深度学习、机器学习、自然语言处理NLP优秀文章整理

本是古典何须时尚 2022-12-12 13:50 183阅读 0赞

[LSTM 为何如此有效？这五个秘密是你要知道的][LSTM]

[文本挖掘实操课！手把手教你用文本挖掘剖析54万首诗歌][54]

[从发展历史视角解析Transformer：从全连接CNN到Transformer][Transformer_CNN_Transformer]

[生成对抗网络（GAN）的数学原理全解][GAN]

[超细节的BERT/Transformer知识点][BERT_Transformer]

[基于知识引入的情感分析][Link 1]

[深度学习中Attention Mechanism详细介绍：原理、分类及应用][Attention Mechanism]

[融合外部知识的常识问答][Link 2]

[张俊林：对比学习「Contrastive Learning」研究进展精要][Contrastive Learning]

[一文掌握《对比学习（Contrastive Learning）》要旨，详述MoCo和SimCLR算法][Contrastive Learning_MoCo_SimCLR]

[人大团队研究：面向文本生成，预训练模型进展梳理][Link 3]

[如何解决文本分类中的样本不均衡问题][Link 4]

[【Hello NLP】CS224n笔记\[7\]:整理了12小时，只为让你20分钟搞懂Seq2seq][Hello NLP_CS224n_7_12_20_Seq2seq]

[清华团队让 AI 写诗“更上一层楼”，诗歌图灵测试迷惑近半数玩家][AI]

[Knowledge In PLM: 语言模型可以作为一种知识库吗？][Knowledge In PLM_]

[一文了解预训练语言模型][Link 5]

[吊打BERT、GPT、DALL·E，跨模态榜单新霸主诞生！][BERT_GPT_DALL_E]

[FAIR研究员：重新思考BatchNorm中的"Batch"][FAIR_BatchNorm_Batch]

[深度学习中的Normalization你懂了多少？一个公式概括其思想与框架][Normalization]

[为什么机器学习算法难以优化？一文详解算法优化内部机制][Link 6]

[浅谈贝叶斯统计][Link 7][添加链接描述][Link 8]

[神经网络中，设计loss function有哪些技巧?][loss function]

[语言模型微调领域有哪些最新进展？一文详解最新趋势][Link 9]

[深入理解“预训练”][Link 10]

[多样性文本生成任务的研究进展][Link 11]

[如何利用多任务学习提升模型性能？][Link 12]

[深度学习多目标优化的多个loss应该如何权衡][loss]

[【CMU博士论文】可控文本生成，附107页pdf与Slides][Link 8]

[对比学习有多火？文本聚类都被刷爆了…][Link 13]

[预训练语言模型fine-tuning近期进展概述][fine-tuning]

[基于深度学习的文本分类6大算法-原理、结构、论文、源码打包分享][6_-]

