Reinforcement Learning (RL) is a machine learning (ML) technique to learn sequential
decision-making in complex problems. RL is inspired by trial-and-error based human/animal …

Replay in the brain has been viewed as rehearsal or, more recently, as sampling from a
transition model. Here, we propose a new hypothesis: that replay is able to implement a form …

Recent efforts have incorporated large language models (LLMs) with external resources (eg,
the Internet) or internal control flows (eg, prompt chaining) for tasks requiring grounding or …

Sequential decision making, commonly formalized as Markov Decision Process (MDP)
optimization, is an important challenge in artificial intelligence. Two key approaches to this …

Experience replay is central to off-policy algorithms in deep reinforcement learning (RL), but
there remain significant gaps in our understanding. We therefore present a systematic and …

A hallmark of human intelligence is the ability to plan multiple steps into the future,. Despite
decades of research,–, it is still debated whether skilled decision-makers plan more steps …

We propose a novel policy update that combines regularized policy optimization with model
learning as an auxiliary loss. The update (henceforth Muesli) matches MuZero's state-of-the …

AlphaZero is a powerful reinforcement learning algorithm based on approximate policy
iteration and tree search. However, AlphaZero can fail to improve its policy network, if not …

Model-based planning is often thought to be necessary for deep, careful reasoning and
generalization in artificial agents. While recent successes of model-based reinforcement …

Abstract The combination of Monte-Carlo tree search (MCTS) with deep reinforcement
learning has led to groundbreaking results in artificial intelligence. However, AlphaZero, the …