{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:LIAPXWD24COATQIBUVRXAB5GKH","short_pith_number":"pith:LIAPXWD2","schema_version":"1.0","canonical_sha256":"5a00fbd87ae09c09c101a5637007a651f4c7012224855124a643c32a97d940d0","source":{"kind":"arxiv","id":"2505.21061","version":1},"attestation_state":"computed","paper":{"title":"LPOI: Listwise Preference Optimization for Vision Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Fatemeh Pesaran zadeh, Gunhee Kim, Yoojin Oh","submitted_at":"2025-05-27T11:47:28Z","abstract_excerpt":"Aligning large VLMs with human preferences is a challenging task, as methods like RLHF and DPO often overfit to textual information or exacerbate hallucinations. Although augmenting negative image samples partially addresses these pitfalls, no prior work has employed listwise preference optimization for VLMs, due to the complexity and cost of constructing listwise image samples. In this work, we propose LPOI, the first object-aware listwise preference optimization developed for reducing hallucinations in VLMs. LPOI identifies and masks a critical object in the image, and then interpolates the "},"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":"2505.21061","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-05-27T11:47:28Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"4b9ad3bcf327dd6b1ae93969f9642dd9d26aa5e9b93d1a0fd3e492e11799a1a3","abstract_canon_sha256":"08d527c9ec9df31839bb0eac76e00eb8247161715289c61a7be32c6e5fcc2ff0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:10:23.037342Z","signature_b64":"kc2cDYS++tNykXGYXifXUEWYO1GPyp0HSDUNaiNE4jXEm3AaBRyssK+8ynQfKQr3lcd0nl9XLnXxXeXPi2CiDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5a00fbd87ae09c09c101a5637007a651f4c7012224855124a643c32a97d940d0","last_reissued_at":"2026-07-05T11:10:23.036768Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:10:23.036768Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LPOI: Listwise Preference Optimization for Vision Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Fatemeh Pesaran zadeh, Gunhee Kim, Yoojin Oh","submitted_at":"2025-05-27T11:47:28Z","abstract_excerpt":"Aligning large VLMs with human preferences is a challenging task, as methods like RLHF and DPO often overfit to textual information or exacerbate hallucinations. Although augmenting negative image samples partially addresses these pitfalls, no prior work has employed listwise preference optimization for VLMs, due to the complexity and cost of constructing listwise image samples. In this work, we propose LPOI, the first object-aware listwise preference optimization developed for reducing hallucinations in VLMs. LPOI identifies and masks a critical object in the image, and then interpolates the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.21061","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/2505.21061/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":"2505.21061","created_at":"2026-07-05T11:10:23.036849+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.21061v1","created_at":"2026-07-05T11:10:23.036849+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.21061","created_at":"2026-07-05T11:10:23.036849+00:00"},{"alias_kind":"pith_short_12","alias_value":"LIAPXWD24COA","created_at":"2026-07-05T11:10:23.036849+00:00"},{"alias_kind":"pith_short_16","alias_value":"LIAPXWD24COATQIB","created_at":"2026-07-05T11:10:23.036849+00:00"},{"alias_kind":"pith_short_8","alias_value":"LIAPXWD2","created_at":"2026-07-05T11:10:23.036849+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/LIAPXWD24COATQIBUVRXAB5GKH","json":"https://pith.science/pith/LIAPXWD24COATQIBUVRXAB5GKH.json","graph_json":"https://pith.science/api/pith-number/LIAPXWD24COATQIBUVRXAB5GKH/graph.json","events_json":"https://pith.science/api/pith-number/LIAPXWD24COATQIBUVRXAB5GKH/events.json","paper":"https://pith.science/paper/LIAPXWD2"},"agent_actions":{"view_html":"https://pith.science/pith/LIAPXWD24COATQIBUVRXAB5GKH","download_json":"https://pith.science/pith/LIAPXWD24COATQIBUVRXAB5GKH.json","view_paper":"https://pith.science/paper/LIAPXWD2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.21061&json=true","fetch_graph":"https://pith.science/api/pith-number/LIAPXWD24COATQIBUVRXAB5GKH/graph.json","fetch_events":"https://pith.science/api/pith-number/LIAPXWD24COATQIBUVRXAB5GKH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LIAPXWD24COATQIBUVRXAB5GKH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LIAPXWD24COATQIBUVRXAB5GKH/action/storage_attestation","attest_author":"https://pith.science/pith/LIAPXWD24COATQIBUVRXAB5GKH/action/author_attestation","sign_citation":"https://pith.science/pith/LIAPXWD24COATQIBUVRXAB5GKH/action/citation_signature","submit_replication":"https://pith.science/pith/LIAPXWD24COATQIBUVRXAB5GKH/action/replication_record"}},"created_at":"2026-07-05T11:10:23.036849+00:00","updated_at":"2026-07-05T11:10:23.036849+00:00"}