{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:LTWHYP5BK3AXMP4XSVVRIRNI6Q","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":"f3a288909e3ddee0339012bde8bc17676df114c7ef398a2f4b48d6150606d5d6","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-05-11T16:49:36Z","title_canon_sha256":"e178d81e90584fcf37865ee23101edc46bb6f4f5cdd66e76965a755c80b267b4"},"schema_version":"1.0","source":{"id":"2605.10832","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.10832","created_at":"2026-06-08T01:04:07Z"},{"alias_kind":"arxiv_version","alias_value":"2605.10832v2","created_at":"2026-06-08T01:04:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.10832","created_at":"2026-06-08T01:04:07Z"},{"alias_kind":"pith_short_12","alias_value":"LTWHYP5BK3AX","created_at":"2026-06-08T01:04:07Z"},{"alias_kind":"pith_short_16","alias_value":"LTWHYP5BK3AXMP4X","created_at":"2026-06-08T01:04:07Z"},{"alias_kind":"pith_short_8","alias_value":"LTWHYP5B","created_at":"2026-06-08T01:04:07Z"}],"graph_snapshots":[{"event_id":"sha256:f1a7310dfe649462dbe5ccf85d55516fd39cd0e3e59136d9d4c9fba16bb0de16","target":"graph","created_at":"2026-06-08T01:04:07Z","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":"Across 8 multimodal deep search benchmarks, ODE improves the Qwen3-VL-8B agent from 24.9% to 39.0% on average, surpassing Gemini-2.5 Pro in standard agent-workflow setting (37.9%). At 30B, ODE raises the average score from 30.6% to 41.5%."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That rollouts from the current policy produce training data that accurately identifies and fills the precise capability gaps without introducing self-reinforcing biases or training instability."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"A new image-bank harness and closed-loop on-policy data evolution method raises multimodal agent performance on visual search benchmarks from 24.9% to 39.0% for an 8B model and from 30.6% to 41.5% for a 30B model."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"On-policy data evolution from agent rollouts boosts multimodal deep search performance from 24.9% to 39% on average."}],"snapshot_sha256":"5bb6ee0af1ca64bfb9cf6447b9a1bea750562ba4153f266f382f115c32e5b2a0"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-20T05:22:00.342835Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T14:34:47.442439Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T10:31:17.587837Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T08:57:25.750037Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.10832/integrity.json","findings":[],"snapshot_sha256":"d5d3cb52ab2d139c70b8feedfeceeb355eb45c3f7582d09b4f05e0f3292d6e6a","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Multimodal deep search requires an agent to solve open-world problems by chaining search, tool use, and visual reasoning over evolving textual and visual context. Two bottlenecks limit current systems. First, existing tool-use harnesses treat images returned by search, browsing, or transformation as transient outputs, so intermediate visual evidence cannot be re-consumed by later tools. Second, training data is usually built by fixed curation recipes that cannot track the target agent's evolving capability. To address these challenges, we first introduce a visual-native agent harness centered ","authors_text":"Chenxin Li, Guanting Dong, Hangyu Guo, Hongru Wang, Junting Lu, Shijue Huang, Shuang Chen, Xinyu Geng, Yi R. Fung, Zhaochen Su, Zhenyu Li","cross_cats":[],"headline":"On-policy data evolution from agent rollouts boosts multimodal deep search performance from 24.9% to 39% on average.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-05-11T16:49:36Z","title":"Towards On-Policy Data Evolution for Visual-Native Multimodal Deep Search Agents"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.10832","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-12T04:10:59.872308Z","id":"7a3df81e-29a5-422c-ad95-61721d7662e6","model_set":{"reader":"grok-4.3"},"one_line_summary":"A new image-bank harness and closed-loop on-policy data evolution method raises multimodal agent performance on visual search benchmarks from 24.9% to 39.0% for an 8B model and from 30.6% to 41.5% for a 30B model.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"On-policy data evolution from agent rollouts boosts multimodal deep search performance from 24.9% to 39% on average.","strongest_claim":"Across 8 multimodal deep search benchmarks, ODE improves the Qwen3-VL-8B agent from 24.9% to 39.0% on average, surpassing Gemini-2.5 Pro in standard agent-workflow setting (37.9%). At 30B, ODE raises the average score from 30.6% to 41.5%.","weakest_assumption":"That rollouts from the current policy produce training data that accurately identifies and fills the precise capability gaps without introducing self-reinforcing biases or training instability."}},"verdict_id":"7a3df81e-29a5-422c-ad95-61721d7662e6"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:9c2908ac010601009fbc8c90471910b618a906a2b13d64604eb99ccc6c473cd7","target":"record","created_at":"2026-06-08T01:04:07Z","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":"f3a288909e3ddee0339012bde8bc17676df114c7ef398a2f4b48d6150606d5d6","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-05-11T16:49:36Z","title_canon_sha256":"e178d81e90584fcf37865ee23101edc46bb6f4f5cdd66e76965a755c80b267b4"},"schema_version":"1.0","source":{"id":"2605.10832","kind":"arxiv","version":2}},"canonical_sha256":"5cec7c3fa156c1763f97956b1445a8f42f43b136365004d76ce88b6d7247b0c5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5cec7c3fa156c1763f97956b1445a8f42f43b136365004d76ce88b6d7247b0c5","first_computed_at":"2026-06-08T01:04:07.308478Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-08T01:04:07.308478Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"mQjsZL7omd1buuO+1dRQDhpA0Sco6PmjfKeJQLj9WGAAQa8vmb3HMHiE5WBWBc6Y326HAlBD1OzKYRtmWBaIDQ==","signature_status":"signed_v1","signed_at":"2026-06-08T01:04:07.309241Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.10832","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9c2908ac010601009fbc8c90471910b618a906a2b13d64604eb99ccc6c473cd7","sha256:f1a7310dfe649462dbe5ccf85d55516fd39cd0e3e59136d9d4c9fba16bb0de16"],"state_sha256":"ebfc331661f72b527f5dc605c81e4191cffe58ac8a3471babe6d1c36ae6ed498"}