{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:SKZPZCCUSY5JCQ7A7UXQIXYXPJ","short_pith_number":"pith:SKZPZCCU","schema_version":"1.0","canonical_sha256":"92b2fc8854963a9143e0fd2f045f177a57beffe01b717e46bbc7b845fb31de7a","source":{"kind":"arxiv","id":"2105.04522","version":4},"attestation_state":"computed","paper":{"title":"Generalized Jensen-Shannon Divergence Loss for Learning with Noisy Labels","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Erik Englesson, Hossein Azizpour","submitted_at":"2021-05-10T17:19:38Z","abstract_excerpt":"Prior works have found it beneficial to combine provably noise-robust loss functions e.g., mean absolute error (MAE) with standard categorical loss function e.g. cross entropy (CE) to improve their learnability. Here, we propose to use Jensen-Shannon divergence as a noise-robust loss function and show that it interestingly interpolate between CE and MAE with a controllable mixing parameter. Furthermore, we make a crucial observation that CE exhibit lower consistency around noisy data points. Based on this observation, we adopt a generalized version of the Jensen-Shannon divergence for multiple"},"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":"2105.04522","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-05-10T17:19:38Z","cross_cats_sorted":["cs.CV","stat.ML"],"title_canon_sha256":"29ae934702a33ef614aa1fa81290929ac8b307f72123444c558d31ce5127df08","abstract_canon_sha256":"5e39f96babce6268f031b4e58681b53b377a57349d69eb9241e1abe279b28bcf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:27:02.442482Z","signature_b64":"aBrzGEmKCH5uEima7y+EGL/aA7lJ5nJdvcQsTBpF6EkKrX296QeXpQp5VfisT3/g0/DaJUsXRyiCc3QwOHfsDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"92b2fc8854963a9143e0fd2f045f177a57beffe01b717e46bbc7b845fb31de7a","last_reissued_at":"2026-07-05T03:27:02.442032Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:27:02.442032Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Generalized Jensen-Shannon Divergence Loss for Learning with Noisy Labels","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Erik Englesson, Hossein Azizpour","submitted_at":"2021-05-10T17:19:38Z","abstract_excerpt":"Prior works have found it beneficial to combine provably noise-robust loss functions e.g., mean absolute error (MAE) with standard categorical loss function e.g. cross entropy (CE) to improve their learnability. Here, we propose to use Jensen-Shannon divergence as a noise-robust loss function and show that it interestingly interpolate between CE and MAE with a controllable mixing parameter. Furthermore, we make a crucial observation that CE exhibit lower consistency around noisy data points. Based on this observation, we adopt a generalized version of the Jensen-Shannon divergence for multiple"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2105.04522","kind":"arxiv","version":4},"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/2105.04522/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":"2105.04522","created_at":"2026-07-05T03:27:02.442092+00:00"},{"alias_kind":"arxiv_version","alias_value":"2105.04522v4","created_at":"2026-07-05T03:27:02.442092+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2105.04522","created_at":"2026-07-05T03:27:02.442092+00:00"},{"alias_kind":"pith_short_12","alias_value":"SKZPZCCUSY5J","created_at":"2026-07-05T03:27:02.442092+00:00"},{"alias_kind":"pith_short_16","alias_value":"SKZPZCCUSY5JCQ7A","created_at":"2026-07-05T03:27:02.442092+00:00"},{"alias_kind":"pith_short_8","alias_value":"SKZPZCCU","created_at":"2026-07-05T03:27:02.442092+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.08339","citing_title":"SHeRL-FL: When Representation Learning Meets Split Learning in Hierarchical Federated Learning","ref_index":37,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SKZPZCCUSY5JCQ7A7UXQIXYXPJ","json":"https://pith.science/pith/SKZPZCCUSY5JCQ7A7UXQIXYXPJ.json","graph_json":"https://pith.science/api/pith-number/SKZPZCCUSY5JCQ7A7UXQIXYXPJ/graph.json","events_json":"https://pith.science/api/pith-number/SKZPZCCUSY5JCQ7A7UXQIXYXPJ/events.json","paper":"https://pith.science/paper/SKZPZCCU"},"agent_actions":{"view_html":"https://pith.science/pith/SKZPZCCUSY5JCQ7A7UXQIXYXPJ","download_json":"https://pith.science/pith/SKZPZCCUSY5JCQ7A7UXQIXYXPJ.json","view_paper":"https://pith.science/paper/SKZPZCCU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2105.04522&json=true","fetch_graph":"https://pith.science/api/pith-number/SKZPZCCUSY5JCQ7A7UXQIXYXPJ/graph.json","fetch_events":"https://pith.science/api/pith-number/SKZPZCCUSY5JCQ7A7UXQIXYXPJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SKZPZCCUSY5JCQ7A7UXQIXYXPJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SKZPZCCUSY5JCQ7A7UXQIXYXPJ/action/storage_attestation","attest_author":"https://pith.science/pith/SKZPZCCUSY5JCQ7A7UXQIXYXPJ/action/author_attestation","sign_citation":"https://pith.science/pith/SKZPZCCUSY5JCQ7A7UXQIXYXPJ/action/citation_signature","submit_replication":"https://pith.science/pith/SKZPZCCUSY5JCQ7A7UXQIXYXPJ/action/replication_record"}},"created_at":"2026-07-05T03:27:02.442092+00:00","updated_at":"2026-07-05T03:27:02.442092+00:00"}