{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:2UJEODLOTYSLIYCV6VIHHUESRZ","short_pith_number":"pith:2UJEODLO","schema_version":"1.0","canonical_sha256":"d512470d6e9e24b46055f55073d0928e66e6ac29768b1dade8963b0d97b7cb2d","source":{"kind":"arxiv","id":"2503.14895","version":1},"attestation_state":"computed","paper":{"title":"Mitigating Object Hallucinations in MLLMs via Multi-Frequency Perturbations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CV","authors_text":"Guodong Zheng, Jiajun Sun, Qi Zhang, Shuo Li, Tao Gui, Tao Ji, Wenming Tan, Xiaoran Fan, Xuanjing Huang, Yi Lu, Yujiong Shen, Yuming Yang, Zhiheng Xi","submitted_at":"2025-03-19T04:39:45Z","abstract_excerpt":"Recently, multimodal large language models (MLLMs) have demonstrated remarkable performance in visual-language tasks. However, the authenticity of the responses generated by MLLMs is often compromised by object hallucinations. We identify that a key cause of these hallucinations is the model's over-susceptibility to specific image frequency features in detecting objects. In this paper, we introduce Multi-Frequency Perturbations (MFP), a simple, cost-effective, and pluggable method that leverages both low-frequency and high-frequency features of images to perturb visual feature representations "},"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":"2503.14895","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-03-19T04:39:45Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"f24358572b67d7fbab6ad6373c5c6ec95e495139df69dc7299ec9aaf0ca0a460","abstract_canon_sha256":"4c9b6346b8af07dcda2be0df243f1b41d74b24cde262652983eb9286074b98e4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:34:26.797409Z","signature_b64":"pFYyg1DvQwRB8xSrjogS2rwOwA919NV0eO0U5WaBjYpsBxwiICDy/ez382eVc2Fe99XJDA/5WImuav52ySTdCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d512470d6e9e24b46055f55073d0928e66e6ac29768b1dade8963b0d97b7cb2d","last_reissued_at":"2026-07-05T10:34:26.796107Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:34:26.796107Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mitigating Object Hallucinations in MLLMs via Multi-Frequency Perturbations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CV","authors_text":"Guodong Zheng, Jiajun Sun, Qi Zhang, Shuo Li, Tao Gui, Tao Ji, Wenming Tan, Xiaoran Fan, Xuanjing Huang, Yi Lu, Yujiong Shen, Yuming Yang, Zhiheng Xi","submitted_at":"2025-03-19T04:39:45Z","abstract_excerpt":"Recently, multimodal large language models (MLLMs) have demonstrated remarkable performance in visual-language tasks. However, the authenticity of the responses generated by MLLMs is often compromised by object hallucinations. We identify that a key cause of these hallucinations is the model's over-susceptibility to specific image frequency features in detecting objects. In this paper, we introduce Multi-Frequency Perturbations (MFP), a simple, cost-effective, and pluggable method that leverages both low-frequency and high-frequency features of images to perturb visual feature representations "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.14895","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/2503.14895/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":"2503.14895","created_at":"2026-07-05T10:34:26.796287+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.14895v1","created_at":"2026-07-05T10:34:26.796287+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.14895","created_at":"2026-07-05T10:34:26.796287+00:00"},{"alias_kind":"pith_short_12","alias_value":"2UJEODLOTYSL","created_at":"2026-07-05T10:34:26.796287+00:00"},{"alias_kind":"pith_short_16","alias_value":"2UJEODLOTYSLIYCV","created_at":"2026-07-05T10:34:26.796287+00:00"},{"alias_kind":"pith_short_8","alias_value":"2UJEODLO","created_at":"2026-07-05T10:34:26.796287+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.29805","citing_title":"Clearer Sight, Fewer Lies: Oriented Pickup Preference Optimization for Multimodal Hallucination Mitigation","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29805","citing_title":"Clearer Sight, Fewer Lies: Oriented Pickup Preference Optimization for Multimodal Hallucination Mitigation","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12424","citing_title":"Decoding by Perturbation: Mitigating MLLM Hallucinations via Dynamic Textual Perturbation","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2604.04500","citing_title":"Saliency-R1: Enforcing Interpretable and Faithful Vision-language Reasoning via Saliency-map Alignment Reward","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22822","citing_title":"DO-Bench: An Attributable Benchmark for Diagnosing Object Hallucination in Vision-Language Models","ref_index":21,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2UJEODLOTYSLIYCV6VIHHUESRZ","json":"https://pith.science/pith/2UJEODLOTYSLIYCV6VIHHUESRZ.json","graph_json":"https://pith.science/api/pith-number/2UJEODLOTYSLIYCV6VIHHUESRZ/graph.json","events_json":"https://pith.science/api/pith-number/2UJEODLOTYSLIYCV6VIHHUESRZ/events.json","paper":"https://pith.science/paper/2UJEODLO"},"agent_actions":{"view_html":"https://pith.science/pith/2UJEODLOTYSLIYCV6VIHHUESRZ","download_json":"https://pith.science/pith/2UJEODLOTYSLIYCV6VIHHUESRZ.json","view_paper":"https://pith.science/paper/2UJEODLO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.14895&json=true","fetch_graph":"https://pith.science/api/pith-number/2UJEODLOTYSLIYCV6VIHHUESRZ/graph.json","fetch_events":"https://pith.science/api/pith-number/2UJEODLOTYSLIYCV6VIHHUESRZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2UJEODLOTYSLIYCV6VIHHUESRZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2UJEODLOTYSLIYCV6VIHHUESRZ/action/storage_attestation","attest_author":"https://pith.science/pith/2UJEODLOTYSLIYCV6VIHHUESRZ/action/author_attestation","sign_citation":"https://pith.science/pith/2UJEODLOTYSLIYCV6VIHHUESRZ/action/citation_signature","submit_replication":"https://pith.science/pith/2UJEODLOTYSLIYCV6VIHHUESRZ/action/replication_record"}},"created_at":"2026-07-05T10:34:26.796287+00:00","updated_at":"2026-07-05T10:34:26.796287+00:00"}