{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:FFZKSTMR2G5MIUBLZGEG2R6P6U","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"118bf059e09fbe7927df304ab6d8bb2c863f08dcaba997245484db0b223c2087","cross_cats_sorted":["cs.AI","cs.CL","cs.IR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-02T06:40:42Z","title_canon_sha256":"d08b1190e6ccce64ca8ad55466d40ca9103daa69a66afaeaa7dd9130256fe2f7"},"schema_version":"1.0","source":{"id":"2605.01284","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.01284","created_at":"2026-05-26T01:03:31Z"},{"alias_kind":"arxiv_version","alias_value":"2605.01284v2","created_at":"2026-05-26T01:03:31Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.01284","created_at":"2026-05-26T01:03:31Z"},{"alias_kind":"pith_short_12","alias_value":"FFZKSTMR2G5M","created_at":"2026-05-26T01:03:31Z"},{"alias_kind":"pith_short_16","alias_value":"FFZKSTMR2G5MIUBL","created_at":"2026-05-26T01:03:31Z"},{"alias_kind":"pith_short_8","alias_value":"FFZKSTMR","created_at":"2026-05-26T01:03:31Z"}],"graph_snapshots":[{"event_id":"sha256:470b284130bf706af0fafbb79fbcb01fe94010b2219e800847caa64b56bb0dc5","target":"graph","created_at":"2026-05-26T01:03:31Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"fine-tuned Qwen3-VL-8B-Instruct achieves robust performance, significantly outperforming text-based baselines in scenarios requiring visual layout understanding, while establishing a retriever-agnostic solution for pixel-level interpretable iRAG."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That vision-language models applied to raw document screenshots can reliably recover spatial logic and layout cues that text conversion discards, without format-specific parsing or additional supervision beyond the fine-tuning described."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"CoE applies vision-language models directly to document screenshots to deliver pixel-level bounding-box attribution for evidence in iterative retrieval-augmented generation, outperforming text baselines on visual-layout tasks."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Vision-language models can deliver pixel-level visual evidence chains for iterative retrieval-augmented generation by operating directly on document screenshots."}],"snapshot_sha256":"aac36e72c896125b8a1411355e904d9260963437b70889c9d8e7cd40bb9a88a5"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-20T18:35:45.366112Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T17:27:00.327473Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.01284/integrity.json","findings":[],"snapshot_sha256":"650733d44fd774141e603140059a529820898b3fa431f48d1df8b89b55a1e878","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Iterative Retrieval-Augmented Generation (iRAG) has emerged as a powerful paradigm for answering complex multi-hop questions by progressively retrieving and reasoning over external documents. However, current systems predominantly operate on parsed text, which creates two critical bottlenecks: (1) \\textit{Coarse-grained attribution}, where users are burdened with manually locating evidence within lengthy documents based on vague text-level citations; and (2) \\textit{Visual semantic loss}, where the conversion of visually rich documents (e.g., slides, PDFs with charts) into text discards spatia","authors_text":"Di Liang, Peiyang Liu, Wei Ye, Xi Wang, Ziqiang Cui","cross_cats":["cs.AI","cs.CL","cs.IR"],"headline":"Vision-language models can deliver pixel-level visual evidence chains for iterative retrieval-augmented generation by operating directly on document screenshots.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-02T06:40:42Z","title":"Chain of Evidence: Pixel-Level Visual Attribution for Iterative Retrieval-Augmented Generation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.01284","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-09T14:42:48.086500Z","id":"ee448b55-1017-4652-b5e6-87e9ee16626d","model_set":{"reader":"grok-4.3"},"one_line_summary":"CoE applies vision-language models directly to document screenshots to deliver pixel-level bounding-box attribution for evidence in iterative retrieval-augmented generation, outperforming text baselines on visual-layout tasks.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Vision-language models can deliver pixel-level visual evidence chains for iterative retrieval-augmented generation by operating directly on document screenshots.","strongest_claim":"fine-tuned Qwen3-VL-8B-Instruct achieves robust performance, significantly outperforming text-based baselines in scenarios requiring visual layout understanding, while establishing a retriever-agnostic solution for pixel-level interpretable iRAG.","weakest_assumption":"That vision-language models applied to raw document screenshots can reliably recover spatial logic and layout cues that text conversion discards, without format-specific parsing or additional supervision beyond the fine-tuning described."}},"verdict_id":"ee448b55-1017-4652-b5e6-87e9ee16626d"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:41dc2a8abf4423821f42c4664306a2c36bcca9cc142d90ca339756672d0c8824","target":"record","created_at":"2026-05-26T01:03:31Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"118bf059e09fbe7927df304ab6d8bb2c863f08dcaba997245484db0b223c2087","cross_cats_sorted":["cs.AI","cs.CL","cs.IR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-02T06:40:42Z","title_canon_sha256":"d08b1190e6ccce64ca8ad55466d40ca9103daa69a66afaeaa7dd9130256fe2f7"},"schema_version":"1.0","source":{"id":"2605.01284","kind":"arxiv","version":2}},"canonical_sha256":"2972a94d91d1bac4502bc9886d47cff5086052b64ce46b316a36ffb3550198e7","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2972a94d91d1bac4502bc9886d47cff5086052b64ce46b316a36ffb3550198e7","first_computed_at":"2026-05-26T01:03:31.680631Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-26T01:03:31.680631Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"6/BtKp86jUs0NFzghgZgCTthmcRYz19eFVhpP7QByZMhm43RUMKabxqYyOw+kuw7cIxmLv+XBNR5A4SEHfDbDw==","signature_status":"signed_v1","signed_at":"2026-05-26T01:03:31.681437Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.01284","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:41dc2a8abf4423821f42c4664306a2c36bcca9cc142d90ca339756672d0c8824","sha256:470b284130bf706af0fafbb79fbcb01fe94010b2219e800847caa64b56bb0dc5"],"state_sha256":"d304cb3d2933db37e774f19f8c1fc31ffc2b971c1810d56aacf482cb7dbf9927"}