{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:GBKI5NPL6WU4CFHEDEUJPXC4TF","short_pith_number":"pith:GBKI5NPL","schema_version":"1.0","canonical_sha256":"30548eb5ebf5a9c114e4192897dc5c99782bda7ad7b87373664702f5e9fffa0b","source":{"kind":"arxiv","id":"2606.22158","version":1},"attestation_state":"computed","paper":{"title":"Improving Reasoning in Vision-Language Models via Perception Verified Self-Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Sadbhawna Thakur, Sonam Gupta, Sourabh Sharma","submitted_at":"2026-06-20T17:33:07Z","abstract_excerpt":"Achieving human-like reasoning in Vision-Language Models (VLMs) remains a long-standing challenge. Recent approaches leverage Chain-of-Thought (CoT) rationales generated by human annotators or proprietary models to improve reasoning, which is costly and difficult to scale. Self-training offers a promising alternative by using models own outputs as supervision. However, existing methods often suffer from visual hallucinations -- where rationales describe non-existent visual content, and language shortcuts -- where predictions rely on textual priors rather than true visual grounding, as rational"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.22158","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-06-20T17:33:07Z","cross_cats_sorted":[],"title_canon_sha256":"daa5ffa614144acb6d83d8496763d6037b720d1f83b1b2001ba78b126ade3b27","abstract_canon_sha256":"ed320cc57fdf501bdaa4a6350f5a0e1d7b724af7b3884273d0d80db9197202f4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T02:13:29.876969Z","signature_b64":"8s3swfIKKZbrKsOm1zDRutm6mMNihUX5SWbNhfEwl1dTDH4fsvudxzQ/RACeVOfF8PJcTNutxYtEHDD98+GEDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"30548eb5ebf5a9c114e4192897dc5c99782bda7ad7b87373664702f5e9fffa0b","last_reissued_at":"2026-06-23T02:13:29.876590Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T02:13:29.876590Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improving Reasoning in Vision-Language Models via Perception Verified Self-Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Sadbhawna Thakur, Sonam Gupta, Sourabh Sharma","submitted_at":"2026-06-20T17:33:07Z","abstract_excerpt":"Achieving human-like reasoning in Vision-Language Models (VLMs) remains a long-standing challenge. Recent approaches leverage Chain-of-Thought (CoT) rationales generated by human annotators or proprietary models to improve reasoning, which is costly and difficult to scale. Self-training offers a promising alternative by using models own outputs as supervision. However, existing methods often suffer from visual hallucinations -- where rationales describe non-existent visual content, and language shortcuts -- where predictions rely on textual priors rather than true visual grounding, as rational"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.22158","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2606.22158/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.22158","created_at":"2026-06-23T02:13:29.876652+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.22158v1","created_at":"2026-06-23T02:13:29.876652+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.22158","created_at":"2026-06-23T02:13:29.876652+00:00"},{"alias_kind":"pith_short_12","alias_value":"GBKI5NPL6WU4","created_at":"2026-06-23T02:13:29.876652+00:00"},{"alias_kind":"pith_short_16","alias_value":"GBKI5NPL6WU4CFHE","created_at":"2026-06-23T02:13:29.876652+00:00"},{"alias_kind":"pith_short_8","alias_value":"GBKI5NPL","created_at":"2026-06-23T02:13:29.876652+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GBKI5NPL6WU4CFHEDEUJPXC4TF","json":"https://pith.science/pith/GBKI5NPL6WU4CFHEDEUJPXC4TF.json","graph_json":"https://pith.science/api/pith-number/GBKI5NPL6WU4CFHEDEUJPXC4TF/graph.json","events_json":"https://pith.science/api/pith-number/GBKI5NPL6WU4CFHEDEUJPXC4TF/events.json","paper":"https://pith.science/paper/GBKI5NPL"},"agent_actions":{"view_html":"https://pith.science/pith/GBKI5NPL6WU4CFHEDEUJPXC4TF","download_json":"https://pith.science/pith/GBKI5NPL6WU4CFHEDEUJPXC4TF.json","view_paper":"https://pith.science/paper/GBKI5NPL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.22158&json=true","fetch_graph":"https://pith.science/api/pith-number/GBKI5NPL6WU4CFHEDEUJPXC4TF/graph.json","fetch_events":"https://pith.science/api/pith-number/GBKI5NPL6WU4CFHEDEUJPXC4TF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GBKI5NPL6WU4CFHEDEUJPXC4TF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GBKI5NPL6WU4CFHEDEUJPXC4TF/action/storage_attestation","attest_author":"https://pith.science/pith/GBKI5NPL6WU4CFHEDEUJPXC4TF/action/author_attestation","sign_citation":"https://pith.science/pith/GBKI5NPL6WU4CFHEDEUJPXC4TF/action/citation_signature","submit_replication":"https://pith.science/pith/GBKI5NPL6WU4CFHEDEUJPXC4TF/action/replication_record"}},"created_at":"2026-06-23T02:13:29.876652+00:00","updated_at":"2026-06-23T02:13:29.876652+00:00"}