{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:RV5VLLILUVVKRIPJEYICPAKEEC","short_pith_number":"pith:RV5VLLIL","schema_version":"1.0","canonical_sha256":"8d7b55ad0ba56aa8a1e926102781442098693c46f9b36df4b9f6aa3352a63bdf","source":{"kind":"arxiv","id":"2203.15332","version":1},"attestation_state":"computed","paper":{"title":"Balanced Multimodal Learning via On-the-fly Gradient Modulation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Andong Deng, Di Hu, Dong Wang, Xiaokang Peng, Yake Wei","submitted_at":"2022-03-29T08:26:38Z","abstract_excerpt":"Multimodal learning helps to comprehensively understand the world, by integrating different senses. Accordingly, multiple input modalities are expected to boost model performance, but we actually find that they are not fully exploited even when the multimodal model outperforms its uni-modal counterpart. Specifically, in this paper we point out that existing multimodal discriminative models, in which uniform objective is designed for all modalities, could remain under-optimized uni-modal representations, caused by another dominated modality in some scenarios, e.g., sound in blowing wind event, "},"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":"2203.15332","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2022-03-29T08:26:38Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"224283b97089a239cee143268fb8c51eb185000b48b2e6a1048b3cd9367170f6","abstract_canon_sha256":"88f740c976115d25033e4e094986f403df73f46a2ff13123c9795a3298c0e90b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:09:24.185719Z","signature_b64":"Lsv+KUIu72qgMBbpJdNpM7VY9RVw+GIFoSUd9bsY3LgFcanozHbgLdj6ewhJMCiiBed52ozMFDaWr+kH0zQvBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8d7b55ad0ba56aa8a1e926102781442098693c46f9b36df4b9f6aa3352a63bdf","last_reissued_at":"2026-07-05T04:09:24.185311Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:09:24.185311Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Balanced Multimodal Learning via On-the-fly Gradient Modulation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Andong Deng, Di Hu, Dong Wang, Xiaokang Peng, Yake Wei","submitted_at":"2022-03-29T08:26:38Z","abstract_excerpt":"Multimodal learning helps to comprehensively understand the world, by integrating different senses. Accordingly, multiple input modalities are expected to boost model performance, but we actually find that they are not fully exploited even when the multimodal model outperforms its uni-modal counterpart. Specifically, in this paper we point out that existing multimodal discriminative models, in which uniform objective is designed for all modalities, could remain under-optimized uni-modal representations, caused by another dominated modality in some scenarios, e.g., sound in blowing wind event, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.15332","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/2203.15332/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":"2203.15332","created_at":"2026-07-05T04:09:24.185409+00:00"},{"alias_kind":"arxiv_version","alias_value":"2203.15332v1","created_at":"2026-07-05T04:09:24.185409+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.15332","created_at":"2026-07-05T04:09:24.185409+00:00"},{"alias_kind":"pith_short_12","alias_value":"RV5VLLILUVVK","created_at":"2026-07-05T04:09:24.185409+00:00"},{"alias_kind":"pith_short_16","alias_value":"RV5VLLILUVVKRIPJ","created_at":"2026-07-05T04:09:24.185409+00:00"},{"alias_kind":"pith_short_8","alias_value":"RV5VLLIL","created_at":"2026-07-05T04:09:24.185409+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.01856","citing_title":"Boosting Multimodal Federated Learning via Chained Modality Optimization","ref_index":12,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RV5VLLILUVVKRIPJEYICPAKEEC","json":"https://pith.science/pith/RV5VLLILUVVKRIPJEYICPAKEEC.json","graph_json":"https://pith.science/api/pith-number/RV5VLLILUVVKRIPJEYICPAKEEC/graph.json","events_json":"https://pith.science/api/pith-number/RV5VLLILUVVKRIPJEYICPAKEEC/events.json","paper":"https://pith.science/paper/RV5VLLIL"},"agent_actions":{"view_html":"https://pith.science/pith/RV5VLLILUVVKRIPJEYICPAKEEC","download_json":"https://pith.science/pith/RV5VLLILUVVKRIPJEYICPAKEEC.json","view_paper":"https://pith.science/paper/RV5VLLIL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2203.15332&json=true","fetch_graph":"https://pith.science/api/pith-number/RV5VLLILUVVKRIPJEYICPAKEEC/graph.json","fetch_events":"https://pith.science/api/pith-number/RV5VLLILUVVKRIPJEYICPAKEEC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RV5VLLILUVVKRIPJEYICPAKEEC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RV5VLLILUVVKRIPJEYICPAKEEC/action/storage_attestation","attest_author":"https://pith.science/pith/RV5VLLILUVVKRIPJEYICPAKEEC/action/author_attestation","sign_citation":"https://pith.science/pith/RV5VLLILUVVKRIPJEYICPAKEEC/action/citation_signature","submit_replication":"https://pith.science/pith/RV5VLLILUVVKRIPJEYICPAKEEC/action/replication_record"}},"created_at":"2026-07-05T04:09:24.185409+00:00","updated_at":"2026-07-05T04:09:24.185409+00:00"}