{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:6NOJPEYCZJSAOTA7PUXEYWKD4Y","short_pith_number":"pith:6NOJPEYC","schema_version":"1.0","canonical_sha256":"f35c979302ca64074c1f7d2e4c5943e61f325ee92c660548f5628702e75fdeae","source":{"kind":"arxiv","id":"2103.11275","version":3},"attestation_state":"computed","paper":{"title":"Self-supervised Representation Learning with Relative Predictive Coding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.IT","math.IT"],"primary_cat":"cs.LG","authors_text":"Han Zhao, Louis-Philippe Morency, Martin Q. Ma, Muqiao Yang, Ruslan Salakhutdinov, Yao-Hung Hubert Tsai","submitted_at":"2021-03-21T01:04:24Z","abstract_excerpt":"This paper introduces Relative Predictive Coding (RPC), a new contrastive representation learning objective that maintains a good balance among training stability, minibatch size sensitivity, and downstream task performance. The key to the success of RPC is two-fold. First, RPC introduces the relative parameters to regularize the objective for boundedness and low variance. Second, RPC contains no logarithm and exponential score functions, which are the main cause of training instability in prior contrastive objectives. We empirically verify the effectiveness of RPC on benchmark vision and spee"},"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":"2103.11275","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-03-21T01:04:24Z","cross_cats_sorted":["cs.IT","math.IT"],"title_canon_sha256":"ad64a094e10952254d5f9be60a860f565ada1b3617ab836af9cd72bbfb326949","abstract_canon_sha256":"f9332f236fbb7c0222e923e681c9651287bc881525760bfba68cca7fda20c664"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:31:23.223029Z","signature_b64":"cZiYqb5j1epgdAl0wEjY/KBAzcXVSDTjo3FC71ndrTnm+uRAQ1JcTqv490EFB/Ta4ZlC+3JkRK8s/Y2M4sBZCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f35c979302ca64074c1f7d2e4c5943e61f325ee92c660548f5628702e75fdeae","last_reissued_at":"2026-07-05T02:31:23.222540Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:31:23.222540Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Self-supervised Representation Learning with Relative Predictive Coding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.IT","math.IT"],"primary_cat":"cs.LG","authors_text":"Han Zhao, Louis-Philippe Morency, Martin Q. Ma, Muqiao Yang, Ruslan Salakhutdinov, Yao-Hung Hubert Tsai","submitted_at":"2021-03-21T01:04:24Z","abstract_excerpt":"This paper introduces Relative Predictive Coding (RPC), a new contrastive representation learning objective that maintains a good balance among training stability, minibatch size sensitivity, and downstream task performance. The key to the success of RPC is two-fold. First, RPC introduces the relative parameters to regularize the objective for boundedness and low variance. Second, RPC contains no logarithm and exponential score functions, which are the main cause of training instability in prior contrastive objectives. We empirically verify the effectiveness of RPC on benchmark vision and spee"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2103.11275","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/2103.11275/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":"2103.11275","created_at":"2026-07-05T02:31:23.222597+00:00"},{"alias_kind":"arxiv_version","alias_value":"2103.11275v3","created_at":"2026-07-05T02:31:23.222597+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2103.11275","created_at":"2026-07-05T02:31:23.222597+00:00"},{"alias_kind":"pith_short_12","alias_value":"6NOJPEYCZJSA","created_at":"2026-07-05T02:31:23.222597+00:00"},{"alias_kind":"pith_short_16","alias_value":"6NOJPEYCZJSAOTA7","created_at":"2026-07-05T02:31:23.222597+00:00"},{"alias_kind":"pith_short_8","alias_value":"6NOJPEYC","created_at":"2026-07-05T02:31:23.222597+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.01334","citing_title":"HOLA: Holistic Multi-Modal Alignment for Open-Set 3D Recognition","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21130","citing_title":"Self-Predictive Representation for Autonomous UAV Object-Goal Navigation","ref_index":41,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6NOJPEYCZJSAOTA7PUXEYWKD4Y","json":"https://pith.science/pith/6NOJPEYCZJSAOTA7PUXEYWKD4Y.json","graph_json":"https://pith.science/api/pith-number/6NOJPEYCZJSAOTA7PUXEYWKD4Y/graph.json","events_json":"https://pith.science/api/pith-number/6NOJPEYCZJSAOTA7PUXEYWKD4Y/events.json","paper":"https://pith.science/paper/6NOJPEYC"},"agent_actions":{"view_html":"https://pith.science/pith/6NOJPEYCZJSAOTA7PUXEYWKD4Y","download_json":"https://pith.science/pith/6NOJPEYCZJSAOTA7PUXEYWKD4Y.json","view_paper":"https://pith.science/paper/6NOJPEYC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2103.11275&json=true","fetch_graph":"https://pith.science/api/pith-number/6NOJPEYCZJSAOTA7PUXEYWKD4Y/graph.json","fetch_events":"https://pith.science/api/pith-number/6NOJPEYCZJSAOTA7PUXEYWKD4Y/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6NOJPEYCZJSAOTA7PUXEYWKD4Y/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6NOJPEYCZJSAOTA7PUXEYWKD4Y/action/storage_attestation","attest_author":"https://pith.science/pith/6NOJPEYCZJSAOTA7PUXEYWKD4Y/action/author_attestation","sign_citation":"https://pith.science/pith/6NOJPEYCZJSAOTA7PUXEYWKD4Y/action/citation_signature","submit_replication":"https://pith.science/pith/6NOJPEYCZJSAOTA7PUXEYWKD4Y/action/replication_record"}},"created_at":"2026-07-05T02:31:23.222597+00:00","updated_at":"2026-07-05T02:31:23.222597+00:00"}