Deep Hunt — Issue #58
We’re back after a long time — here are the highlights! NVIDIA Launches Year-Long Research Residency Program; EMNLP just ends in Brussels; Google open sources BERT — State of Art Pre-Training for NLP; Reinforcement Learning with Prediction Based Rewards
News
Google launches $25 million AI Impact Challenge
Selected organizations will receive customized support to help bring their ideas to life: coaching from Google’s AI experts, Google.org grant funding from a $25M pool, credit and consulting from Google Cloud, and more.
Sentieo raises $19M to be the AI-powered Bloomberg Terminal
Investment in AI and related companies has picked up a lot and Sentieo — an investment research software suite that uses AI to scan financial documents, analyze alternative data sets and create visualizations, is the latest to pick up huge funding.
NVIDIA Launches Year-Long Research Residency Program
During the one-year, paid program, residents will be paired with an NVIDIA research scientist on a joint project and have the opportunity to publish and present their findings at prominent research conferences such as CVPR, IUCLR and ICML.
Articles
EMNLP —short for Conference on Empirical Methods in Natural Language Processing, happened this week in Belgium and here’s a take from Claudia Hauff on what were some interesting papers presented there.
Tutorials, Tools and Tips
Playing Mortal Kombat with TensorFlow.js. Transfer learning and data augmentation
Here’s a step by step tutorial Minko Gechev on how to train your own model and play Mortal Kombat using posture detection.
Open Sourcing BERT: State-of-the-Art Pre-training for Natural Language Processing
Google opensources a new technique for NLP pre-training called Bidirectional Encoder Representations from Transformers, or BERT. With this release, anyone in the world can train their own state-of-the-art question answering system (or a variety of other models) in about 30 minutes on a single Cloud TPU, or in a few hours using a single GPU.
Trask set out with a goal to write the easiest intro possible to Deep Learning and this repository is an effort in that direction. It accompanies his upcoming book — Grokking Deep Learning. So, do check it out.
Research
Reinforcement Learning with Prediction-Based Rewards
Yet another game — Montezuma’s revenge is conquered by machines. Random Network Distillation by Open AI achieves state-of-the-art performance, periodically finds all 24 rooms and solves the first level without using demonstrations or having access to the underlying state of the game.
If you like what you are reading, please follow and recommend to your friends or give a shoutout on Twitter! I’m glad to hear your suggestions and recommendations @deephunt_in or in comments below!