](https:\u002F\u002Farxiv.org\u002Fabs\u002F1808.03867)\n\n:3rd_place_medal: :page_facing_up: [Single Headed Attention RNN: Stop Thinking With Your Head](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1911.11423.pdf)\n\n### :alien: GANs\n\n:1st_place_medal: :page_facing_up: [Generative Adversarial Nets - Goodfellow et al.](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1406.2661v1.pdf)\n\n:books: GAN Rabbit Hole -> [GAN Papers](https:\u002F\u002Fgithub.com\u002Fzhangqianhui\u002FAdversarialNetsPapers)\n\n### :o::heavy_minus_sign::o: GNNs (Graph Neural Networks)\n\n:3rd_place_medal: :page_facing_up: [A Comprehensive Survey on Graph Neural Networks](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1901.00596.pdf)\n\n---\n\n### :man_health_worker: :syringe: Medical AI :pill: :microscope:\n\n[Machine learning classifiers and fMRI: a tutorial overview - by Francisco et al.](https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fpmc\u002Farticles\u002FPMC2892746\u002Fpdf\u002Fnihms100405.pdf)\n\n---\n\n## :point_down: Cool Stuff :point_down:\n\n:loud_sound: :page_facing_up: [SoundNet: Learning Sound\nRepresentations from Unlabeled Video](http:\u002F\u002Fsoundnet.csail.mit.edu\u002F)\n\n:art: :page_facing_up: [CAN: Creative Adversarial NetworksGenerating “Art” by Learning About Styles andDeviating from Style Norms](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1706.07068)\n\n:art: :page_facing_up: [Deep Painterly Harmonization](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1804.03189)\n\n- [Github Code](https:\u002F\u002Fgithub.com\u002Fluanfujun\u002Fdeep-painterly-harmonization)\n\n:man_dancing: :dancer: :page_facing_up: [Everybody Dance Now](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1808.07371)\n\n- [Everybody Dance Now - Youtube Video](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=PCBTZh41Ris)\n\n:soccer: [Soccer on Your Tabletop](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1806.00890)\n\n:blonde_woman: :haircut_woman: :page_facing_up: [SC-FEGAN: Face Editing Generative Adversarial Network with User's Sketch and Color](https:\u002F\u002Farxiv.org\u002Fabs\u002F1902.06838)\n\n- [Github Code](https:\u002F\u002Fgithub.com\u002Frun-youngjoo\u002FSC-FEGAN)\n\n:camera_flash: :page_facing_up: [Handheld Mobile Photography in Very Low Light](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1910.11336v1)\n\n:japanese_castle: :mosque: :page_facing_up: [Learning Deep Features for Scene Recognitionusing Places Database](http:\u002F\u002Fplaces.csail.mit.edu\u002Fplaces_NIPS14.pdf)\n\n:bullettrain_front: :bullettrain_side: :page_facing_up: [High-Speed Tracking withKernelized Correlation Filters](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1404.7584)\n\n:clapper: :page_facing_up: [Recent progress in semantic image segmentation](https:\u002F\u002Farxiv.org\u002Fftp\u002Farxiv\u002Fpapers\u002F1809\u002F1809.10198)\n\nRabbit hole -> :loud_sound: :globe_with_meridians: [Analytics Vidhya Top 10 Audio Processing Tasks and their papers](https:\u002F\u002Fwww.analyticsvidhya.com\u002Fblog\u002F2018\u002F01\u002F10-audio-processing-projects-applications\u002F)\n\n:blonde_man: -> :older_man: :page_facing_up: :page_facing_up: [Face Aging With Condintional GANS](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1702.01983)\n\n:blonde_man: -> :older_man: :page_facing_up: :page_facing_up: [Dual Conditional GANs for Face Aging