Google's Decision Transformer
Reinforcement Learning

Reinforcement Learning Transformed: Transformers succeed at reinforcemend learning tasks.

Transformers have matched or exceeded earlier architectures in language modeling and image classification. New work shows they can achieve state-of-the-art results in some reinforcement learning tasks as well.
Animated chart shows how AI can help robots locate key spatial coordinates.
Reinforcement Learning

Finding Useful Points in Space: Keypoint3D Helps Robots Locate Spatial Coordinates

A new machine learning method aims to improve a machine’s ability to determine and locate points of interest.
Animated video showing an image generator imitating brush strokes
Reinforcement Learning

Different Strokes for Robot Folks: Transformer-Based Image Generator Imitates Painters

A neural network can make a photo resemble a painting via neural style transfer, but it can also learn to reproduce an image by applying brush strokes. A new method taught a system this painterly skill without any training data.
Series of images explaining how the system Eva works
Reinforcement Learning

Who Needs a Covid Test? AI Decides: Greek Border Patrol Used AI to Screen for Covid

Greece’s border agents last year had enough Covid tests to swab only 17 percent of people who sought to enter the country. They managed the shortage by using an AI system to flag high-risk visitors.
Animation showing a simulated football team and how it works
Reinforcement Learning

Team Players: Football-Playing AI Blends Individual and Group Skills

Playing a team sport involves a fluid blend of individual and group skills. Researchers integrated both types of action into realistic humanoid agents that play football (known as soccer in the U.S.).
Series of images showing some of the findings of the new study by researchers at Stanford’s Human AI Institute
Reinforcement Learning

Weak Foundations Make Weak Models: Foundation AI Models Pass Flaws to Fine-Tuned Variants

A new study examines a major strain of recent research: huge models pretrained on immense quantities of uncurated, unlabeled data and then fine-tuned on a smaller, curated corpus.
Sequence of famous arcade games' scenes
Reinforcement Learning

Solve RL With This One Weird Trick: How to get better performance from reinforcement learning.

The previous state-of-the-art model for playing vintage Atari games took advantage of a number of advances in reinforcement learning (RL). The new champion is a basic RL architecture plus a trick borrowed from image generation.
Forbidden sign over a robot's hand solving a Rubik's Cube
Reinforcement Learning

Bye Bye Bots: OpenAI quit robotics to focus on AGI.

The independent research lab OpenAI wowed technology watchers in 2019 with a robotic hand that solved Rubik’s Cube. Now it has disbanded the team that built it. OpenAI cofounder Wojciech Zaremba revealed that OpenAI shuttered its robotics program last October.
A four-legged robot walking over difficult and changing terrain
Reinforcement Learning

Walking the Dog: Training a robot to walk over unsteady terrain with RL.

A reinforcement learning system enabled a four-legged robot to amble over unfamiliar, rapidly changing terrain.
Automated player learning by watching recorded gameplay
Reinforcement Learning

Behavioral Cloning Shootout: AI learns to play Counter Strike Global Offensive.

Neural networks have learned to play video games like Dota 2 via reinforcement learning by playing for the equivalent of thousands of years (compressed into far less time). In new work, an automated player learned not by playing for millennia but by watching a few days’ worth of recorded gameplay.
On the left, the policy is being trained from scratch, and on the right, a pre-trained policy is being fine-tuned
Reinforcement Learning

Computers Making Computers: How Google used AI to help design its TPU v4 chip.

A neural network wrote the blueprint for upcoming computer chips that will accelerate deep learning itself. Google engineers used a reinforcement learning system to arrange the billions of minuscule transistors in an upcoming version of its Tensor Processing Unit (TPU) chips.
Surgical robots performing different actions
Reinforcement Learning

Medical AI Gets a Grip: An AI System controlled DaVinci surgical robots.

Surgical robots perform millions of delicate operations annually under human control. Now they’re getting ready to operate on their own.
Video sequence showing military drones working
Reinforcement Learning

Drones For Defense: How companies like Anduril are developing military drones.

Drone startups are taking aim at military customers. As large tech companies have backed away from defense work, startups like Anduril, Shield AI, and Teal are picking up the slack. They’re developing autonomous fliers specifically for military operations.
Graphs and data related to recurrent neural nets (RNNs)
Reinforcement Learning

Performance Guaranteed: How deep learning networks can become Bayes-optimal.

Bayes-optimal algorithms always make the best decisions given their training and input, if certain assumptions hold true. New work shows that some neural networks can approach this kind of performance.
Graphs related to world models
Reinforcement Learning

It’s a Small World Model After All: More efficient world models for reinforcement learning

World models, which learn a compressed representation of a dynamic environment like, say, a video game, have delivered top results in reinforcement learning. A new method makes them much smaller.

Subscribe to The Batch

Stay updated with weekly AI News and Insights delivered to your inbox