{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:P6ONOIG6JJICVTJLRHSXWGEXZT","short_pith_number":"pith:P6ONOIG6","schema_version":"1.0","canonical_sha256":"7f9cd720de4a502acd2b89e57b1897ccc1e16b107065fd48273194a7df56602d","source":{"kind":"arxiv","id":"2312.04554","version":1},"attestation_state":"computed","paper":{"title":"Improved Visual Grounding through Self-Consistent Explanations","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Alexander C. Berg, Paola Cascante-Bonilla, Ruozhen He, Vicente Ordonez, Ziyan Yang","submitted_at":"2023-12-07T18:59:22Z","abstract_excerpt":"Vision-and-language models trained to match images with text can be combined with visual explanation methods to point to the locations of specific objects in an image. Our work shows that the localization --\"grounding\"-- abilities of these models can be further improved by finetuning for self-consistent visual explanations. We propose a strategy for augmenting existing text-image datasets with paraphrases using a large language model, and SelfEQ, a weakly-supervised strategy on visual explanation maps for paraphrases that encourages self-consistency. Specifically, for an input textual phrase, "},"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":"2312.04554","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2023-12-07T18:59:22Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"a5f9c803197cda0735c4c41e12cb74401d6f4d125bcc154900ae4858ffebaa8a","abstract_canon_sha256":"84921f235db1626ca484519aa1132ea4c905d5da2b1ac286b7df85584328ad4d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:21:35.921922Z","signature_b64":"JVcDEef8LPBPrZjhGPwP+TinW24Prl/GycFLC7/5xKgwrx/ullRT2EWI4mYGjorGSZuKbbKgDOKIPDyZszioDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7f9cd720de4a502acd2b89e57b1897ccc1e16b107065fd48273194a7df56602d","last_reissued_at":"2026-07-05T07:21:35.921410Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:21:35.921410Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improved Visual Grounding through Self-Consistent Explanations","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Alexander C. Berg, Paola Cascante-Bonilla, Ruozhen He, Vicente Ordonez, Ziyan Yang","submitted_at":"2023-12-07T18:59:22Z","abstract_excerpt":"Vision-and-language models trained to match images with text can be combined with visual explanation methods to point to the locations of specific objects in an image. Our work shows that the localization --\"grounding\"-- abilities of these models can be further improved by finetuning for self-consistent visual explanations. We propose a strategy for augmenting existing text-image datasets with paraphrases using a large language model, and SelfEQ, a weakly-supervised strategy on visual explanation maps for paraphrases that encourages self-consistency. Specifically, for an input textual phrase, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.04554","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/2312.04554/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":"2312.04554","created_at":"2026-07-05T07:21:35.921485+00:00"},{"alias_kind":"arxiv_version","alias_value":"2312.04554v1","created_at":"2026-07-05T07:21:35.921485+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.04554","created_at":"2026-07-05T07:21:35.921485+00:00"},{"alias_kind":"pith_short_12","alias_value":"P6ONOIG6JJIC","created_at":"2026-07-05T07:21:35.921485+00:00"},{"alias_kind":"pith_short_16","alias_value":"P6ONOIG6JJICVTJL","created_at":"2026-07-05T07:21:35.921485+00:00"},{"alias_kind":"pith_short_8","alias_value":"P6ONOIG6","created_at":"2026-07-05T07:21:35.921485+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/P6ONOIG6JJICVTJLRHSXWGEXZT","json":"https://pith.science/pith/P6ONOIG6JJICVTJLRHSXWGEXZT.json","graph_json":"https://pith.science/api/pith-number/P6ONOIG6JJICVTJLRHSXWGEXZT/graph.json","events_json":"https://pith.science/api/pith-number/P6ONOIG6JJICVTJLRHSXWGEXZT/events.json","paper":"https://pith.science/paper/P6ONOIG6"},"agent_actions":{"view_html":"https://pith.science/pith/P6ONOIG6JJICVTJLRHSXWGEXZT","download_json":"https://pith.science/pith/P6ONOIG6JJICVTJLRHSXWGEXZT.json","view_paper":"https://pith.science/paper/P6ONOIG6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2312.04554&json=true","fetch_graph":"https://pith.science/api/pith-number/P6ONOIG6JJICVTJLRHSXWGEXZT/graph.json","fetch_events":"https://pith.science/api/pith-number/P6ONOIG6JJICVTJLRHSXWGEXZT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/P6ONOIG6JJICVTJLRHSXWGEXZT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/P6ONOIG6JJICVTJLRHSXWGEXZT/action/storage_attestation","attest_author":"https://pith.science/pith/P6ONOIG6JJICVTJLRHSXWGEXZT/action/author_attestation","sign_citation":"https://pith.science/pith/P6ONOIG6JJICVTJLRHSXWGEXZT/action/citation_signature","submit_replication":"https://pith.science/pith/P6ONOIG6JJICVTJLRHSXWGEXZT/action/replication_record"}},"created_at":"2026-07-05T07:21:35.921485+00:00","updated_at":"2026-07-05T07:21:35.921485+00:00"}