{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:V6FO5VMUUFJSHSFRGJMOFIG3ZP","short_pith_number":"pith:V6FO5VMU","schema_version":"1.0","canonical_sha256":"af8aeed594a15323c8b13258e2a0dbcbf282fef13b828e8611a498acd564b864","source":{"kind":"arxiv","id":"2605.10989","version":2},"attestation_state":"computed","paper":{"title":"SURGE: Surrogate Gradient Adaptation in Binary Neural Networks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Binary neural networks train more accurately when a learnable surrogate gradient uses an auxiliary full-precision branch to reduce mismatch.","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Baochang Zhang, Boyu Liu, Canyu Chen, Haoyu Huang, Linlin Yang, Xuhui Liu, Yanjing Li, Yuguang Yang, Zhongqian Fu","submitted_at":"2026-05-09T09:52:38Z","abstract_excerpt":"The training of Binary Neural Networks (BNNs) is fundamentally based on gradient approximation for non-differentiable binarization operations (e.g., sign function). However, prevailing methods including the Straight-Through Estimator (STE) and its improved variants, rely on hand-crafted designs that suffer from gradient mismatch problem and information loss induced by fixed-range gradient clipping. To address this, we propose SURrogate GradiEnt Adaptation (SURGE), a novel learnable gradient compensation framework with theoretical grounding. SURGE mitigates gradient mismatch through auxiliary b"},"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":true,"formal_links_present":false},"canonical_record":{"source":{"id":"2605.10989","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-09T09:52:38Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"1b015046d7a5cfebfacbfb405d8f00f88db4ab80a6a8aef77987b0e9b4c4f95e","abstract_canon_sha256":"2d08372d9a4c7b48dbc14e28644d640bf9ae342b1f3921f992fd13232fc40b07"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:00:42.650843Z","signature_b64":"Jqul8IKxidIPJXRUoHA3Hw2maaxYpI3ZDfLDsXz5dszJRW/JtGRA/c9dHLYmcaMox2KZhcwX2gbNwAvtUqUHDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"af8aeed594a15323c8b13258e2a0dbcbf282fef13b828e8611a498acd564b864","last_reissued_at":"2026-05-20T00:00:42.650092Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:00:42.650092Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SURGE: Surrogate Gradient Adaptation in Binary Neural Networks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Binary neural networks train more accurately when a learnable surrogate gradient uses an auxiliary full-precision branch to reduce mismatch.","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Baochang Zhang, Boyu Liu, Canyu Chen, Haoyu Huang, Linlin Yang, Xuhui Liu, Yanjing Li, Yuguang Yang, Zhongqian Fu","submitted_at":"2026-05-09T09:52:38Z","abstract_excerpt":"The training of Binary Neural Networks (BNNs) is fundamentally based on gradient approximation for non-differentiable binarization operations (e.g., sign function). However, prevailing methods including the Straight-Through Estimator (STE) and its improved variants, rely on hand-crafted designs that suffer from gradient mismatch problem and information loss induced by fixed-range gradient clipping. To address this, we propose SURrogate GradiEnt Adaptation (SURGE), a novel learnable gradient compensation framework with theoretical grounding. SURGE mitigates gradient mismatch through auxiliary b"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments on image classification, object detection, and language understanding tasks demonstrate that SURGE performs best over state-of-the-art methods.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The full-precision auxiliary branch in DPGC provides bias-reduced gradient estimates beyond the first-order STE approximation without introducing new mismatches or instabilities during training.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"SURGE introduces a dual-path gradient compensator and adaptive scaler to improve surrogate gradient estimation in binarized neural network training.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Binary neural networks train more accurately when a learnable surrogate gradient uses an auxiliary full-precision branch to reduce mismatch.