{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/arxiv-2602-10863","title":"ICA: Information-Aware Credit Assignment for Visually Grounded Long-Horizon Information-Seeking Agents","arxiv_id":"2602.10863","date":"2026-02-11","proceeding":null,"authors":["Cong Pang","Xuyu Feng","Yujie Yi","Jiaqi Su","Zixuan Chen","Jiawei Hong","Tiankuo Yao","Nang Yuan","Jiapeng Luo","Lewei Lu","Xin Lou"],"abstract":"Long-horizon reinforcement learning for information seeking agents remains difficult because terminal rewards reveal whether the final answer is correct, but not which acquired information enabled it. This difficulty is amplified by text-derived webpage observations, where parsing, truncation, and summarization often produce incomplete and unstable content representations across trajectories. We propose an evidence-centric framework for web agent learning that represents information acquired through tools as identifiable units for comparison across trajectories. In particular, fetched webpages are represented as rendered snapshots, preserving layout and multimodal content as stable content-level observations. Building on these units, we introduce Information-Aware Credit Assignment, a post hoc reward propagation method that estimates turn-level utility scores from rollout success rates and assigns dense rewards to intermediate steps that introduced high-utility information. Integrated with GSPO, our method consistently improves performance on BrowseComp, GAIA, Xbench-DS, and Seal-0. Code and datasets will be released at https://github.com/pc-inno/ICA_MM_deepsearch.","url_abs":"https://arxiv.org/abs/2602.10863","url_pdf":"https://arxiv.org/pdf/2602.10863","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2602.10863","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2602.10863"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/pc-inno/ICA_MM_deepsearch","reach":null}],"summary":{"unverified":4},"by_repo_kind":{"found_in_text":{"samples":4,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":4,"samples":[{"code_sha256_prefix":"7dcb327feafec1b5","entry":"load_done_ids","repo":"pc-inno/ICA_MM_deepsearch","repo_kind":"found_in_text","path":"llm_judge.py","file_url":"https://github.com/pc-inno/ICA_MM_deepsearch/blob/HEAD/llm_judge.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"7dcb327feafec1b5"}},{"code_sha256_prefix":"8362cbe6a0734d97","entry":"read_input","repo":"pc-inno/ICA_MM_deepsearch","repo_kind":"found_in_text","path":"run_inference.py","file_url":"https://github.com/pc-inno/ICA_MM_deepsearch/blob/HEAD/run_inference.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"8362cbe6a0734d97"}},{"code_sha256_prefix":"f45882c6806cb8d4","entry":"read_jsonl","repo":"pc-inno/ICA_MM_deepsearch","repo_kind":"found_in_text","path":"llm_judge.py","file_url":"https://github.com/pc-inno/ICA_MM_deepsearch/blob/HEAD/llm_judge.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"f45882c6806cb8d4"}},{"code_sha256_prefix":"91442ee48ebfc784","entry":"strip_think","repo":"pc-inno/ICA_MM_deepsearch","repo_kind":"found_in_text","path":"llm_judge.py","file_url":"https://github.com/pc-inno/ICA_MM_deepsearch/blob/HEAD/llm_judge.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"91442ee48ebfc784"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.LG","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}