{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:OALDKVU3HHEAZ3NA7GNIDILJPC","short_pith_number":"pith:OALDKVU3","schema_version":"1.0","canonical_sha256":"701635569b39c80ceda0f99a81a1697895f2bd68451fe9c7a8328ab8ec224031","source":{"kind":"arxiv","id":"2109.04290","version":3},"attestation_state":"computed","paper":{"title":"Improving Video-Text Retrieval by Multi-Stream Corpus Alignment and Dual Softmax Loss","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Dong Shen, Fan Yang, Hezheng Lin, Xiangyu Wu, Xing Cheng","submitted_at":"2021-09-09T14:10:43Z","abstract_excerpt":"Employing large-scale pre-trained model CLIP to conduct video-text retrieval task (VTR) has become a new trend, which exceeds previous VTR methods. Though, due to the heterogeneity of structures and contents between video and text, previous CLIP-based models are prone to overfitting in the training phase, resulting in relatively poor retrieval performance. In this paper, we propose a multi-stream Corpus Alignment network with single gate Mixture-of-Experts (CAMoE) and a novel Dual Softmax Loss (DSL) to solve the two heterogeneity. The CAMoE employs Mixture-of-Experts (MoE) to extract multi-per"},"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":"2109.04290","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2021-09-09T14:10:43Z","cross_cats_sorted":[],"title_canon_sha256":"9543a88a30f0883a3da61b17bba1be244a4527eb03f8c4d711376ebd945ec02e","abstract_canon_sha256":"3c98e9cde690a3338a19e971fbbaed754d49b7c9540be94ac6887ac6439fc149"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:33:49.665246Z","signature_b64":"PYs+crR/iETqA5EBtT5eX1zSU07UKkzlYKQGsVpuOjzd2fvGuXWausIPSDjhscuDS4zpbL3YdijDGLOx0RHIAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"701635569b39c80ceda0f99a81a1697895f2bd68451fe9c7a8328ab8ec224031","last_reissued_at":"2026-07-05T03:33:49.664690Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:33:49.664690Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improving Video-Text Retrieval by Multi-Stream Corpus Alignment and Dual Softmax Loss","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Dong Shen, Fan Yang, Hezheng Lin, Xiangyu Wu, Xing Cheng","submitted_at":"2021-09-09T14:10:43Z","abstract_excerpt":"Employing large-scale pre-trained model CLIP to conduct video-text retrieval task (VTR) has become a new trend, which exceeds previous VTR methods. Though, due to the heterogeneity of structures and contents between video and text, previous CLIP-based models are prone to overfitting in the training phase, resulting in relatively poor retrieval performance. In this paper, we propose a multi-stream Corpus Alignment network with single gate Mixture-of-Experts (CAMoE) and a novel Dual Softmax Loss (DSL) to solve the two heterogeneity. The CAMoE employs Mixture-of-Experts (MoE) to extract multi-per"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.04290","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/2109.04290/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":"2109.04290","created_at":"2026-07-05T03:33:49.664772+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.04290v3","created_at":"2026-07-05T03:33:49.664772+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.04290","created_at":"2026-07-05T03:33:49.664772+00:00"},{"alias_kind":"pith_short_12","alias_value":"OALDKVU3HHEA","created_at":"2026-07-05T03:33:49.664772+00:00"},{"alias_kind":"pith_short_16","alias_value":"OALDKVU3HHEAZ3NA","created_at":"2026-07-05T03:33:49.664772+00:00"},{"alias_kind":"pith_short_8","alias_value":"OALDKVU3","created_at":"2026-07-05T03:33:49.664772+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2607.00858","citing_title":"MoVA: Learning Asymmetric Dual Projections for Modular Long Video-Text Alignment","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2503.13821","citing_title":"Stitch-a-Demo: Video Demonstrations from Multistep Descriptions","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2212.03191","citing_title":"InternVideo: General Video Foundation Models via Generative and Discriminative Learning","ref_index":91,"is_internal_anchor":false},{"citing_arxiv_id":"2204.00598","citing_title":"Socratic Models: Composing Zero-Shot Multimodal Reasoning with Language","ref_index":73,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00826","citing_title":"Understanding the Performance Plateau in Text-to-Video Retrieval: A Comprehensive Empirical and Linguistic Analysis","ref_index":59,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OALDKVU3HHEAZ3NA7GNIDILJPC","json":"https://pith.science/pith/OALDKVU3HHEAZ3NA7GNIDILJPC.json","graph_json":"https://pith.science/api/pith-number/OALDKVU3HHEAZ3NA7GNIDILJPC/graph.json","events_json":"https://pith.science/api/pith-number/OALDKVU3HHEAZ3NA7GNIDILJPC/events.json","paper":"https://pith.science/paper/OALDKVU3"},"agent_actions":{"view_html":"https://pith.science/pith/OALDKVU3HHEAZ3NA7GNIDILJPC","download_json":"https://pith.science/pith/OALDKVU3HHEAZ3NA7GNIDILJPC.json","view_paper":"https://pith.science/paper/OALDKVU3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.04290&json=true","fetch_graph":"https://pith.science/api/pith-number/OALDKVU3HHEAZ3NA7GNIDILJPC/graph.json","fetch_events":"https://pith.science/api/pith-number/OALDKVU3HHEAZ3NA7GNIDILJPC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OALDKVU3HHEAZ3NA7GNIDILJPC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OALDKVU3HHEAZ3NA7GNIDILJPC/action/storage_attestation","attest_author":"https://pith.science/pith/OALDKVU3HHEAZ3NA7GNIDILJPC/action/author_attestation","sign_citation":"https://pith.science/pith/OALDKVU3HHEAZ3NA7GNIDILJPC/action/citation_signature","submit_replication":"https://pith.science/pith/OALDKVU3HHEAZ3NA7GNIDILJPC/action/replication_record"}},"created_at":"2026-07-05T03:33:49.664772+00:00","updated_at":"2026-07-05T03:33:49.664772+00:00"}