持续更新中……

[LSTM]: https://mp.weixin.qq.com/s/33F9Q39DPq22nuTZUleavw
[54]: https://mp.weixin.qq.com/s/B-fsEVViwHFMvCW6pGb-LQ
[Transformer_CNN_Transformer]: https://mp.weixin.qq.com/s/h1fdIHQiPt6MfaaQjKe9zQ
[GAN]: https://mp.weixin.qq.com/s/9ZVP23lij-G40QNJ2ZKgIA
[BERT_Transformer]: https://mp.weixin.qq.com/s/TB5RRvn-W9fjtxjt6reoSw
[Link 1]: https://mp.weixin.qq.com/s/BrqBEfeHmjnrjW-eejcLRg
[Attention Mechanism]: https://mp.weixin.qq.com/s?__biz=MzIxNDgzNDg3NQ==&mid=2247490339&idx=2&sn=d51481a8e1593929a5209c9cc2e4acca&chksm=97a0d0f7a0d759e14b3f430ed59682c7e5ff42d8ed28219fd6493478c257e1c363e7abf4d20c&mpshare=1&scene=1&srcid=1118Mbv86rzUCuH5SZXsDf37&sharer_sharetime=1605681268257&sharer_shareid=aa93c5677c69f70d21796e75d3bfaf99&key=153bf6c0773cb59ed853cdfba10a0102bf6e8811e41b7546db247bafc05fbf15d06273f2479015eb500bd88f3c960db19063d10fc18a573fb6d9af0a9db61cced8c7ed5a3f1a07a86610fbe023ec9aed3416e8df7dfa8efcc2bcd63f2b8f647b4954538d3fddbbbf92541e22e525c7c5342b0f8b9a2abe40bf55ba3051ed75df&ascene=1&uin=MjY0NDUxMTMxMw==&devicetype=Windows%2010%20x64&version=6300002f&lang=zh_CN&exportkey=A%2bYDwcfTfbmsHpi5gLLx/jE=&pass_ticket=vjk9s39MUphAwdlT%2b8wF23%2bQd/2636yB6V8VtG4aUNbYHs5gK2xkFbg5m5CdIZxH&wx_header=0
[Link 2]: https://mp.weixin.qq.com/s/4s_WisgMUaFyPmLanlDXLg
[Contrastive Learning]: https://mp.weixin.qq.com/s/7rL0dbDCUpmJArOZll8n3w
[Contrastive Learning_MoCo_SimCLR]: https://mp.weixin.qq.com/s/2a6xKL61u0M6mD1UMJjV3Q
[Link 3]: https://mp.weixin.qq.com/s/tBQ-i50KrMgT0MiPfNI5bQ
[Link 4]: https://mp.weixin.qq.com/s/PPR9LVWPNypgS-4ujPkQvg
[Hello NLP_CS224n_7_12_20_Seq2seq]: https://mp.weixin.qq.com/s/czTgcG489-5VXftwoSr9KQ
[AI]: https://mp.weixin.qq.com/s/-_v5uKADu7LkH0mzITDGXw
[Knowledge In PLM_]: https://mp.weixin.qq.com/s/uH6z_3gQwoh6MTL2rPR9ew
[Link 5]: https://mp.weixin.qq.com/s/RP51ddihw3nWHOT2G2lQtA
[BERT_GPT_DALL_E]: https://mp.weixin.qq.com/s/7NYe59gKu6-js32tfy4xBw
[FAIR_BatchNorm_Batch]: https://mp.weixin.qq.com/s/6OkobO8Y7TELa1QAki18RA
[Normalization]: https://mp.weixin.qq.com/s/s3eKrh9MZ1ttK2LMa-43KA
[Link 6]: https://mp.weixin.qq.com/s/p4uIAHnMOcxWkTsLozkYpg
[Link 7]: https://editor.csdn.net/md/?articleId=108950432
[Link 8]: https://mp.weixin.qq.com/s/jRsJ-KXNTFn8JMA6NYqWAA
[loss function]: https://mp.weixin.qq.com/s/sZ8aCkIlFO5uo7QE6acqpA
[Link 9]: https://mp.weixin.qq.com/s/XVZSAxaWM30t9rOeXYM03A
[Link 10]: https://mp.weixin.qq.com/s/A6JXMdW3nB4Xvy0SRMykpw
[Link 11]: https://mp.weixin.qq.com/s/rFUYJINuJYZ934rxEhTffQ
[Link 12]: https://mp.weixin.qq.com/s/d4pUY-CMUlikzrfzunwNcA
[loss]: https://mp.weixin.qq.com/s/c7IOB1cF6d0XySKsw6nzvg
[Link 13]: https://mp.weixin.qq.com/s/BWK_tKR6If7Lww8SrRB5Ww
[fine-tuning]: https://mp.weixin.qq.com/s/-_ZNV1jJIqP3YFyykV_BGA
[6_-]: https://mp.weixin.qq.com/s/76PXOTESh3lkbTgH_7xE1A