{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:XG3TOQVQWYK2DGHUC6SUGRHOFU","short_pith_number":"pith:XG3TOQVQ","schema_version":"1.0","canonical_sha256":"b9b73742b0b615a198f417a54344ee2d0a8ade8b7ba04272ec03ae53aea78181","source":{"kind":"arxiv","id":"2505.04960","version":1},"attestation_state":"computed","paper":{"title":"Learning Item Representations Directly from Multimodal Features for Effective Recommendation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.MM"],"primary_cat":"cs.IR","authors_text":"Dusit Niyato, Xiaoxiong Zhang, Xin Zhou, Zhiqi Shen","submitted_at":"2025-05-08T05:42:22Z","abstract_excerpt":"Conventional multimodal recommender systems predominantly leverage Bayesian Personalized Ranking (BPR) optimization to learn item representations by amalgamating item identity (ID) embeddings with multimodal features. Nevertheless, our empirical and theoretical findings unequivocally demonstrate a pronounced optimization gradient bias in favor of acquiring representations from multimodal features over item ID embeddings. As a consequence, item ID embeddings frequently exhibit suboptimal characteristics despite the convergence of multimodal feature parameters. Given the rich informational conte"},"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":"2505.04960","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2025-05-08T05:42:22Z","cross_cats_sorted":["cs.MM"],"title_canon_sha256":"2307578a19021d3fd5686c5b9333be14a59cf8ef741fe6853eb3f94a4c8ba6d8","abstract_canon_sha256":"f8924648dce271f00b379c70d7ad29b50db1f087e6e8981a073fe42a218ca5bc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:00:16.681256Z","signature_b64":"uvTAvI/Q8GjLgt+4WS167mGiURLOpp4Vf6seRf1EUDdUeYZieg+kHei7VEXALCEEBw61R3YrTdS9fs5nkeBjAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b9b73742b0b615a198f417a54344ee2d0a8ade8b7ba04272ec03ae53aea78181","last_reissued_at":"2026-07-05T11:00:16.680724Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:00:16.680724Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Learning Item Representations Directly from Multimodal Features for Effective Recommendation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.MM"],"primary_cat":"cs.IR","authors_text":"Dusit Niyato, Xiaoxiong Zhang, Xin Zhou, Zhiqi Shen","submitted_at":"2025-05-08T05:42:22Z","abstract_excerpt":"Conventional multimodal recommender systems predominantly leverage Bayesian Personalized Ranking (BPR) optimization to learn item representations by amalgamating item identity (ID) embeddings with multimodal features. Nevertheless, our empirical and theoretical findings unequivocally demonstrate a pronounced optimization gradient bias in favor of acquiring representations from multimodal features over item ID embeddings. As a consequence, item ID embeddings frequently exhibit suboptimal characteristics despite the convergence of multimodal feature parameters. Given the rich informational conte"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.04960","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/2505.04960/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":"2505.04960","created_at":"2026-07-05T11:00:16.680790+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.04960v1","created_at":"2026-07-05T11:00:16.680790+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.04960","created_at":"2026-07-05T11:00:16.680790+00:00"},{"alias_kind":"pith_short_12","alias_value":"XG3TOQVQWYK2","created_at":"2026-07-05T11:00:16.680790+00:00"},{"alias_kind":"pith_short_16","alias_value":"XG3TOQVQWYK2DGHU","created_at":"2026-07-05T11:00:16.680790+00:00"},{"alias_kind":"pith_short_8","alias_value":"XG3TOQVQ","created_at":"2026-07-05T11:00:16.680790+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.18044","citing_title":"Modality-Aware Identity Construction and Counterfactual Structure Learning for ID-Free Multimodal Recommendation","ref_index":17,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XG3TOQVQWYK2DGHUC6SUGRHOFU","json":"https://pith.science/pith/XG3TOQVQWYK2DGHUC6SUGRHOFU.json","graph_json":"https://pith.science/api/pith-number/XG3TOQVQWYK2DGHUC6SUGRHOFU/graph.json","events_json":"https://pith.science/api/pith-number/XG3TOQVQWYK2DGHUC6SUGRHOFU/events.json","paper":"https://pith.science/paper/XG3TOQVQ"},"agent_actions":{"view_html":"https://pith.science/pith/XG3TOQVQWYK2DGHUC6SUGRHOFU","download_json":"https://pith.science/pith/XG3TOQVQWYK2DGHUC6SUGRHOFU.json","view_paper":"https://pith.science/paper/XG3TOQVQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.04960&json=true","fetch_graph":"https://pith.science/api/pith-number/XG3TOQVQWYK2DGHUC6SUGRHOFU/graph.json","fetch_events":"https://pith.science/api/pith-number/XG3TOQVQWYK2DGHUC6SUGRHOFU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XG3TOQVQWYK2DGHUC6SUGRHOFU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XG3TOQVQWYK2DGHUC6SUGRHOFU/action/storage_attestation","attest_author":"https://pith.science/pith/XG3TOQVQWYK2DGHUC6SUGRHOFU/action/author_attestation","sign_citation":"https://pith.science/pith/XG3TOQVQWYK2DGHUC6SUGRHOFU/action/citation_signature","submit_replication":"https://pith.science/pith/XG3TOQVQWYK2DGHUC6SUGRHOFU/action/replication_record"}},"created_at":"2026-07-05T11:00:16.680790+00:00","updated_at":"2026-07-05T11:00:16.680790+00:00"}