Machine Learning Weekly Review №4 — the source of latest credible papers, videos and projects on machine learning for scientists and engineers.

Recommended this week


  1. Sketch-RNN-JS — javascript implementation for sketch-rnn (i.e. a Generative Model for Vector Drawings) by hard maru.
  2. Malware Env for OpenAI Gym — create an AI that bypasses machine learning static-analysis malware detection.
  3. OpenPose — A Real-Time Multi-Person Keypoint Detection And Multi-Threading C++ Library.
  4. Texture — an open science manuscript editor. As open as LaTeX and as simple as a classic word processor.
  5. Densely Connected Convolutional Networks implemented in Keras, PyTorch, MXNet, TF and more.
  6. Tensorflow implementation of Enhanced Deep Residual Networks for Single Image Super-Resolution.
  7. Tiny Face Detector. CVPR 2017.


  1. Annotating Object Instances with a Polygon-RNN”. Given a bounding box, automatically predicts the polygon outlining the object instance inside the box. CVPR 2017.
  2. A correspondence between thermodynamics and inference”. Proposes an approach that unifies the Bayesian, frequentist and information-based paradigms of statistics by achieving coherent inference between them.
  3. Learning from Simulated and Unsupervised Images through Adversarial Training”. Paper by Apple Researches that earned “Best Paper Prize” CVPR 2017.
  4. When Neurons Fail”. Proves a tight bounds on the number of neurons that can fail without harming the result of a computation.
  5. “Mimicking Word Embeddings using Subword RNNs”. Paper by Yuval Pinter infers out-of-vocabulary(OOV) word embeddings from pre-trained models, without the originating corpus.
  6. Going beyond average for reinforcement learning” by Marc G. Bellemare, DeepMind.
  7. A Tutorial on Thompson Sampling” by DeepMind Daniel Russo
  8. Learning Cross-modal Embeddings for Cooking Recipes and Food Images”. CVPR 2017.


  1. PyData Seattle 2017” and “PyData Berlin 2017” playlists.
  2. Analyzing eye-movement and pupil-size data with Python DataMatrix.


  1. A collection of papers and books for Getting Started with Genetic Analysis. From Prof. John Storey, Princeton University.
  2. Deep Learning — The Straight Dope”. An interactive book on deep learning, in concept and in MXNet by Zack Chase Lipton.
  3. “How to train your own Object Detector with TensorFlow’s Object Detector API” by Dat Tran.
  4. Bayesian Neural Networks with Random Inputs for Model Based Reinforcement Learning” by José Miguel Hernández Lobato.

You would also like

Facebook Group: Machine Learning Review