1. Judea Pearl上次在NIPS有一张令人唏嘘的照片,不过现在他又回来了,发了新书也给了一个访谈,说深度学习就像是curve fitting(我觉得没错,lol)
链接:https://www.quantamagazine.org/to-build-truly-intelligent-machines-teach-them-cause-and-effect-20180515/
causal inference我觉得是非常重要的,这里有个简介
ML beyond Curve Fitting: An Intro to Causal Inference and do-Calculus
链接:http://www.inference.vc/untitled/
2. Google 相关
2.1 Google Machine Learning Practica培训教程
链接:https://developers.google.com/machine-learning/practica/
2.2 Google用RL搜索数据增强的方法,这个也是AutoML继architect和optimizer之后的新进展,非常有意思的工作
AutoAugment: Learning Augmentation Policies from Data
链接:https://arxiv.org/pdf/1805.09501.pdf
3. pinterest海量数据推荐
Pixie: A System for Recommending 3+ Billion Items to 200+ Million Users in Real-Time
链接:http://delivery.acm.org/10.1145/3190000/3186183/p1775-eksombatchai.pdf?ip=180.168.218.163&id=3186183&acc=TRUSTED&key=4D4702B0C3E38B35%2E4D4702B0C3E38B35%2E4D4702B0C3E38B35%2EE47D41B086F0CDA3&__acm__=1527227693_300e8dd8aee7ce1ffd8dc318fd2cd31c
4. OpenAI Gym Retro加了很多新游戏,有很多红白机的,比如冒险岛,马戏团等等
链接:https://blog.openai.com/gym-retro/
5. Intel Nervana开源NLP库
链接:https://github.com/NervanaSystems/nlp-architect
6. POS tagging实战
链接:https://becominghuman.ai/part-of-speech-tagging-tutorial-with-the-keras-deep-learning-library-d7f93fa05537
7. 四种推荐引擎方法
链接:https://towardsdatascience.com/the-4-recommendation-engines-that-can-predict-your-movie-tastes-109dc4e10c52
8. adversarial攻击实战
"I Pity the fool", Deep Learning style
链接:https://blog.godatadriven.com/rod-fool-neural-network?utm_campaign=Revue%20newsletter&utm_medium=Newsletter&utm_source=Deep%20Learning%20Weekly
9. 模型评价metrics
Choosing the Right Metric for Evaluating ML Models
part1: https://towardsdatascience.com/choosing-the-right-metric-for-machine-learning-models-part-1-a99d7d7414e4
part2:https://towardsdatascience.com/choosing-the-right-metric-for-evaluating-machine-learning-models-part-2-86d5649a5428
本文分享自 机器学习人工学weekly 微信公众号,前往查看
如有侵权,请联系 cloudcommunity@tencent.com 删除。
本文参与 腾讯云自媒体同步曝光计划 ,欢迎热爱写作的你一起参与!