{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:GPHXZZDEGZKNG7DHFAGG5RAYPH","short_pith_number":"pith:GPHXZZDE","schema_version":"1.0","canonical_sha256":"33cf7ce4643654d37c67280c6ec41879e703376cfab5c7b3fc6ad4566b357303","source":{"kind":"arxiv","id":"1905.12681","version":5},"attestation_state":"computed","paper":{"title":"What Makes Training Multi-Modal Classification Networks Hard?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Du Tran, Matt Feiszli, Weiyao Wang","submitted_at":"2019-05-29T19:10:06Z","abstract_excerpt":"Consider end-to-end training of a multi-modal vs. a single-modal network on a task with multiple input modalities: the multi-modal network receives more information, so it should match or outperform its single-modal counterpart. In our experiments, however, we observe the opposite: the best single-modal network always outperforms the multi-modal network. This observation is consistent across different combinations of modalities and on different tasks and benchmarks.\n  This paper identifies two main causes for this performance drop: first, multi-modal networks are often prone to overfitting due"},"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":"1905.12681","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2019-05-29T19:10:06Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"7e4e5e5299d5f944046751a4b972e1b0831fae2fcbfa9fe9fee1701c88b7e959","abstract_canon_sha256":"45dde9b4e8af14ee31aed2bf41e2daf6da186360aa73df1aa05c9d65658b2edf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:52:26.665279Z","signature_b64":"WK/TFqXiBFTYKL4Vq7bMe26zOX2nXNZd3j2UDfT2wA39q37cWxtqMNA2ThC9xUCkb00NqsQdTU7diRUU9bavBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"33cf7ce4643654d37c67280c6ec41879e703376cfab5c7b3fc6ad4566b357303","last_reissued_at":"2026-07-05T00:52:26.664944Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:52:26.664944Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"What Makes Training Multi-Modal Classification Networks Hard?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Du Tran, Matt Feiszli, Weiyao Wang","submitted_at":"2019-05-29T19:10:06Z","abstract_excerpt":"Consider end-to-end training of a multi-modal vs. a single-modal network on a task with multiple input modalities: the multi-modal network receives more information, so it should match or outperform its single-modal counterpart. In our experiments, however, we observe the opposite: the best single-modal network always outperforms the multi-modal network. This observation is consistent across different combinations of modalities and on different tasks and benchmarks.\n  This paper identifies two main causes for this performance drop: first, multi-modal networks are often prone to overfitting due"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1905.12681","kind":"arxiv","version":5},"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/1905.12681/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":"1905.12681","created_at":"2026-07-05T00:52:26.665000+00:00"},{"alias_kind":"arxiv_version","alias_value":"1905.12681v5","created_at":"2026-07-05T00:52:26.665000+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1905.12681","created_at":"2026-07-05T00:52:26.665000+00:00"},{"alias_kind":"pith_short_12","alias_value":"GPHXZZDEGZKN","created_at":"2026-07-05T00:52:26.665000+00:00"},{"alias_kind":"pith_short_16","alias_value":"GPHXZZDEGZKNG7DH","created_at":"2026-07-05T00:52:26.665000+00:00"},{"alias_kind":"pith_short_8","alias_value":"GPHXZZDE","created_at":"2026-07-05T00:52:26.665000+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.32023","citing_title":"FLORA: A deep learning approach to predict forest attributes from heterogeneous LiDAR data","ref_index":22,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GPHXZZDEGZKNG7DHFAGG5RAYPH","json":"https://pith.science/pith/GPHXZZDEGZKNG7DHFAGG5RAYPH.json","graph_json":"https://pith.science/api/pith-number/GPHXZZDEGZKNG7DHFAGG5RAYPH/graph.json","events_json":"https://pith.science/api/pith-number/GPHXZZDEGZKNG7DHFAGG5RAYPH/events.json","paper":"https://pith.science/paper/GPHXZZDE"},"agent_actions":{"view_html":"https://pith.science/pith/GPHXZZDEGZKNG7DHFAGG5RAYPH","download_json":"https://pith.science/pith/GPHXZZDEGZKNG7DHFAGG5RAYPH.json","view_paper":"https://pith.science/paper/GPHXZZDE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1905.12681&json=true","fetch_graph":"https://pith.science/api/pith-number/GPHXZZDEGZKNG7DHFAGG5RAYPH/graph.json","fetch_events":"https://pith.science/api/pith-number/GPHXZZDEGZKNG7DHFAGG5RAYPH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GPHXZZDEGZKNG7DHFAGG5RAYPH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GPHXZZDEGZKNG7DHFAGG5RAYPH/action/storage_attestation","attest_author":"https://pith.science/pith/GPHXZZDEGZKNG7DHFAGG5RAYPH/action/author_attestation","sign_citation":"https://pith.science/pith/GPHXZZDEGZKNG7DHFAGG5RAYPH/action/citation_signature","submit_replication":"https://pith.science/pith/GPHXZZDEGZKNG7DHFAGG5RAYPH/action/replication_record"}},"created_at":"2026-07-05T00:52:26.665000+00:00","updated_at":"2026-07-05T00:52:26.665000+00:00"}