Issue 4 — Sequence Models course, Gary Marcus vs Yann LeCun, Deep Learning Matrix Calculus…

On People…
Debate among two leading researchers, Gary Marcus and Yann LeCun, on whether AI needs more innate machinery — Link

Dan Jurafsky, discusses how to process the language of policing and answers “Does This Vehicle Belong to You?” — Link

On Code and Data…
FAIR releases PyTorch implementation of “Poincaré Embeddings for Learning Hierarchical Representations” — Link

An interesting podcast discussing NLP highlights and the state-of-the-art techniques. In this episode, Dan Roth discusses strategies on how to deal with cases where data is lacking — Link

Tutorial on how to use Mechanical Turk for tagging your training datasets — Link

Introducing the “Broad Twitter Corpus”, which is a dataset containing tweets collected over stratified times, places and social uses — Link

Beginner’s tutorial on how to train and visualize word vectors (embeddings) — Link

Tensorflow code used for Capsule model implemented in the paper “Dynamic Routing between Capsules” — Link

On Education and Research… releases new course on sequence models, introducing topics such as GRUs, LSTMs, and Recurrent Neural Networks (RNNs) — Link

Paper: “Beyond Word Importance: Contextual Decomposition to Extract Interactions from LSTMs” — Link

Free online book release by Jeremy Howard discussing the fundamental maths you need for deep learning: “The Matrix Calculus You Need For Deep Learning” — Link

Another beautiful online website for testing your regular expressions — Link

On Industry…
Yann LeCun steps down as Facebook’s head of AI research group, FAIR — Link

Andrew Ng continues his mission to build an AI-powered society with the formation of the AI Fund — Link

Simple tests show the shallowness of the state-of-the-art Google translate — Link

On using NLP and machine learning for generating clinical labels of medical scans — Link

Worthy Mentions…
The “NLP News” newsletter (Poincaré embeddings, trolling trolls, A2C comic, General AI Challenge, heuristics for writing, year of PyTorch, BlazingText, MaskGAN, Moments in Time) — Link

NLP Chronicles #1: Interdisciplinary studies, ethics, communication, MOOCs, inclusiveness … — Link

The “Deep Learning Weekly” newsletter (Efficient ML, Learning Rates, Detectron, Mobile) — Link

Here is the previous release (#3) of the NLP Newsletter (Emotional Chatbot, Google Colab, Detectron, Deep Learning Course, Google Free GPU, …) — Link

Update: Last week’s newsletter reached over 500 people 😍; we hope you love this one too!

If you spot any errors or inaccuracies in this newsletter please open an issue. Submit a pull request if you would like to add important NLP news here.

Source: Deep Learning on Medium