An AI-driven alarm system helps rescue patients before infections become fatal. The problem: Machine learning can spot patterns in electronic health data indicating where a patient’s condition is headed that may be too subtle for doctors and nurses to catch.

Will deep learning discover new medicines? Startups — and big-pharma partners — are betting on it. The problem: In theory, there’s a pharmacological cure for just about any ailment. In practice, discovering those therapies takes years and billions of dollars.

Are your scientist friends intimidated by machine learning? They might be inspired by a primer from one of the world’s premier tech titans. Former Google CEO Eric Schmidt and Cornell PhD candidate Maithra Raghu school scientists in machine learning in a sprawling overview.

Robots are moving into a job that traditionally required the human touch.What’s new: A commercial warehouse that ships electrical supplies deployed AI-driven robotic arms from Covariant, a high-profile Silicon Valley robotics firm.

AlphaGo Zero demonstrates superhuman performance playing Go, chess, and shogi. Models like R2D2 do the same playing classic Atari titles. A new approach to deep reinforcement learning is the first to achieve state-of-the-art results playing both board and video games.

How is it that many people learn to drive a car fairly safely in 20 hours of practice, while current imitation learning algorithms take hundreds of thousands of hours, and reinforcement learning algorithms take millions of hours? Clearly we’re missing something big.

Many people in the AI community focus on achieving flashy results, like building an agent that can win at Go or Jeopardy. This kind of work is impressive in terms of complexity.

We’ve had great success with supervised deep learning on labeled data. Now it’s time to explore other ways to learn: training on unlabeled data, lifelong learning, and especially letting models explore a simulated environment before transferring what they learn to the real world.

The future of machine learning may depend less on amassing ground-truth data than simulating the environment in which a model will operate. Deep learning works like magic with enough high-quality data. When examples are scarce, though, researchers are using simulation to fill the gap.

Reinforcement learning trains models by trial and error. In batch reinforcement learning (BRL), models learn by observing many demonstrations by a variety of actors. But what if one doctor is handier with a scalpel while another excels at suturing?

In reinforcement learning, if researchers want an agent to have an internal representation of its environment, they’ll build and train a world model that it can refer to. New research shows that world models can emerge from standard training, rather than needing to be built separately.

Is there any reason to continue playing games that AI has mastered? Ask the former champions who have been toppled by machines.

Large NLP models like BERT can answer questions about a document thanks to the transformer network, a sequence-processing architecture that retains information across much longer sequences than previous methods. But transformers have had little success in reinforcement learning — until now.

Neuroevolution, which combines neural networks with ideas drawn from Darwin, is gaining momentum. Its advocates claim that they can achieve faster, better results by generating a succession of new models, each slightly different than its predecessors, rather than relying on a purpose-built model.

At the BlizzCon gaming convention last weekend, players of the strategy game StarCraft II stood in line to get walloped by DeepMind’s AI. After training for the better part of a year, the bot has become one of the world’s top players.