Guardado en:
Detalles Bibliográficos
Autores principales: Sridhar, Kaustubh, Dutta, Souradeep, Jayaraman, Dinesh, Lee, Insup
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:https://arxiv.org/abs/2412.04759
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910841958301696
author Sridhar, Kaustubh
Dutta, Souradeep
Jayaraman, Dinesh
Lee, Insup
author_facet Sridhar, Kaustubh
Dutta, Souradeep
Jayaraman, Dinesh
Lee, Insup
contents 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
format Preprint
id arxiv_https___arxiv_org_abs_2412_04759
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle REGENT: A Retrieval-Augmented Generalist Agent That Can Act In-Context in New Environments
Sridhar, Kaustubh
Dutta, Souradeep
Jayaraman, Dinesh
Lee, Insup
Artificial Intelligence
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
title REGENT: A Retrieval-Augmented Generalist Agent That Can Act In-Context in New Environments
topic Artificial Intelligence
url https://arxiv.org/abs/2412.04759