AI Paper Atlas

Research map / Multi-agent systems and games / LLM agents

PatchScout: Thematic Web Data Collection via Information Foraging

Michael West, Eduard Dragut

NeurIPS 2026 · Multi-agent systems and games › LLM agents · approach: Agent data synthesis and behavior studies · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Introduces PatchScout, a method for thematic web data collection that gathers semantically related documents scattered across disconnected web regions, framed via information foraging.

Problem

Researchers need reusable web corpora of documents sharing a common theme but spread across heterogeneous, structurally disconnected web sources; existing traversal-based and query-based approaches are insufficient for this.

Approach

Frames thematic web data collection as an information foraging problem, so the crawler forages across decentralized web patches for theme-relevant documents rather than following link structure or answering a single query.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add re-seeding (Fixed-Reseed vs No-Reseed) (ablation)PatchScout · Universitiescoverage (%): 1.7 -> 16.1Improved
add re-seeding (Fixed-Reseed vs No-Reseed) (ablation)PatchScout · ASPscoverage (%): 1.5 -> 30.8Improved
use policy-guided instead of random patch switching (ablation)PatchScout (Scent-EMA) · Law Schoolscoverage (%): 25.4 -> 86.8Improved
use policy-guided instead of random patch switching (ablation)PatchScout (Scent-EMA) · ASPscoverage (%): 9.0 -> 35.3Improved
generate queries adaptively with feedback instead of upfront (ablation)PatchScout (Scent-EMA) · Fortune 500coverage (%): 6.6 -> 24.6Improved
use scent-based link ranking instead of random link selection (ablation)PatchScout (Realized-EMA) · Universitiesharvest rate (%): 53.5 -> 86.2Improved
use scent-based re-seeding instead of realized-productivity re-seedingPatchScout · Universitiescoverage (%): Realized-Mean 2.4 -> Scent-Mean 20.3Improved
use realized-productivity policyPatchScout (Realized-Mean) · Universitiesharvest rate (%): HR 73.1, coverage 2.4Mixed
use fixed threshold re-seeding instead of finer-grained scent estimatesPatchScout (Threshold) · ASPscoverage (%): 35.3 -> 39.1Improved
PatchScout vs deep-research agents when an authoritative directory existsPatchScout (best) · Law Schoolscoverage (%): Gemini 93.4 vs 86.8Hurt
PatchScout vs deep-research agents on large target setPatchScout (Scent-Mean) · Universitiescoverage (%): 1.8 -> 20.3Improved
use smaller Qwen critic instead of DeepSeek criticPatchScout Critic · 10,000 sampled pagesagreement (%): 86.1, 87.4, 86.9 for 9B, 27B, 397BNo clear effect

Experimental setup

Baselines: traversal-based methods query-based methods

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy