Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces WebChallenger, a web agent framework designed to enhance autonomous web navigation for LLM agents by addressing three cognitive advantages of human navigation: selective attention, persistent memory, and procedural fluency. By utilizing a structured page representation called PageMem, which organizes web pages into a hierarchy of semantic sections, the framework allows agents to efficiently extract relevant information and navigate websites without the need for site-specific adaptations. The results demonstrate that WebChallenger achieves competitive performance on various web navigation benchmarks using off-the-shelf models, significantly reducing inference costs associated with proprietary systems.
WebChallenger achieves competitive web navigation performance with off-the-shelf models, slashing costs while mimicking human cognitive advantages.
Autonomous web navigation remains challenging for LLM agents, and the strongest generalist systems rely on proprietary reasoning models whose inference cost is prohibitive for the repetitive tasks where such agents would be most useful. We argue this gap stems not from insufficient model capability but from agent architectures that fail to replicate three human cognitive advantages: selective attention to relevant page regions, persistent memory of website structure, and procedural fluency with common interaction patterns. We introduce WebChallenger, a web agent framework that addresses each gap through architecture design rather than model scale, built around PageMem: a structured page representation deterministically constructed from the DOM that exposes each page as a hierarchy of semantic sections with short summaries. On this shared substrate we build three mechanisms that mirror the three cognitive advantages: a divide-and-conquer observation pipeline that lets the agent skim section summaries and extract details only from task-relevant regions; a lightweight exploration and memory system that traverses each website once to build a reusable map of pages and element behaviors; and compound action workflows that collapse common multi-step interactions into single agent actions, handling partial state changes automatically. Because all three operate over PageMem, the framework generalizes across websites without site-specific adapters. Using off-the-shelf open-weight models without fine-tuning, our system achieves 56.3% on WebArena, 48.7% on VisualWebArena, 51.0% on Online-Mind2Web, and 70.9% on WorkArena, approaching frontier proprietary systems at a fraction of the cost. Our code is released at https://github.com/jayoohwang1/webchallenger