and Rejuvenation](https:\u002F\u002Fwww.ijcai.org\u002Fproceedings\u002F2018\u002F0125.pdf)\n\n:balance_scale: :page_facing_up: [BAGAN: Data Augmentation with Balancing GAN](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1803.09655)\n\n[labml.ai Annotated PyTorch Paper Implementations](https:\u002F\u002Fnn.labml.ai\u002F)\n\n---\n\n## :newspaper: Cap Stone Projects :newspaper:\n\n[8 Awesome Data Science Capstone Projects](https:\u002F\u002Fwww.analyticsvidhya.com\u002Fblog\u002F2019\u002F04\u002F8-awesome-data-science-capstone-projects-from-praxis-business-school\u002F)\n\n[10 Powerful Applications of Linear Algebra in Data Science](https:\u002F\u002Fwww.analyticsvidhya.com\u002Fblog\u002F2019\u002F07\u002F10-applications-linear-algebra-data-science\u002F)\n\n[Top 5 Interesting Applications of GANs](https:\u002F\u002Fwww.analyticsvidhya.com\u002Fblog\u002F2019\u002F04\u002Ftop-5-interesting-applications-gans-deep-learning\u002F)\n\n[Deep Learning Applications a beginner can build in minutes ](https:\u002F\u002Fwww.analyticsvidhya.com\u002Fblog\u002F2017\u002F02\u002F5-deep-learning-applications-beginner-python\u002F)\n\n---\n\n#### CHANGELOG\n\n2019-10-28 Started `must-read-papers-for-ml` repo\n\n2019-10-29 Added analytics vidhya use case studies article links\n\n2019-10-30 Added Outlier\u002FAnomaly detection paper, separated Boosting, CNN, Object Detection, NLP papers, and added Image captioning papers \n\n2019-10-31 Added Famous Blogs from Deep and Machine Learning Researchers\n\n2019-11-1 Fixed markdown issues, added contribution guideline\n\n2019-11-20 Added Recommender Surveys, and Papers\n\n2019-12-12 Added R-CNN variants, PoseNets, GNNs\n\n2020-02-23 Added GRU paper\n","# 数据科学、机器学习和深度学习必读论文\n### 精选的数据科学、机器学习和深度学习论文、综述及文章合集，均为必读清单中的内容。\n\n---\n\n> 注意：:construction: 正在更新中，请告诉我您希望添加哪些额外的论文、文章或博客，我会在此处补充。\n\n### 使用方法  \n> :point_right: :star: 这个仓库\n\n\n\n## 贡献\n- :point_right: :arrows_clockwise: 如果链接失效，或者我遗漏了任何重要的论文、博客或文章，请随时 [提交 Pull Request](https:\u002F\u002Fgithub.com\u002Fhurshd0\u002Fmust-read-papers-for-ml\u002Fpulls)。\n\n[![维护](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FMaintained%3F-yes-green.svg)](https:\u002F\u002Fgithub.com\u002Fhurshd0\u002Fmust-read-papers-for-ml\u002Fgraphs\u002Fcommit-activity)\n\n### :point_down: 请阅读此内容 :point_down:\n\n- :point_right: 阅读包含大量数学公式的论文很困难，需要时间和精力去理解。关键在于坚持和动力，不要轻易放弃。不妨多读几遍，直到真正领会并为之震撼。\n\n\n:1st_place_medal: - 第一次阅读\n\n:2nd_place_medal: - 第二次阅读 \n\n:3rd_place_medal: - 第三次阅读\n\n---\n\n## 数据科学\n\n### :bar_chart: 数据预处理与探索性数据分析\n\n:1st_place_medal: :page_facing_up:[数据预处理——整洁数据——哈德利·威克姆著](https:\u002F\u002Fvita.had.co.nz\u002Fpapers\u002Ftidy-data.pdf)\n\n### :notebook: 通用数据科学\n\n:1st_place_medal: :page_facing_up: [统计建模：两种文化——莱奥·布雷曼著](https:\u002F\u002Fprojecteuclid.org\u002Fdownload\u002Fpdf_1\u002Feuclid.ss\u002F1009213726)\n\n:2nd_place_medal: :page_facing_up: [拉什蒙曲线与体积研究：机器学习中泛化与模型简洁性的新视角](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1908.01755.pdf)\n\n- :video_camera: [KDD 2019 辛西娅·鲁丁主题演讲](https:\u002F\u002Fyoutu.be\u002FwL4X4lG20sM)\n\n:1st_place_medal: :page_facing_up: [频率主义与贝叶斯主义：杰克·范德普拉斯的Python入门教程](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1411.5018.pdf)\n\n---\n\n## 机器学习\n\n### :dart: 通用机器学习\n\n:1st_place_medal: :page_facing_up: [机器学习中的模型评估、模型选择与算法选择——塞巴斯蒂安·拉斯奇卡著](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1811.12808.pdf)\n\n:1st_place_medal: :page_facing_up: [机器学习简要介绍——冈纳尔·拉茨奇著](https:\u002F\u002Fevents.ccc.de\u002Fcongress\u002F2004\u002Ffahrplan\u002Ffiles\u002F105-machine-learning-paper.pdf)\n\n:3rd_place_medal: :page_facing_up: [共轭梯度法简介：无需痛苦的讲解——乔纳森·理查德·谢伍克著](http:\u002F\u002Fwww.cs.cmu.edu\u002F~quake-papers\u002Fpainless-conjugate-gradient.pdf)\n\n:3rd_place_medal: :page_facing_up: [随机种子对模型稳定性的影响研究](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1909.10447)\n\n### :mag: 异常值\u002F异常检测\n\n:1st_place_medal: :newspaper: [异常检测：综述](https:\u002F\u002Fpdfs.semanticscholar.org\u002F912b\u002F0b7879ca99bf654a26bbb0d50d4b8e0ed6c0.pdf)\n\n### :rocket: 提升算法\n\n:2nd_place_medal: :page_facing_up: [XGBoost：可扩展的树提升系统](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1603.02754.pdf)\n\n:2nd_place_medal: :page_facing_up: [LightGBM：高效的梯度提升决策树](https:\u002F\u002Fpapers.nips.cc\u002Fpaper\u002F6907-lightgbm-a-highly-efficient-gradient-boosting-decision-tree.pdf)\n\n:2nd_place_medal: :page_facing_up: [AdaBoost与分类器界的超级碗——自适应提升算法教程](http:\u002F\u002Fwww.inf.fu-berlin.de\u002Finst\u002Fag-ki\u002Fadaboost4.pdf)\n\n:3rd_place_medal: :page_facing_up: [贪婪函数逼近：梯度提升机](https:\u002F\u002Fprojecteuclid.org\u002Fdownload\u002Fpdf_1\u002Feuclid.aos\u002F1013203451)\n\n\n### :book: 解密黑盒机器学习\n\n:3rd_place_medal: :page_facing_up: [窥探黑盒：用个体条件期望图可视化统计学习](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1309.6392.pdf)\n\n:3rd_place_medal: :page_facing_up: [数据夏普利值：机器学习中数据的公平估值](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1904.02868.pdf)\n\n### :scissors: 降维 \n\n:1st_place_medal: :page_facing_up: [主成分分析教程](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1404.1100.pdf)\n\n:2nd_place_medal: :page_facing_up: [如何有效使用t-SNE](https:\u002F\u002Fdistill.pub\u002F2016\u002Fmisread-tsne\u002F)\n\n:3rd_place_medal: :page_facing_up: [使用t-SNE进行数据可视化](https:\u002F\u002Flvdmaaten.github.io\u002Fpublications\u002Fpapers\u002FJMLR_2008.pdf)\n\n\n### :chart_with_upwards_trend: 优化\n\n:1st_place_medal: :page_facing_up: [贝叶斯优化教程](https:\u002F\u002Farxiv.org\u002Fabs\u002F1807.02811)\n\n:2nd_place_medal: :page_facing_up: [摆脱人为干预：贝叶斯优化综述](https:\u002F\u002Fwww.cs.ox.ac.uk\u002Fpeople\u002Fnando.defreitas\u002Fpublications\u002FBayesOptLoop.pdf)\n\n---\n\n### 著名博客\n\n[塞巴斯蒂安·拉斯奇卡](https:\u002F\u002Fsebastianraschka.com\u002Fblog\u002Findex.html)\n[奇普·休恩](https:\u002F\u002Fhuyenchip.com\u002Fblog\u002F)\n\n---\n\n### :8ball: :crystal_ball: 推荐系统\n\n#### 综述\n\n:1st_place_medal: :page_facing_up: [协同过滤技术综述](http:\u002F\u002Fdownloads.hindawi.com\u002Farchive\u002F2009\u002F421425.pdf)\n\n:1st_place_medal: :page_facing_up: [协同过滤推荐系统](http:\u002F\u002Fciteseerx.ist.psu.edu\u002Fviewdoc\u002Fdownload?doi=10.1.1.130.4520&rep=rep1&type=pdf)\n\n:1st_place_medal: :page_facing_up: [基于深度学习的推荐系统：综述与新视角](https:\u002F\u002Fsci-hub.tw\u002F10.1145\u002F3285029)\n\n:1st_place_medal: :page_facing_up: :thinking: :star: [可解释的推荐：综述与新视角](https:\u002F\u002Farxiv.org\u002Fabs\u002F1804.11192) :star:\n\n#### 案例研究\n\n:2nd_place_medal: :page_facing_up: [Netflix推荐系统：算法、商业价值与创新](http:\u002F\u002Fdelivery.acm.org\u002F10.1145\u002F2850000\u002F2843948\u002Fa13-gomez-uribe.pdf)\n\n- :globe_with_meridians: Netflix Medium博客\n  - [Netflix推荐：超越五星评价 第1部分](https:\u002F\u002Fmedium.com\u002Fnetflix-techblog\u002Fnetflix-recommendations-beyond-the-5-stars-part-2-d9b96aa399f5)\n  - [Netflix推荐：超越五星评价 第2部分](https:\u002F\u002Fmedium.com\u002Fnetflix-techblog\u002Fnetflix-recommendations-beyond-the-5-stars-part-2-d9b96aa399f5)\n\n:2nd_place_medal: :page_facing_up: [亚马逊推荐系统二十年回顾](https:\u002F\u002Fpdfs.semanticscholar.org\u002F0f06\u002Fd328f6deb44e5e67408e0c16a8c7356330d1.pdf)\n\n:2nd_place_medal: :globe_with_meridians: [Spotify为何如此了解你？](https:\u002F\u002Fmedium.com\u002Fs\u002Fstory\u002Fspotifys-discover-weekly-how-machine-learning-finds-your-new-music-19a41ab76efe)\n\n:point_right: 更深入的研究， :closed_book: [推荐系统手册](https:\u002F\u002Fwww.amazon.com\u002FRecommender-Systems-Handbook-Francesco-Ricci\u002Fdp\u002F1489976361) \n\n---\n\n### 著名的深度学习博客 :cowboy_hat_face:\n\n:globe_with_meridians: [斯坦福UFLDL深度学习教程](http:\u002F\u002Fufldl.stanford.edu\u002Ftutorial\u002F)\n\n:globe_with_meridians: [Distill.pub](https:\u002F\u002Fdistill.pub\u002F)\n\n:globe_with_meridians: [Colah的博客](http:\u002F\u002Fcolah.github.io\u002F)\n\n:globe_with_meridians: [Andrej Karpathy](https:\u002F\u002Fkarpathy.github.io\u002F)\n\n:globe_with_meridians: [Zack Lipton](http:\u002F\u002Fzacklipton.com\u002Farticles\u002F)\n\n:globe_with_meridians: [Sebastian Ruder](https:\u002F\u002Fruder.io\u002F)\n\n:globe_with_meridians: [Jay Alammar](http:\u002F\u002Fjalammar.github.io\u002F)\n\n---\n\n## :books: 神经网络与深度学习 神经网络\n\n:star: :1st_place_medal: :newspaper: [深度学习所需的矩阵微积分 - Terence Parr 和 Jeremy Howard](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1802.01528.pdf) :star:\n\n:1st_place_medal: :newspaper: [深度学习 - Yann LeCun、Yoshua Bengio 和 Geoffrey Hinton](https:\u002F\u002Fwww.cs.toronto.edu\u002F~hinton\u002Fabsps\u002FNatureDeepReview.pdf)\n\n:1st_place_medal: :page_facing_up: [深度学习中的泛化](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1710.05468.pdf)\n\n:1st_place_medal: :page_facing_up: [人工神经网络中学习的拓扑结构](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1902.08160v1.pdf)\n\n:1st_place_medal: :page_facing_up: [Dropout：防止神经网络过拟合的简单方法](https:\u002F\u002Fwww.cs.toronto.edu\u002F~hinton\u002Fabsps\u002FJMLRdropout.pdf)\n\n:2nd_place_medal: :page_facing_up: [多项式回归作为神经网络的替代方案](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1806.06850)\n\n:2nd_place_medal: :globe_with_meridians: [神经网络动物园](https:\u002F\u002Fwww.asimovinstitute.org\u002Fneural-network-zoo\u002F?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more)\n\n:2nd_place_medal: :globe_with_meridians: [使用TensorFlow进行深度学习的图像补全](http:\u002F\u002Fbamos.github.io\u002F2016\u002F08\u002F09\u002Fdeep-completion\u002F?utm_source=mybridge&utm_medium=blog&utm_campaign=read_more)\n\n:2nd_place_medal: :page_facing_up: [批归一化：通过减少内部协变量偏移加速深度网络训练](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1502.03167v3)\n\n:3rd_place_medal: :page_facing_up: [卷积神经网络中类别不平衡问题的系统性研究](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1710.05381)\n\n:3rd_place_medal: :page_facing_up: [所有神经网络都是平等的](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1905.10854)\n\n:3rd_place_medal: :page_facing_up: [Adam：一种随机优化方法](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1412.6980)\n\n:3rd_place_medal: :page_facing_up: [AutoML：现状综述](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1908.00709v1)\n\n### :framed_picture: CNNs\n\n:1st_place_medal: :page_facing_up: [可视化与理解卷积网络 - Andrej Karpathy、Justin Johnson 和 Li Fei-Fei](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1311.2901.pdf)\n\n:2nd_place_medal: :page_facing_up: [用于图像识别的深度残差学习](https:\u002F\u002Fwww.cv-foundation.org\u002Fopenaccess\u002Fcontent_cvpr_2016\u002Fpapers\u002FHe_Deep_Residual_Learning_CVPR_2016_paper.pdf)\n\n:2nd_place_medal: :page_facing_up:[AlexNet-ImageNet分类与深度卷积神经网络](https:\u002F\u002Fpapers.nips.cc\u002Fpaper\u002F4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf)\n\n:2nd_place_medal: :page_facing_up:[VGG Net-用于大规模图像识别的超深卷积网络](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1409.1556v6.pdf)\n\n:3rd_place_medal: :page_facing_up: [用于特征提取的深度卷积神经网络的数学理论](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1512.06293)\n\n:3rd_place_medal: :page_facing_up: [使用卷积神经网络进行大规模视频分类](https:\u002F\u002Fwww.cv-foundation.org\u002Fopenaccess\u002Fcontent_cvpr_2014\u002Fpapers\u002FKarpathy_Large-scale_Video_Classification_2014_CVPR_paper.pdf)\n\n:3rd_place_medal: :page_facing_up: [自下而上和自上而下的注意力机制在图像字幕生成和视觉问答中的应用](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1707.07998.pdf)\n\n### :black_circle: CapsNet :trident:\t\n\n:1st_place_medal: :page_facing_up: [胶囊之间的动态路由](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1710.09829.pdf)\n\n- 解释性博客，[\"什么是CapsNet，或胶囊网络？\"](https:\u002F\u002Fmedium.com\u002Fai%C2%B3-theory-practice-business\u002Funderstanding-hintons-capsule-networks-part-i-intuition-b4b559d1159b)\n\n- [Aureline Geron的胶囊网络教程](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=pPN8d0E3900&t=1199s)\n\n###  :national_park: :speech_balloon: 图像字幕生成\n\n:1st_place_medal: :page_facing_up: [展示与讲述：神经网络图像字幕生成器](https:\u002F\u002Farxiv.org\u002Fabs\u002F1411.4555)\n\n:2nd_place_medal: :page_facing_up: [通过联合学习对齐与翻译实现神经机器翻译](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1409.0473v7)\n\n:2nd_place_medal: :page_facing_up: [StyleNet：以风格生成吸引人的视觉字幕](https:\u002F\u002Fwww.microsoft.com\u002Fen-us\u002Fresearch\u002Fuploads\u002Fprod\u002F2017\u002F06\u002FGenerating-Attractive-Visual-Captions-with-Styles)\n\n:2nd_place_medal: :page_facing_up: [展示、关注与讲述：结合视觉注意力的神经网络图像字幕生成](https:\u002F\u002Farxiv.org\u002Fabs\u002F1502.03044)\n\n:2nd_place_medal: :page_facing_up: [图像字幕生成器中图像应放置的位置](https:\u002F\u002Farxiv.org\u002Fabs\u002F1703.09137)\n\n:2nd_place_medal: :page_facing_up: [Dank Learning：利用深度神经网络生成表情包](https:\u002F\u002Farxiv.org\u002Fabs\u002F1806.04510)\n\n### :car: :walking_man: 物体检测 :eagle: :football:\n\n:2nd_place_medal: :page_facing_up:[ResNet-用于图像识别的深度残差学习](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1512.03385)\n\n:2nd_place_medal: :page_facing_up: [YOLO-你只需看一次：统一的实时物体检测](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1506.02640)\n\n:2nd_place_medal: :page_facing_up: [Microsoft COCO：上下文中的常见物体](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1405.0312)\n\n- [COCO数据集](http:\u002F\u002Fcocodataset.org\u002F#home)\n\n:2nd_place_medal: :page_facing_up: [(R-CNN) 用于精确物体检测和语义分割的丰富特征层次结构](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1311.2524.pdf)\n\n:2nd_place_medal: :page_facing_up: [Fast R-CNN](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1504.08083.pdf)\n\n- :computer: [Papers with Code](https:\u002F\u002Fwww.paperswithcode.com\u002Fpaper\u002Ffast-r-cnn)\n\n:2nd_place_medal: :page_facing_up: [Faster R-CNN](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1506.01497v3.pdf)\n\n- :computer: [Papers with Code](https:\u002F\u002Fwww.paperswithcode.com\u002Fpaper\u002Fmask-r-cnn)\n\n:2nd_place_medal: :page_facing_up: [Mask R-CNN](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1703.06870.pdf)\n\n- :computer: [Papers with Code](https:\u002F\u002Fwww.paperswithcode.com\u002Fpaper\u002Fmask-r-cnn)\n\n### :car: :walking_man: :couple: 姿态检测 :runner: :dancer:\n\n:2nd_place_medal: :page_facing_up: [DensePose：野外密集人体姿态估计](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1802.00434v1.pdf)\n\n- :computer: [Papers with Code](https:\u002F\u002Fwww.paperswithcode.com\u002Fpaper\u002Fdensepose-dense-human-pose-estimation-in-the)\n\n:2nd_place_medal: :page_facing_up:  [用于实例级人体分析的Parsing R-CNN](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1811.12596v1.pdf)\n\n- :computer: [Papers with Code](https:\u002F\u002Fwww.paperswithcode.com\u002Fpaper\u002Fparsing-r-cnn-for-instance-level-human)\n\n### :abcd: :symbols: 深度NLP :currency_exchange: :1234:\n\n:1st_place_medal: :page_facing_up: [自然语言处理中神经网络模型入门](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1510.00726.pdf)\n\n:1st_place_medal: :page_facing_up: [门控循环神经网络在序列建模中的经验评估](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1412.3555.pdf)\n\n:1st_place_medal: :page_facing_up: [关于神经机器翻译的特性：编码器-解码器方法](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1409.1259.pdf)\n\n:1st_place_medal: :page_facing_up: [LSTM：搜索空间之旅——克劳斯·格雷夫等著](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1503.04069.pdf)\n\n:1st_place_medal: :page_facing_up: [递归神经网络用于序列学习的批判性评论](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1506.00019.pdf)\n\n:1st_place_medal: :page_facing_up: [递归网络的可视化与理解](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1506.02078.pdf)\n\n:star: :1st_place_medal: :page_facing_up: [注意力就是一切](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1706.03762) :star:\n\n:1st_place_medal: :page_facing_up: [递归网络架构的实证探索](http:\u002F\u002Fproceedings.mlr.press\u002Fv37\u002Fjozefowicz15.pdf)\n\n:1st_place_medal: :page_facing_up: [OpenAI（GPT-2）语言模型是无监督的多任务学习者](https:\u002F\u002Fpaperswithcode.com\u002Fpaper\u002Flanguage-models-are-unsupervised-multitask)\n\n:1st_place_medal: :page_facing_up: [BERT：面向语言理解的深度双向Transformer预训练](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1810.04805)\n\n- [Google BERT公告](https:\u002F\u002Fai.googleblog.com\u002F2018\u002F11\u002Fopen-sourcing-bert-state-of-art-pre.html)\n\n:3rd_place_medal: :page_facing_up: [NLP中的参数高效迁移学习](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1902.00751)\n\n:3rd_place_medal: :page_facing_up: [卷积神经网络用于句子分类的敏感性分析及实践指南](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1510.03820v4.pdf)\n\n:3rd_place_medal: :page_facing_up: [深度学习模型在命名实体识别领域的最新进展综述](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1910.11470v1)\n\n:3rd_place_medal: :page_facing_up: [卷积神经网络用于句子分类](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1408.5882v2)\n\n:3rd_place_medal: :page_facing_up: [普适注意力：用于序列到序列预测的二维卷积神经网络](https:\u002F\u002Farxiv.org\u002Fabs\u002F1808.03867)\n\n:3rd_place_medal: :page_facing_up: [单头注意力RNN：停止用大脑思考](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1911.11423.pdf)\n\n### :alien: GANs\n\n:1st_place_medal: :page_facing_up: [生成对抗网络——古德费洛等著](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1406.2661v1.pdf)\n\n:books: GAN深坑 -> [GAN论文](https:\u002F\u002Fgithub.com\u002Fzhangqianhui\u002FAdversarialNetsPapers)\n\n### :o::heavy_minus_sign::o: GNNs（图神经网络）\n\n:3rd_place_medal: :page_facing_up: [图神经网络综合综述](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1901.00596.pdf)\n\n---\n\n### :man_health_worker: :syringe: 医疗AI :pill: :microscope:\n\n[机器学习分类器与fMRI：弗朗西斯科等人的教程概述](https:\u002F\u002Fwww.ncbi.nlm.nih.gov\u002Fpmc\u002Farticles\u002FPMC2892746\u002Fpdf\u002Fnihms100405.pdf)\n\n---\n\n## :point_down: 