Enhancing Web Agents with a Hierarchical Memory Tree

TL;DR

This paper introduces a Hierarchical Memory Tree (HMT) that structures web agent memory into levels to improve generalization across unseen websites, addressing workflow mismatches caused by flat memory structures.

Contribution

The paper proposes a novel hierarchical memory framework that decouples planning from execution, enhancing web agents' ability to generalize across diverse and unseen websites.

Findings

HMT outperforms flat-memory methods in cross-website tasks.

Structured memory improves logical consistency and task success.

HMT enhances robustness in complex web interactions.

Abstract

Large language model-based web agents have shown strong potential in automating web interactions through advanced reasoning and instruction following. While retrieval-based memory derived from historical trajectories enables these agents to handle complex, long-horizon tasks, current methods struggle to generalize across unseen websites. We identify that this challenge arises from the flat memory structures that entangle high-level task logic with site-specific action details. This entanglement induces a workflow mismatch in new environments, where retrieved contents are conflated with current web, leading to logically inconsistent execution. To address this, we propose Hierarchical Memory Tree (HMT), a structured framework designed to explicitly decouple logical planning from action execution. HMT constructs a three-level hierarchy from raw trajectories via an automated abstraction pipeline: the Intent level maps diverse user instructions to standardized task goals; the Stage level defines reusable semantic subgoals characterized by observable pre-conditions and post-conditions; and the Action level stores action patterns paired with transferable semantic element descriptions. Leveraging this structure, we develop a stage-aware inference mechanism comprising a Planner and an Actor. By explicitly validating pre-conditions, the Planner aligns the current state with the correct logical subgoal to prevent workflow mismatch, while the Actor grounds actions by matching the stored semantic descriptions to the target page. Experimental results on Mind2Web and WebArena show that HMT significantly outperforms flat-memory methods, particularly in cross-website and cross-domain scenarios, highlighting the necessity of structured memory for robust generalization of web agents.