Reinforcement Learning - The Batch

The table shows MAI-Thinking-1 leading in several benchmarks, compared to other AI models.

Machine Learning Research

Microsoft Strikes Out on Its Own: Microsoft revealed MAI-Thinking-1, a Claude Sonnet 4.6-sized reasoning model developed without distillation

Microsoft, once OpenAI’s exclusive partner and still a major reseller of other companies’ AI models, built its own reasoning model from scratch.

Flowchart illustrates the POPE method, transitioning from guided to unguided problem-solving in reinforcement learning.

Machine Learning Research

Reinforcement Learning With Hints: Privileged On-Policy Exploration (POPE) trains models to expand on partial solutions

Reinforcement learning can’t train a model to solve a difficult problem if the model doesn’t discover all the right steps.

Infographic showing Nvidia's chip design flow, highlighting placer, router, and optimization stages.

Machine Learning Research

How Nvidia Uses AI to Design Chips: Chipmaker's models design circuits, verify designs, and test new layouts

Nvidia’s chief scientist dreams of telling an AI model to design a new GPU, then skiing for a couple days while the system does the job.

Flowchart shows data reordering, probability sampling, and effective gradient updating in reinforcement learning.

Machine Learning Research

Faster Reinforcement Learning: New technique auto-selects training examples to speed up fine-tuning

Fine-tuning large language models via reinforcement learning is computationally expensive, but researchers found a way to streamline the process.

Researcher adjusts sensors on a quadruped robot disguised as a Tibetan antelope in a high-altitude reserve.

Robotics

Robot Antelope Joins Herd: Chinese scientists disguise modified robot dog as antelope to study herd behavior

Researchers in China disguised a quadruped robot as a Tibetan antelope to help study the animals close-up.

Machine Learning Research

Robo-Football From Simulation to Reality: Reinforcement learning powers humanoid robots to play football

Humanoid robots can play football (known as soccer in the United States) in the real world, thanks to reinforcement learning.

Machine Learning Research

Learning Language by Exploration: Agent develops language skills through simulated exploration tasks

Machine learning models typically learn language by training on tasks like predicting the next word in a given text. Researchers trained a language model in a less focused, more human-like way.

Assembly pseudocode before and after applying the AlphaDev swap move

Machine Learning Research

AI Builds Better Sorting Algorithms: AlphaDev, a new system for high-speed sorting of lists and numbers

Online sorting algorithms run trillions of times a day to organize lists according to users’ interests. New work found faster alternatives. Daniel J. Mankowitz and colleagues at Google developed AlphaDev, a system that learned to generate algorithms that sort three...

Machine Learning Research

Stratego Master: DeepNash, the RL system that plays Stratego like a master

Reinforcement learning agents have mastered games like Go that provide complete information about the state of the game to players. They’ve also excelled at Texas Hold ’Em poker, which provides incomplete information, as few cards are revealed.

Machine Learning Research

Bug Finder: A system that provides feedback with near human-level accuracy

One challenge to making online education available worldwide is evaluating an immense volume of student work. Especially difficult is evaluating interactive computer programming assignments such as coding a game.

Science

Letting Chatbots See Your Data: Coding framework LlamaIndex enables data interaction with LLMs

A new coding framework lets you pipe your own data into large language models. LlamaIndex streamlines the coding involved in enabling developers to summarize, reason over, and otherwise manipulate data from documents, databases, and apps using models like GPT-4.

Machine Learning Research

Optimizing Matrix Multiplication: AlphaTensor for faster matrix multiplication, explained

Matrix multiplication is executed so often in deep learning, video games, and scientific computing that even a slight acceleration can save substantial amounts of processing time. New work finds ways to speed up this crucial operation.