有趣的东西 :point_down:\n\n:loud_sound: :page_facing_up: [SoundNet：从无标签视频中学习声音表示](http:\u002F\u002Fsoundnet.csail.mit.edu\u002F)\n\n:art: :page_facing_up: [CAN：通过学习风格并偏离风格规范来生成“艺术”的创造性对抗网络](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1706.07068)\n\n:art: :page_facing_up: [深度绘画风格化](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1804.03189)\n\n- [Github代码](https:\u002F\u002Fgithub.com\u002Fluanfujun\u002Fdeep-painterly-harmonization)\n\n:man_dancing: :dancer: :page_facing_up: [大家一起跳舞吧](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1808.07371)\n\n- [大家一起跳舞吧——YouTube视频](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=PCBTZh41Ris)\n\n:soccer: [桌面上的足球](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1806.00890)\n\n:blonde_woman: :haircut_woman: :page_facing_up: [SC-FEGAN：基于用户草图和颜色的脸部编辑生成对抗网络](https:\u002F\u002Farxiv.org\u002Fabs\u002F1902.06838)\n\n- [Github代码](https:\u002F\u002Fgithub.com\u002Frun-youngjoo\u002FSC-FEGAN)\n\n:camera_flash: :page_facing_up: [极低光照下的手持移动摄影](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1910.11336v1)\n\n:japanese_castle: :mosque: :page_facing_up: [使用Places数据库学习场景识别的深度特征](http:\u002F\u002Fplaces.csail.mit.edu\u002Fplaces_NIPS14.pdf)\n\n:bullettrain_front: :bullettrain_side: :page_facing_up: [基于核相关滤波器的高速目标跟踪](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1404.7584)\n\n:clapper: :page_facing_up: [语义图像分割的最新进展](https:\u002F\u002Farxiv.org\u002Fftp\u002Farxiv\u002Fpapers\u002F1809\u002F1809.10198)\n\n深坑 -> :loud_sound: :globe_with_meridians: [Analytics Vidhya十大音频处理任务及其论文](https:\u002F\u002Fwww.analyticsvidhya.com\u002Fblog\u002F2018\u002F01\u002F10-audio-processing-projects-applications\u002F)\n\n:blonde_man: -> :older_man: :page_facing_up: :page_facing_up: [使用条件GAN进行人脸老化](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1702.01983)\n\n:blonde_man: -> :older_man: :page_facing_up: :page_facing_up: [用于人脸老化与年轻化的双重条件GAN](https:\u002F\u002Fwww.ijcai.org\u002Fproceedings\u002F2018\u002F0125.pdf)\n\n:balance_scale: :page_facing_up: [BAGAN：基于平衡GAN的数据增强](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1803.09655)\n\n[labml.ai注释版PyTorch论文实现](https:\u002F\u002Fnn.labml.ai\u002F)\n\n---\n\n## :newspaper: 综合项目 :newspaper:\n\n[8个超赞的数据科学综合项目](https:\u002F\u002Fwww.analyticsvidhya.com\u002Fblog\u002F2019\u002F04\u002F8-awesome-data-science-capstone-projects-from-praxis-business-school\u002F)\n\n[线性代数在数据科学中的10大强大应用](https:\u002F\u002Fwww.analyticsvidhya.com\u002Fblog\u002F2019\u002F07\u002F10-applications-linear-algebra-data-science\u002F)\n\n[GANs的5大有趣应用](https:\u002F\u002Fwww.analyticsvidhya.com\u002Fblog\u002F2019\u002F04\u002Ftop-5-interesting-applications-gans-deep-learning\u002F)\n\n[初学者可在几分钟内构建的深度学习应用](https:\u002F\u002Fwww.analyticsvidhya.com\u002Fblog\u002F2017\u002F02\u002F5-deep-learning-applications-beginner-python\u002F)\n\n---\n\n#### 更改记录\n\n2019年10月28日 启动`must-read-papers-for-ml`仓库\n\n2019年10月29日 添加Analytics Vidhya用例研究文章链接\n\n2019年10月30日 添加异常检测论文，将提升、CNN、目标检测、NLP论文分开，并添加图像字幕生成论文\n\n2019年10月31日 添加深度学习和机器学习研究者的知名博客\n\n2019年11月1日 修复Markdown问题，添加贡献指南\n\n2019年11月20日 添加推荐系统调查及论文\n\n2019年12月12日 添加R-CNN变体、PoseNets、GNNs\n\n2020年2月23日 添加GRU论文","# must-read-papers-for-ml 快速上手指南\n\n`must-read-papers-for-ml` 并非一个需要编译或安装依赖的软件库，而是一个**精选的数据科学、机器学习和深度学习论文、综述及文章清单**。它本质上是一个静态资源仓库，旨在为开发者提供高质量的学习路径。\n\n以下是获取和使用该资源的最简指南。\n\n## 环境准备\n\n本项目无需特定的操作系统或复杂的软件依赖，只需具备以下条件即可：\n\n*   **操作系统**：Windows, macOS, 或 Linux 均可。