TextQL, Colorless Green RNNs, ConvAI2, Machine Learning Yearning, Meta-Learning Tutorial, Tinn, World Models,…
Welcome to the 10th Issue of NLP Newsletter! Here are this week’s notable NLP news!
On People…
Andrew Ng plans to soon roll-out first draft of his book, Machine Learning Yearning — Link
David Ha et al., presents “World Models”, where they aim to answer “Can agents learn inside of their own dreams” — Link
An amazing tutorial by Jaan Altosaar on everything there is to know about variational autoencoders — Link
On Education and Research…
This interesting paper presents a “quantitative analysis on the use of skin tone modifiers on emoji on Twitter” — Link
Paper discussing experiments on neural machine translation using transformer model and Tensor2Tensor framework — Link
Exciting paper on word embeddings applies to massive source of medical data — Link
Step-by-step implementation of “Attention Is All You Need” using PyTorch — Link
An impressive piece teaching how to implement Meta-Learning (learning to learn), which is a very recent trend in machine learning — Link
Here is a nice piece providing details on the most important and conventional text processing techniques important in text mining — Link
On Code and Data…
TextQL is a library that allows you to execute against CSV or TSV text files — Link
A neat and very detailed tutorial on how to use pandas dataframes — Link
Tinn (Tiny Neural Network) is a simple dependency-free neural network library written in C — Link
Code for the paper entitled “Colorless green recurrent networks dream hierarchically” (by Facebook) — Link
On Industry…
LSTM implementation: Intel Skylake vs NVIDIA V100 — Link
PyTorch now installable through pip — Link
Helping kids develop language skills using NLP (by MIT) — Link
Quotes of the Week…
“Essentially, all models are wrong, but some are useful” — George E.P.Box
“It’s not news that NSA etc is using #NLProc techniques to trawl through language data — — and anything world-readable on social media is presumably considered fair game.” — Emily Bender
Worthy Mentions…
Using machine learning to discover neural network optimizers (Google Research Blog) — Link
ConvAI2 conversation chatbot competition by NIPS2018 — Link
Analysis of scientific publications using distributed word representations — Link
Natural Language Processing News by Sebastian Ruder — Link
