机器学习人工学weekly-2018/7/1

1. 网红Andrey Karpathy分享在Tesla的工作，非常有趣

Building the Software 2.0 Stack by Andrej Karpathy from Tesla

链接：https://www.figure-eight.com/building-the-software-2-0-stack-by-andrej-karpathy-from-tesla/

2. OpenAI Five打Dota 2

链接：https://blog.openai.com/openai-five/?utm_source=ActiveCampaign&utm_medium=email&utm_content=Transforming+Industries+With+AI+++IoT%2C+OpenAI+Improves+Language+Understanding%2C+China+s+AI-Infused+Retail+Store&utm_campaign=Weekly+Newsletter+06+27+2018+Issue+99

3. Uber相关

3.1 Uber Eats送外卖

How Trip Inferences and Machine Learning Optimize Delivery Times on Uber Eats

链接：https://eng.uber.com/uber-eats-trip-optimization/?utm_campaign=ARCHITECHT&utm_medium=email&utm_source=ARCHITECHT_35

3.2 Uber防fraud

Advanced Technologies for Detecting and Preventing Fraud at Uber

链接：https://eng.uber.com/advanced-technologies-detecting-preventing-fraud-uber/?utm_campaign=ARCHITECHT&utm_medium=email&utm_source=ARCHITECHT_35

4. Stitch Fix用matrix factorization理解用户和style的关系

Understanding Latent Style

链接：https://multithreaded.stitchfix.com/blog/2018/06/28/latent-style/

5. 深度学习做视频动作识别

Deep Learning for Videos: A 2018 Guide to Action Recognition

链接：http://blog.qure.ai/notes/deep-learning-for-videos-action-recognition-review

6. NLP实战

A Practitioner's Guide to Natural Language Processing

链接：https://towardsdatascience.com/a-practitioners-guide-to-natural-language-processing-part-i-processing-understanding-text-9f4abfd13e72

7. feature selection工具

A Feature Selection Tool for Machine Learning in Python

链接：https://towardsdatascience.com/a-feature-selection-tool-for-machine-learning-in-python-b64dd23710f0

8. PyTorch/Keras对比

Keras or PyTorch as your first deep learning framework

链接：https://deepsense.ai/keras-or-pytorch/

9. bias/variance tradeoff的可视化讨论

Model Tuning and the Bias-Variance Tradeoff

链接：http://www.r2d3.us/visual-intro-to-machine-learning-part-2/

10. DQN踢FIFA足球

Using Deep Q-Learning in FIFA 18 to perfect the art of free-kicks

链接：https://towardsdatascience.com/using-deep-q-learning-in-fifa-18-to-perfect-the-art-of-free-kicks-f2e4e979ee66

11. 用CycleGAN把fortnite的场景渲染成吃鸡的场景

Turning Fortnite into PUBG with Deep Learning (CycleGAN)

链接：https://towardsdatascience.com/turning-fortnite-into-pubg-with-deep-learning-cyclegan-2f9d339dcdb0

12. inverse RL介绍

Learning from humans: what is inverse reinforcement learning?

链接：https://thegradient.pub/learning-from-humans-what-is-inverse-reinforcement-learning/

13. 结合ML和运筹学优化出行时间

Travel Time Optimization With Machine Learning And Genetic Algorithm

链接：https://towardsdatascience.com/travel-time-optimization-with-machine-learning-and-genetic-algorithm-71b40a3a4c2

14. 用fastText对knowledge graph的实体关系分类

Using fastText and Comet.ml to classify relationships in Knowledge Graphs

链接：https://medium.com/comet-ml/using-fasttext-and-comet-ml-to-classify-relationships-in-knowledge-graphs-e73d27b40d67

15. architecture search

15.1 可differentiate的architect search search，通过使用双层优化

DARTS: Differentiable Architecture Search

链接：https://arxiv.org/pdf/1806.09055v1.pdf

15.2 百度省资源的NAS

Resource-Efficient Neural Architect

链接：https://arxiv.org/pdf/1806.07912.pdf

16. BAIR blog，让机器人看一次视频演示就学会做任务

One-Shot Imitation from Watching Videos

链接：http://bair.berkeley.edu/blog/2018/06/28/daml/

17. object detection详细教程

One-shot object detection

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

18. 各类工具

18.1 这个网站非常不错，专门收集论文和代码

链接：https://paperswithcode.com/

18.2 小米发布移动端深度学习inference框架MACE

链接：https://github.com/XiaoMi/mace

18.3 metacar开车RL环境

链接：https://www.metacar-project.com/

18.4 intel发布模型压缩工具distiller

链接：https://github.com/NervanaSystems/distiller