\n*   **前置依赖**：\n    *   现代 Web 浏览器（用于在线阅读或下载 PDF）。\n    *   PDF 阅读器（如 Adobe Acrobat, Chrome, Edge 等）。\n    *   （可选）Git 客户端：如果你希望将列表克隆到本地以便离线浏览或贡献内容。\n\n## 安装步骤（获取资源）\n\n你可以通过以下两种方式获取论文清单：\n\n### 方式一：在线浏览（推荐）\n直接访问 GitHub 仓库页面查看整理好的分类列表和链接：\n```text\nhttps:\u002F\u002Fgithub.com\u002Fhurshd0\u002Fmust-read-papers-for-ml\n```\n\n### 方式二：克隆到本地\n如果你习惯在本地管理文档或打算提交 PR 补充论文，可以使用 Git 克隆：\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fhurshd0\u002Fmust-read-papers-for-ml.git\ncd must-read-papers-for-ml\n```\n\n> **提示**：国内用户若遇到克隆速度慢的问题，可使用国内镜像源（如 Gitee 镜像，若有）或配置 Git 代理加速。\n\n## 基本使用\n\n本项目的核心用法是**按图索骥**，根据作者标记的优先级阅读论文。\n\n### 1. 理解优先级标记\n作者在 README 中使用了奖牌图标来建议阅读顺序：\n*   🥇 (`:1st_place_medal:`)：**首选阅读**。该领域最基础或最重要的论文。\n*   🥈 (`:2nd_place_medal:`)：**其次阅读**。进阶内容或重要变体。\n*   🥉 (`:3rd_place_medal:`)：**再次阅读**。补充材料或特定场景下的优化方案。\n\n### 2. 开始学习示例\n假设你想学习 **机器学习中的降维技术 (Dimensionality Reduction)**：\n\n1.  在仓库页面或本地 `README.md` 文件中找到 `## Machine Learning` -> `:scissors: Dimensionality Reduction` 章节。\n2.  **第一步**：点击 🥇 标记的链接阅读 [A Tutorial on Principal Component Analysis](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1404.1100.pdf)（PCA 教程）。\n3.  **第二步**：点击 🥈 标记的链接阅读 [How to Use t-SNE Effectively](https:\u002F\u002Fdistill.pub\u002F2016\u002Fmisread-tsne\u002F)（如何有效使用 t-SNE）。\n4.  **第三步**：结合 🥉 标记的原始论文深入理解算法细节。\n\n### 3. 阅读建议\n正如项目作者所言，阅读包含大量数学公式的论文具有挑战性：\n*   不要因初次看不懂而气馁。\n*   建议反复阅读（\"read once, read twice, read thrice...\"），直到完全理解。\n*   配合项目中提供的著名博客（如 Sebastian Raschka, Chip Huyen, Andrej Karpathy 等）的解读文章辅助理解。\n\n### 4. 参与贡献\n如果你发现链接失效，或有重要的论文缺失，欢迎通过以下方式贡献：\n```bash\n# 在本地修改 README.md 添加新论文链接后\ngit add .\ngit commit -m \"Add new paper: [Paper Title]\"\ngit push origin main\n# 然后在 GitHub 上发起 Pull Request\n```","某金融科技公司的高级算法工程师李明，正负责构建一个高精准度的信贷违约预测模型，急需夯实理论基础并复现前沿算法。\n\n### 没有 must-read-papers-for-ml 时\n- **文献检索如大海捞针**：在 Google Scholar 和 arXiv 上盲目搜索关键词，花费数天筛选，却难以辨别哪些是真正奠定领域基石的经典论文。\n- **知识体系支离破碎**：自学过程中遗漏了如 Leo Breiman 的“统计建模两种文化”等关键思想，导致对模型泛化能力的理解存在盲区。\n- **复现核心算法受阻**：在尝试优化梯度提升树时，因未读到 XGBoost 或 LightGBM 的原始论文，无法深入理解其底层剪枝策略与效率优化细节。\n- **陷入数学恐惧症**：面对晦涩的数学推导缺乏心理建设和循序渐进的阅读指引，容易产生挫败感而半途而废。\n\n### 使用 must-read-papers-for-ml 后\n- **获取权威精选书单**：直接利用仓库中按“数据处理、通用机器学习、 boosting\"分类的金牌榜单，瞬间锁定 Hadley Wickham 和 Sebastian Raschka 等人的必读之作。\n- **构建完整认知框架**：通过研读清单中关于“黑盒模型可视化”和“数据沙普利值”的论文，快速建立起从模型训练到可解释性评估的完整闭环。\n- **精准攻克技术难点**：按图索骥找到 XGBoost 和 AdaBoost 的原始论文，深入掌握算法精髓，成功将模型训练速度提升 40% 且 AUC 指标显著优化。\n- **获得持续学习动力**：遵循 README 中“读一遍不懂就读三遍”的鼓励建议，配合清晰的优先级标记（🥇🥈🥉），从容应对复杂的数学公式。\n\nmust-read-papers-for-ml 将原本耗时数周的文献调研工作压缩至几小时，为工程师提供了一条通往机器学习核心殿堂的捷径。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fhurshd0_must-read-papers-for-ml_2f5a61d7.png","hurshd0","Hursh Desai","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fhurshd0_d6078d78.png","ML\u002FAI Infra Eng | JHU Alum ",null,"Dallas, TX","https:\u002F\u002Fgithub.com\u002Fhurshd0",1340,181,"2026-04-07T12:49:02","MIT",1,"","未说明",{"notes":87,"python":85,"dependencies":88},"该工具并非可执行的软件或代码库，而是一个 curated（精心策划）的机器学习、数据科学和深度学习论文、评论及文章列表。它不需要安装任何运行环境、依赖库或硬件资源。用户只需通过浏览器访问 README 中提供的链接即可阅读相关文献。",[],[90,14,16],"其他",[92,93,94,95,96,97,98,99,100,101,102,103],"deep-learning","machine-learning","data-science","papers","neural-networks","convolutional-networks","recurrent-neural-networks","recommender-system","rnn-lstm","exploratory-data-analysis","data-analysis","generalized-additive-models","2026-03-27T02:49:30.150509","2026-04-08T10:08:02.467445",[],[]]