WebDancer: Towards Autonomous Information Seeking Agency

TL;DR

WebDancer is an end-to-end web agent framework that combines data construction, sampling, supervised fine-tuning, and reinforcement learning to improve autonomous information seeking, demonstrating strong performance on benchmark tasks.

Contribution

This work introduces a novel training paradigm for agentic information seeking agents, integrating multiple stages to enhance generalisation and effectiveness.

Findings

WebDancer outperforms existing benchmarks on GAIA and WebWalkerQA.

The training paradigm significantly improves agent generalisation.

Analysis provides insights for future development of agentic models.

Abstract

Addressing intricate real-world problems necessitates in-depth information seeking and multi-step reasoning. Recent progress in agentic systems, exemplified by Deep Research, underscores the potential for autonomous multi-step research. In this work, we present a cohesive paradigm for building end-to-end agentic information seeking agents from a data-centric and training-stage perspective. Our approach consists of four key stages: (1) browsing data construction, (2) trajectories sampling, (3) supervised fine-tuning for effective cold start, and (4) reinforcement learning for enhanced generalisation. We instantiate this framework in a web agent based on the ReAct, WebDancer. Empirical evaluations on the challenging information seeking benchmarks, GAIA and WebWalkerQA, demonstrate the strong performance of WebDancer, achieving considerable results and highlighting the efficacy of our training paradigm. Further analysis of agent training provides valuable insights and actionable, systematic pathways for developing more capable agentic models. The codes and demo will be released in https://github.com/Alibaba-NLP/WebAgent.