Hi! New AI Weekly is here! Enjoy your weekend reading AI news and don’t forget to share it with your friends 😉
How AI driven by Google’s DeepMind could help save lives and the NHS – Artificial Intelligence could prevent tens of thousands of deaths a year and take pressure off NHS staff, but privacy campaigners are worried. Rob Hastings investigates as he meets the pioneering health teams at Google’s AI firm DeepMind and the Royal Free hospital
Explain yourself, machine. Producing simple text descriptions for AI interpretability. – One big theme in 2017 in AI research was the idea of interpretability. How should AI systems explain their decisions to engender trust in their humans users? Can we trust a decision if we don’t understand the factors that informed it?
AI at Google: principles
Reinforcement Learning from scratch
Realtime Interactive Visualization of Convolutional Neural Networks in Unity
Playing card detection with YOLO
TensorFlow 1.9.0-rc0 released
Realtime tSNE Visualizations with TensorFlow.js – In recent years, the t-distributed Stochastic Neighbor Embedding (tSNE) algorithm has become one of the most used and insightful techniques for exploratory data analysis of high-dimensional data. Used to interpret deep neural network outputs in tools such as the TensorFlow Embedding Projector and TensorBoard, a powerful feature of tSNE is that it reveals clusters of high-dimensional data points at different scales while requiring only minimal tuning of its parameters.
Intel Open Sources NLP Architect – Intel AL Lab has open-sourced a library of natural language processing tool, which will be a help to developers creating chatbots and building skills for virtual assistants. It is one of several AI resources made available since Intel AI Lab was launched last year. NLP Architect is a Python library for exploring deep learning topologies and techniques for natural language processing and natural language understanding and is intended to be a platform for future research and collaboration.
Mix&Match – Agent Curricula for Reinforcement Learning – Mix&Match (M&M) – a training framework designed to facilitate rapid and effective learning in RL agents, especially those that would be too slow or too challenging to train otherwise. The key innovation is a procedure that allows us to automatically form a curriculum over agents. Through such a curriculum authors can progressively train more complex agents by, effectively, bootstrapping from solutions found by simpler agents.
Improving Deep Learning Performance with AutoAugment – The success of deep learning in computer vision can be partially attributed to the availability of large amounts of labeled training data — a model’s performance typically improves as you increase the quality, diversity and the amount of training data. However, collecting enough quality data to train a model to perform well is often prohibitively difficult. One way around this is to hardcode image symmetries into neural network architectures so they perform better or have experts manually design data augmentation methods, like rotation and flipping, that are commonly used to train well-performing vision models. However, until recently, less attention has been paid to finding ways to automatically augment existing data using machine learning.
Playing Atari with Six Neurons – Deep reinforcement learning on Atari games maps pixel directly to actions; internally, the deep neural network bears the responsibility of both extracting useful information and making decisions based on it. Aiming at devoting entire deep networks to decision making alone, authors propose a new method for learning policies and compact state representations separately but simultaneously for policy approximation in reinforcement learning. State representations are generated by a novel algorithm based on Vector Quantization and Sparse Coding, trained online along with the network, and capable of growing its dictionary size over time. This enables networks of only 6 to 18 neurons to learn to play a selection of Atari games with performance comparable to state-of-the-art techniques using evolution strategies on deep networks two orders of magnitude larger.
LPIRC CVPR 2018 – Track one of LPIRC CVPR 2018 is a public competition for fast and accurate on-device image classifiers. It is driven by the need for accurate image classification models that run real-time on mobile devices.
CSS10: A Collection of Single Speaker Speech Datasets for 10 Languages – a collection of single speaker speech datasets for ten languages. It is composed of short audio clips from LibriVox audiobooks and their aligned texts. To validate its quality we train two neural text-to-speech models on each dataset.