SEAL: SEmantic-Augmented Imitation Learning via Language Model

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Gu, Chengyang, Pan, Yuxin, Bai, Haotian, Xiong, Hui, Chen, Yize
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866929525529509888
author Gu, Chengyang
Pan, Yuxin
Bai, Haotian
Xiong, Hui
Chen, Yize
author_facet Gu, Chengyang
Pan, Yuxin
Bai, Haotian
Xiong, Hui
Chen, Yize
contents Hierarchical Imitation Learning (HIL) is a promising approach for tackling long-horizon decision-making tasks. While it is a challenging task due to the lack of detailed supervisory labels for sub-goal learning, and reliance on hundreds to thousands of expert demonstrations. In this work, we introduce SEAL, a novel framework that leverages Large Language Models (LLMs)'s powerful semantic and world knowledge for both specifying sub-goal space and pre-labeling states to semantically meaningful sub-goal representations without prior knowledge of task hierarchies. SEAL employs a dual-encoder structure, combining supervised LLM-guided sub-goal learning with unsupervised Vector Quantization (VQ) for more robust sub-goal representations. Additionally, SEAL incorporates a transition-augmented low-level planner for improved adaptation to sub-goal transitions. Our experiments demonstrate that SEAL outperforms state-of-the-art HIL methods and LLM-based planning approaches, particularly in settings with small expert datasets and complex long-horizon tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2410_02231
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SEAL: SEmantic-Augmented Imitation Learning via Language Model
Gu, Chengyang
Pan, Yuxin
Bai, Haotian
Xiong, Hui
Chen, Yize
Artificial Intelligence
Machine Learning
Systems and Control
Hierarchical Imitation Learning (HIL) is a promising approach for tackling long-horizon decision-making tasks. While it is a challenging task due to the lack of detailed supervisory labels for sub-goal learning, and reliance on hundreds to thousands of expert demonstrations. In this work, we introduce SEAL, a novel framework that leverages Large Language Models (LLMs)'s powerful semantic and world knowledge for both specifying sub-goal space and pre-labeling states to semantically meaningful sub-goal representations without prior knowledge of task hierarchies. SEAL employs a dual-encoder structure, combining supervised LLM-guided sub-goal learning with unsupervised Vector Quantization (VQ) for more robust sub-goal representations. Additionally, SEAL incorporates a transition-augmented low-level planner for improved adaptation to sub-goal transitions. Our experiments demonstrate that SEAL outperforms state-of-the-art HIL methods and LLM-based planning approaches, particularly in settings with small expert datasets and complex long-horizon tasks.
title SEAL: SEmantic-Augmented Imitation Learning via Language Model
topic Artificial Intelligence
Machine Learning
Systems and Control
url https://arxiv.org/abs/2410.02231