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"72687c3fdfd7539a13740308d6a9c557b2a849fc4d6e9e71a8704cd4f1d147b9"},"source":{"id":"2605.10989","kind":"arxiv","version":2},"verdict":{"id":"bc5bbb6d-3359-4593-8909-ecb642af323e","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T17:45:15.360737Z","strongest_claim":"Experiments on image classification, object detection, and language understanding tasks demonstrate that SURGE performs best over state-of-the-art methods.","one_line_summary":"SURGE introduces a dual-path gradient compensator and adaptive scaler to improve surrogate gradient estimation in binarized neural network training.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The full-precision auxiliary branch in DPGC provides bias-reduced gradient estimates beyond the first-order STE approximation without introducing new mismatches or instabilities during training.","pith_extraction_headline":"Binary neural networks train more accurately when a learnable surrogate gradient uses an auxiliary full-precision branch to reduce mismatch."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.10989/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-19T22:33:54.938202Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T14:01:21.770622Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T10:42:35.079079Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"ca90ec4a1b3b3cc7c3c3f17fcbe07abbab950d61e669f743cb4e3207234eef66"},"references":{"count":129,"sample":[{"doi":"","year":null,"title":"Scaling Learning Algorithms Towards","work_id":"bb2761cc-98d0-411b-92f6-803773d64460","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation","work_id":"1fe8c7c8-aff7-4b94-9096-e549d7e60789","ref_index":2,"cited_arxiv_id":"1308.3432","is_internal_anchor":true},{"doi":"","year":null,"title":"Learning sparse neural networks through L\\_0 regularization , author=","work_id":"3379164a-b285-46fb-a1f7-e2faa40cab45","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Differentiable soft quantization: Bridging full-precision and low-bit neural networks , author=","work_id":"f1feb016-058a-4be2-8e69-a25098204a0b","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Binaryconnect: Training deep neural networks with binary weights during propagations , author=","work_id":"e0aa0686-1e81-4f1a-88bd-3e01e1d53ed5","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":129,"snapshot_sha256":"5db2ec13f37ae26fc2e739c2b242a921a7b04b642c4041d8465b906edfc93dfb","internal_anchors":11},"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":"2605.10989","created_at":"2026-05-20T00:00:42.650203+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.10989v2","created_at":"2026-05-20T00:00:42.650203+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.10989","created_at":"2026-05-20T00:00:42.650203+00:00"},{"alias_kind":"pith_short_12","alias_value":"V6FO5VMUUFJS","created_at":"2026-05-20T00:00:42.650203+00:00"},{"alias_kind":"pith_short_16","alias_value":"V6FO5VMUUFJSHSFR","created_at":"2026-05-20T00:00:42.650203+00:00"},{"alias_kind":"pith_short_8","alias_value":"V6FO5VMU","created_at":"2026-05-20T00:00:42.650203+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/V6FO5VMUUFJSHSFRGJMOFIG3ZP","json":"https://pith.science/pith/V6FO5VMUUFJSHSFRGJMOFIG3ZP.json","graph_json":"https://pith.science/api/pith-number/V6FO5VMUUFJSHSFRGJMOFIG3ZP/graph.json","events_json":"https://pith.science/api/pith-number/V6FO5VMUUFJSHSFRGJMOFIG3ZP/events.json","paper":"https://pith.science/paper/V6FO5VMU"},"agent_actions":{"view_html":"https://pith.science/pith/V6FO5VMUUFJSHSFRGJMOFIG3ZP","download_json":"https://pith.science/pith/V6FO5VMUUFJSHSFRGJMOFIG3ZP.json","view_paper":"https://pith.science/paper/V6FO5VMU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.10989&json=true","fetch_graph":"https://pith.science/api/pith-number/V6FO5VMUUFJSHSFRGJMOFIG3ZP/graph.json","fetch_events":"https://pith.science/api/pith-number/V6FO5VMUUFJSHSFRGJMOFIG3ZP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/V6FO5VMUUFJSHSFRGJMOFIG3ZP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/V6FO5VMUUFJSHSFRGJMOFIG3ZP/action/storage_attestation","attest_author":"https://pith.science/pith/V6FO5VMUUFJSHSFRGJMOFIG3ZP/action/author_attestation","sign_citation":"https://pith.science/pith/V6FO5VMUUFJSHSFRGJMOFIG3ZP/action/citation_signature","submit_replication":"https://pith.science/pith/V6FO5VMUUFJSHSFRGJMOFIG3ZP/action/replication_record"}},"created_at":"2026-05-20T00:00:42.650203+00:00","updated_at":"2026-05-20T00:00:42.650203+00:00"}