REGENT: A Retrieval-Augmented Generalist Agent That Can Act In-Context in New Environments
This addresses the problem of deploying adaptable AI agents in digital and real-world settings, offering a more efficient alternative to scaling current architectures.
The paper tackles the challenge of building generalist agents that can adapt to new environments by proposing REGENT, a retrieval-augmented agent that uses in-context learning without finetuning, achieving up to 3x fewer parameters and an order-of-magnitude fewer pre-training datapoints while outperforming state-of-the-art agents.
Building generalist agents that can rapidly adapt to new environments is a key challenge for deploying AI in the digital and real worlds. Is scaling current agent architectures the most effective way to build generalist agents? We propose a novel approach to pre-train relatively small policies on relatively small datasets and adapt them to unseen environments via in-context learning, without any finetuning. Our key idea is that retrieval offers a powerful bias for fast adaptation. Indeed, we demonstrate that even a simple retrieval-based 1-nearest neighbor agent offers a surprisingly strong baseline for today's state-of-the-art generalist agents. From this starting point, we construct a semi-parametric agent, REGENT, that trains a transformer-based policy on sequences of queries and retrieved neighbors. REGENT can generalize to unseen robotics and game-playing environments via retrieval augmentation and in-context learning, achieving this with up to 3x fewer parameters and up to an order-of-magnitude fewer pre-training datapoints, significantly outperforming today's state-of-the-art generalist agents. Website: https://kaustubhsridhar.github.io/regent-research