{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:IHRYIP3CKHOBOQBVSPAYEFWQP4","short_pith_number":"pith:IHRYIP3C","schema_version":"1.0","canonical_sha256":"41e3843f6251dc17403593c18216d07f1d4c9fc4d5fd4f016f4b9a0479025bb1","source":{"kind":"arxiv","id":"2501.12191","version":2},"attestation_state":"computed","paper":{"title":"HEM: a margin-based loss for visual categorisation tasks","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Heiko H. Sch\\\"utt, Michael W. Spratling","submitted_at":"2025-01-21T14:56:47Z","abstract_excerpt":"Training deep neural networks (DNNs) on classification tasks can be performed with a number of different losses, but cross-entropy (CE) loss is the de-facto standard. Here, we propose an alternative loss, high error margin (HEM), which is a margin based loss modified to improve the training dynamics of neural networks. HEM loss is evaluated extensively using a wide range of DNN architectures and benchmark datasets with all experimental settings and training hyper-parameters taken from the literature, and hence, optimised for CE loss. HEM is found to be more effective than CE loss across a rang"},"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":"2501.12191","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2025-01-21T14:56:47Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"8c26413dd67c198767f5df1e74d91093c6a6a7d9a2d199f566b5280c0773252c","abstract_canon_sha256":"52be187706ec02077bde26162b272f18bc579b51c1de15da5175881e2a3a48d2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T02:13:13.766703Z","signature_b64":"kfxbMGkIIxgmpYKYp53AP98nl8xhB4hs8/fcGqaa6zw1F8tQJyuZCK96UDIQCIw5lHgCvrb4hOOtDB5fe6SXDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"41e3843f6251dc17403593c18216d07f1d4c9fc4d5fd4f016f4b9a0479025bb1","last_reissued_at":"2026-06-23T02:13:13.766259Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T02:13:13.766259Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HEM: a margin-based loss for visual categorisation tasks","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Heiko H. Sch\\\"utt, Michael W. Spratling","submitted_at":"2025-01-21T14:56:47Z","abstract_excerpt":"Training deep neural networks (DNNs) on classification tasks can be performed with a number of different losses, but cross-entropy (CE) loss is the de-facto standard. Here, we propose an alternative loss, high error margin (HEM), which is a margin based loss modified to improve the training dynamics of neural networks. HEM loss is evaluated extensively using a wide range of DNN architectures and benchmark datasets with all experimental settings and training hyper-parameters taken from the literature, and hence, optimised for CE loss. HEM is found to be more effective than CE loss across a rang"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.12191","kind":"arxiv","version":2},"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/2501.12191/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":"2501.12191","created_at":"2026-06-23T02:13:13.766321+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.12191v2","created_at":"2026-06-23T02:13:13.766321+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.12191","created_at":"2026-06-23T02:13:13.766321+00:00"},{"alias_kind":"pith_short_12","alias_value":"IHRYIP3CKHOB","created_at":"2026-06-23T02:13:13.766321+00:00"},{"alias_kind":"pith_short_16","alias_value":"IHRYIP3CKHOBOQBV","created_at":"2026-06-23T02:13:13.766321+00:00"},{"alias_kind":"pith_short_8","alias_value":"IHRYIP3C","created_at":"2026-06-23T02:13:13.766321+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IHRYIP3CKHOBOQBVSPAYEFWQP4","json":"https://pith.science/pith/IHRYIP3CKHOBOQBVSPAYEFWQP4.json","graph_json":"https://pith.science/api/pith-number/IHRYIP3CKHOBOQBVSPAYEFWQP4/graph.json","events_json":"https://pith.science/api/pith-number/IHRYIP3CKHOBOQBVSPAYEFWQP4/events.json","paper":"https://pith.science/paper/IHRYIP3C"},"agent_actions":{"view_html":"https://pith.science/pith/IHRYIP3CKHOBOQBVSPAYEFWQP4","download_json":"https://pith.science/pith/IHRYIP3CKHOBOQBVSPAYEFWQP4.json","view_paper":"https://pith.science/paper/IHRYIP3C","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.12191&json=true","fetch_graph":"https://pith.science/api/pith-number/IHRYIP3CKHOBOQBVSPAYEFWQP4/graph.json","fetch_events":"https://pith.science/api/pith-number/IHRYIP3CKHOBOQBVSPAYEFWQP4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IHRYIP3CKHOBOQBVSPAYEFWQP4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IHRYIP3CKHOBOQBVSPAYEFWQP4/action/storage_attestation","attest_author":"https://pith.science/pith/IHRYIP3CKHOBOQBVSPAYEFWQP4/action/author_attestation","sign_citation":"https://pith.science/pith/IHRYIP3CKHOBOQBVSPAYEFWQP4/action/citation_signature","submit_replication":"https://pith.science/pith/IHRYIP3CKHOBOQBVSPAYEFWQP4/action/replication_record"}},"created_at":"2026-06-23T02:13:13.766321+00:00","updated_at":"2026-06-23T02:13:13.766321+00:00"}