Salvato in:
Dettagli Bibliografici
Autori principali: Zhang, Wentao, Wang, Qunbo, Zhao, BoXuan, Zhang, Tao, Wu, Junsheng, Gan, Hongping, Dai, Ling, Deng, Shizhuang, Sun, Shuntong, Liu, Yang
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2512.08366
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915766006185984
author Zhang, Wentao
Wang, Qunbo
Zhao, BoXuan
Zhang, Tao
Wu, Junsheng
Gan, Hongping
Dai, Ling
Deng, Shizhuang
Sun, Shuntong
Liu, Yang
author_facet Zhang, Wentao
Wang, Qunbo
Zhao, BoXuan
Zhang, Tao
Wu, Junsheng
Gan, Hongping
Dai, Ling
Deng, Shizhuang
Sun, Shuntong
Liu, Yang
contents Large language model (LLM) agents often rely on external demonstrations or retrieval-augmented planning, leading to brittleness, poor generalization, and high computational overhead. Inspired by human problem-solving, we propose DuSAR (Dual-Strategy Agent with Reflecting) -- a demonstration-free framework that enables a single frozen LLM to perform co-adaptive reasoning via two complementary strategies: a high-level holistic plan and a context-grounded local policy. These strategies interact through a lightweight reflection mechanism, where the agent continuously assesses progress via a Strategy Fitness Score and dynamically revises its global plan when stuck or refines it upon meaningful advancement, mimicking human metacognitive behavior. On both simulated household (ALFWorld) and real-world web (Mind2Web) environments, DuSAR achieves state-of-the-art performance using only open-source LLMs, substantially outperforming all prior methods without any demonstrations or fine-tuning. Remarkably, it also reduces per-step token consumption by a large margin while maintaining strong task success. Ablation studies confirm the necessity of dual-strategy coordination. Moreover, optional integration of expert demonstrations further boosts performance, highlighting DuSAR's flexibility and compatibility with external knowledge.
format Preprint
id arxiv_https___arxiv_org_abs_2512_08366
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Reflecting with Two Voices: A Co-Adaptive Dual-Strategy Framework for LLM-Based Agent Decision Making
Zhang, Wentao
Wang, Qunbo
Zhao, BoXuan
Zhang, Tao
Wu, Junsheng
Gan, Hongping
Dai, Ling
Deng, Shizhuang
Sun, Shuntong
Liu, Yang
Artificial Intelligence
Large language model (LLM) agents often rely on external demonstrations or retrieval-augmented planning, leading to brittleness, poor generalization, and high computational overhead. Inspired by human problem-solving, we propose DuSAR (Dual-Strategy Agent with Reflecting) -- a demonstration-free framework that enables a single frozen LLM to perform co-adaptive reasoning via two complementary strategies: a high-level holistic plan and a context-grounded local policy. These strategies interact through a lightweight reflection mechanism, where the agent continuously assesses progress via a Strategy Fitness Score and dynamically revises its global plan when stuck or refines it upon meaningful advancement, mimicking human metacognitive behavior. On both simulated household (ALFWorld) and real-world web (Mind2Web) environments, DuSAR achieves state-of-the-art performance using only open-source LLMs, substantially outperforming all prior methods without any demonstrations or fine-tuning. Remarkably, it also reduces per-step token consumption by a large margin while maintaining strong task success. Ablation studies confirm the necessity of dual-strategy coordination. Moreover, optional integration of expert demonstrations further boosts performance, highlighting DuSAR's flexibility and compatibility with external knowledge.
title Reflecting with Two Voices: A Co-Adaptive Dual-Strategy Framework for LLM-Based Agent Decision Making
topic Artificial Intelligence
url https://arxiv.org/abs/2512.08366