Pith. sign in

REVIEW 8 cited by

Multimodal Web Navigation with Instruction-Finetuned Foundation Models

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2305.11854 v4 pith:ILUY5LTU submitted 2023-05-19 cs.LG cs.AIstat.ML

classification cs.LGcs.AIstat.ML
keywords agentmodelmodelsmultimodalnavigationwebgumdemonstrationsfoundation
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

The progress of autonomous web navigation has been hindered by the dependence on billions of exploratory interactions via online reinforcement learning, and domain-specific model designs that make it difficult to leverage generalization from rich out-of-domain data. In this work, we study data-driven offline training for web agents with vision-language foundation models. We propose an instruction-following multimodal agent, WebGUM, that observes both webpage screenshots and HTML pages and outputs web navigation actions, such as click and type. WebGUM is trained by jointly finetuning an instruction-finetuned language model and a vision encoder with temporal and local perception on a large corpus of demonstrations. We empirically demonstrate this recipe improves the agent's ability of grounded multimodal perception, HTML comprehension, and multi-step reasoning, outperforming prior works by a significant margin. On the MiniWoB, we improve over the previous best offline methods by more than 45.8%, even outperforming online-finetuned SoTA, humans, and GPT-4-based agent. On the WebShop benchmark, our 3-billion-parameter model achieves superior performance to the existing SoTA, PaLM-540B. Furthermore, WebGUM exhibits strong positive transfer to the real-world planning tasks on the Mind2Web. We also collect 347K high-quality demonstrations using our trained models, 38 times larger than prior work, and make them available to promote future research in this direction.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 8 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Graph-Enhanced Policy Optimization in LLM Agent Training

    cs.AI 2025-10 conditional novelty 6.0 of 10

    GEPO adds graph-centrality-based intrinsic rewards, dynamic discounts, and two-level advantage shaping to group-based RL, improving LLM agent success on ALFWorld, WebShop, and a private Workbench benchmark.

  2. MMBench-GUI: Hierarchical Multi-Platform Evaluation Framework for GUI Agents

    cs.CV 2025-07 conditional novelty 6.0 of 10

    MMBench-GUI provides a multi-platform, four-level benchmark and an efficiency-aware metric, and its experiments indicate visual grounding is the main bottleneck for current GUI agents.

  3. Build the web for agents, not agents for the web

    cs.LG 2025-06 conditional novelty 6.0 of 10

    The paper proposes a paradigm shift: design a standardized Agentic Web Interface for AI agents, rather than adapting agents to human-facing websites.

  4. Thinking vs. Doing: Agents that Reason by Scaling Test-Time Interaction

    cs.LG 2025-06 conditional novelty 6.0 of 10

    Scaling the number of interaction steps, trained via a curriculum over rollout horizon, improves web-agent task success and outperforms scaling per-step reasoning under fixed token budgets.

  5. TAPO: Transition-Aware Policy Optimization for LLM Agents

    cs.LG 2026-07 conditional novelty 5.0 of 10

    Interleaving action-conditioned next-observation supervision with group RL on a shared LLM backbone consistently lifts long-horizon agent success over pure policy optimization.

  6. Turbocharging Web Automation: The Impact of Compressed History States

    cs.CL 2025-07 conditional novelty 5.0 of 10

    A Perceiver-style history compressor that summarizes past web states into 256-token representations improves Mind2Web and WebLINX web-automation accuracy by 1.2-5.4% over a no-history baseline.

  7. WebArXiv: Evaluating Multimodal Agents on Time-Invariant arXiv Tasks

    cs.IR 2025-07 conditional novelty 5.0 of 10

    WebArXiv is a time-invariant 275-task benchmark for multimodal web agents on arXiv, plus a dynamic-reflection prompting method that modestly improves success rates.

  8. LineRetriever: Planning-Aware Observation Reduction for Web Agents

    cs.CL 2025-06 conditional novelty 4.0 of 10

    LineRetriever uses a small LM to select relevant lines from web page observations, cutting context by up to 73% with only small success-rate drops on web agent benchmarks.

Pith tools