机器学习人工学weekly-2018/6/17

注意下面很多链接需要科学上网，无奈国情如此

1. Twitter内部DeepBird平台使用TF

Twitter meets TensorFlow

链接：https://blog.twitter.com/engineering/en_us/topics/insights/2018/twittertensorflow.html

2. DeepMind通过YouTube游戏视频来攻克Atari里面最难的3个游戏

Playing hard exploration games by watching YouTube

链接：https://arxiv.org/pdf/1805.11592.pdf

2.X graph networks

链接：https://arxiv.org/pdf/1806.01261.pdf

3. OpenAI继续推进unsupervised learning在NLP上的应用

Improving Language Understanding with Unsupervised Learning

链接：https://blog.openai.com/language-unsupervised/?utm_campaign=ARCHITECHT&utm_medium=email&utm_source=ARCHITECHT_34

4. Ian Goodfellow好文讲对adversarial attack攻击的防守

Defense Against the Dark Arts

链接：https://arxiv.org/pdf/1806.04169.pdf

5. kaggle kernel: 预测世界杯比赛冠军

Predicting the winner of the 2018 FIFA World Cup

链接：https://www.kaggle.com/agostontorok/soccer-world-cup-2018-winner?utm_medium=email&utm_source=mailchimp&utm_campaign=datanotes-20160614

6. 网红Chris Olah访谈，讲模型解释性

视频链接：https://www.youtube.com/watch?v=3o4VzEyJ0WA

配合文章链接：https://distill.pub/2018/building-blocks/

7. 水果Create ML教程

Training a Text Classifier with Create ML and the Natural Language Framework

链接：https://flight.school/articles/classifying-programming-languages-with-createml/?utm_campaign=Revue%20newsletter&utm_medium=Newsletter&utm_source=Deep%20Learning%20Weekly

8. InfoGAN讨论

链接：http://www.depthfirstlearning.com/2018/InfoGAN#2-generative-adversarial-networks-gan

9. object detection长篇好文

one-shot object detection

链接：http://machinethink.net/blog/object-detection/?utm_campaign=Revue%20newsletter&utm_medium=Newsletter&utm_source=Deep%20Learning%20Weekly

10. 教程/资源

10.1 attention教程

链接：https://github.com/greentfrapp/attention-primer

10.2 tf eager mode教程

链接：https://github.com/madalinabuzau/tensorflow-eager-tutorials

10.3 ml5.js，基于tf.js的js库

链接：https://ml5js.org/