{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:6NHKRCJ3TH5ONSDO7YM2TGRF43","short_pith_number":"pith:6NHKRCJ3","schema_version":"1.0","canonical_sha256":"f34ea8893b99fae6c86efe19a99a25e6c6a275e877e9b679c2bfefcd391db45d","source":{"kind":"arxiv","id":"2209.06954","version":3},"attestation_state":"computed","paper":{"title":"Correlation Information Bottleneck: Towards Adapting Pretrained Multimodal Models for Robust Visual Question Answering","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jingjing Jiang, Nanning Zheng, Ziyi Liu","submitted_at":"2022-09-14T22:04:10Z","abstract_excerpt":"Benefiting from large-scale pretrained vision language models (VLMs), the performance of visual question answering (VQA) has approached human oracles. However, finetuning such models on limited data often suffers from overfitting and poor generalization issues, leading to a lack of model robustness. In this paper, we aim to improve input robustness from an information bottleneck perspective when adapting pretrained VLMs to the downstream VQA task. Input robustness refers to the ability of models to defend against visual and linguistic input variations, as well as shortcut learning involved in "},"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":"2209.06954","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2022-09-14T22:04:10Z","cross_cats_sorted":[],"title_canon_sha256":"7eb48f5a0a1d3d24a9d842c5ea95a5677cdf1b014e8d69b7e3c2969ef053758d","abstract_canon_sha256":"f2dd29695b826b65e82b0fd8eaf1266739443b9e47744e69d6910bfebdbec827"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:07:32.732211Z","signature_b64":"7iZWy0GX7PA+WA7AdJwfyYSvuoCm2RPvZHMzMmhHefTQkk4VY54QfyATp0vRH/n43nObNnDRkBhp4VRbgUypDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f34ea8893b99fae6c86efe19a99a25e6c6a275e877e9b679c2bfefcd391db45d","last_reissued_at":"2026-07-05T06:07:32.731795Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:07:32.731795Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Correlation Information Bottleneck: Towards Adapting Pretrained Multimodal Models for Robust Visual Question Answering","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jingjing Jiang, Nanning Zheng, Ziyi Liu","submitted_at":"2022-09-14T22:04:10Z","abstract_excerpt":"Benefiting from large-scale pretrained vision language models (VLMs), the performance of visual question answering (VQA) has approached human oracles. However, finetuning such models on limited data often suffers from overfitting and poor generalization issues, leading to a lack of model robustness. In this paper, we aim to improve input robustness from an information bottleneck perspective when adapting pretrained VLMs to the downstream VQA task. Input robustness refers to the ability of models to defend against visual and linguistic input variations, as well as shortcut learning involved in "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2209.06954","kind":"arxiv","version":3},"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/2209.06954/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":"2209.06954","created_at":"2026-07-05T06:07:32.731849+00:00"},{"alias_kind":"arxiv_version","alias_value":"2209.06954v3","created_at":"2026-07-05T06:07:32.731849+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2209.06954","created_at":"2026-07-05T06:07:32.731849+00:00"},{"alias_kind":"pith_short_12","alias_value":"6NHKRCJ3TH5O","created_at":"2026-07-05T06:07:32.731849+00:00"},{"alias_kind":"pith_short_16","alias_value":"6NHKRCJ3TH5ONSDO","created_at":"2026-07-05T06:07:32.731849+00:00"},{"alias_kind":"pith_short_8","alias_value":"6NHKRCJ3","created_at":"2026-07-05T06:07:32.731849+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.22979","citing_title":"Probabilistic Prototype Calibration of Vision-Language Models for Generalized Few-shot Semantic Segmentation","ref_index":16,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6NHKRCJ3TH5ONSDO7YM2TGRF43","json":"https://pith.science/pith/6NHKRCJ3TH5ONSDO7YM2TGRF43.json","graph_json":"https://pith.science/api/pith-number/6NHKRCJ3TH5ONSDO7YM2TGRF43/graph.json","events_json":"https://pith.science/api/pith-number/6NHKRCJ3TH5ONSDO7YM2TGRF43/events.json","paper":"https://pith.science/paper/6NHKRCJ3"},"agent_actions":{"view_html":"https://pith.science/pith/6NHKRCJ3TH5ONSDO7YM2TGRF43","download_json":"https://pith.science/pith/6NHKRCJ3TH5ONSDO7YM2TGRF43.json","view_paper":"https://pith.science/paper/6NHKRCJ3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2209.06954&json=true","fetch_graph":"https://pith.science/api/pith-number/6NHKRCJ3TH5ONSDO7YM2TGRF43/graph.json","fetch_events":"https://pith.science/api/pith-number/6NHKRCJ3TH5ONSDO7YM2TGRF43/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6NHKRCJ3TH5ONSDO7YM2TGRF43/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6NHKRCJ3TH5ONSDO7YM2TGRF43/action/storage_attestation","attest_author":"https://pith.science/pith/6NHKRCJ3TH5ONSDO7YM2TGRF43/action/author_attestation","sign_citation":"https://pith.science/pith/6NHKRCJ3TH5ONSDO7YM2TGRF43/action/citation_signature","submit_replication":"https://pith.science/pith/6NHKRCJ3TH5ONSDO7YM2TGRF43/action/replication_record"}},"created_at":"2026-07-05T06:07:32.731849+00:00","updated_at":"2026-07-05T06:07:32.731849+00:00"}