{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:3Q7EBGTCG5CVZCQ3VKBOE3QCRA","short_pith_number":"pith:3Q7EBGTC","schema_version":"1.0","canonical_sha256":"dc3e409a6237455c8a1baa82e26e02882d21b2fe09db8e4ed57669ecf2fa73ef","source":{"kind":"arxiv","id":"2608.03919","version":1},"attestation_state":"computed","paper":{"title":"Low-Dimensional High-Leverage Subspace Optimization: Beyond Full-Parameter Coupled Training for Neural Network Quantization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Junbiao Pang, Peng Xia, Zheng Huang","submitted_at":"2026-08-04T16:50:04Z","abstract_excerpt":"Low-bit quantization suffers severe accuracy degradation on compact networks, rooted in the dominant full-parameter coupled training paradigm that ignores parameter subspace heterogeneity. Their limited feature redundancy leaves little room to absorb quantization errors. Conventional pipelines adopt monolithic optimization: PTQ reconstructs fixed pretrained models without improving inherent quantization friendliness; QAT updates all parameters jointly, suffering from gradient coupling between backbone weights and calibration parameters. In this paper, we identify normalization affine parameter"},"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":"2608.03919","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-08-04T16:50:04Z","cross_cats_sorted":[],"title_canon_sha256":"686e21c8addba783702a3c3a107df9bd9660ab902b1522ec6a9289bed19f2289","abstract_canon_sha256":"daa05c1dc3f51a4d831f02620c354f71835c47852d9d1f65bca4f56b0e9bb112"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-05T01:37:30.965093Z","signature_b64":"9Cd4S8wND98xCZko25OIoNKkhhUfmKEdd8Z3pWOgJtfXHyoJ8OFl9+r0IuCal+ZJNUnkWFSQr93UXOpVZnyuAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dc3e409a6237455c8a1baa82e26e02882d21b2fe09db8e4ed57669ecf2fa73ef","last_reissued_at":"2026-08-05T01:37:30.963597Z","signature_status":"signed_v1","first_computed_at":"2026-08-05T01:37:30.963597Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Low-Dimensional High-Leverage Subspace Optimization: Beyond Full-Parameter Coupled Training for Neural Network Quantization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Junbiao Pang, Peng Xia, Zheng Huang","submitted_at":"2026-08-04T16:50:04Z","abstract_excerpt":"Low-bit quantization suffers severe accuracy degradation on compact networks, rooted in the dominant full-parameter coupled training paradigm that ignores parameter subspace heterogeneity. Their limited feature redundancy leaves little room to absorb quantization errors. Conventional pipelines adopt monolithic optimization: PTQ reconstructs fixed pretrained models without improving inherent quantization friendliness; QAT updates all parameters jointly, suffering from gradient coupling between backbone weights and calibration parameters. In this paper, we identify normalization affine parameter"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.03919","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/2608.03919/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":"2608.03919","created_at":"2026-08-05T01:37:30.964118+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.03919v1","created_at":"2026-08-05T01:37:30.964118+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.03919","created_at":"2026-08-05T01:37:30.964118+00:00"},{"alias_kind":"pith_short_12","alias_value":"3Q7EBGTCG5CV","created_at":"2026-08-05T01:37:30.964118+00:00"},{"alias_kind":"pith_short_16","alias_value":"3Q7EBGTCG5CVZCQ3","created_at":"2026-08-05T01:37:30.964118+00:00"},{"alias_kind":"pith_short_8","alias_value":"3Q7EBGTC","created_at":"2026-08-05T01:37:30.964118+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/3Q7EBGTCG5CVZCQ3VKBOE3QCRA","json":"https://pith.science/pith/3Q7EBGTCG5CVZCQ3VKBOE3QCRA.json","graph_json":"https://pith.science/api/pith-number/3Q7EBGTCG5CVZCQ3VKBOE3QCRA/graph.json","events_json":"https://pith.science/api/pith-number/3Q7EBGTCG5CVZCQ3VKBOE3QCRA/events.json","paper":"https://pith.science/paper/3Q7EBGTC"},"agent_actions":{"view_html":"https://pith.science/pith/3Q7EBGTCG5CVZCQ3VKBOE3QCRA","download_json":"https://pith.science/pith/3Q7EBGTCG5CVZCQ3VKBOE3QCRA.json","view_paper":"https://pith.science/paper/3Q7EBGTC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.03919&json=true","fetch_graph":"https://pith.science/api/pith-number/3Q7EBGTCG5CVZCQ3VKBOE3QCRA/graph.json","fetch_events":"https://pith.science/api/pith-number/3Q7EBGTCG5CVZCQ3VKBOE3QCRA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3Q7EBGTCG5CVZCQ3VKBOE3QCRA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3Q7EBGTCG5CVZCQ3VKBOE3QCRA/action/storage_attestation","attest_author":"https://pith.science/pith/3Q7EBGTCG5CVZCQ3VKBOE3QCRA/action/author_attestation","sign_citation":"https://pith.science/pith/3Q7EBGTCG5CVZCQ3VKBOE3QCRA/action/citation_signature","submit_replication":"https://pith.science/pith/3Q7EBGTCG5CVZCQ3VKBOE3QCRA/action/replication_record"}},"created_at":"2026-08-05T01:37:30.964118+00:00","updated_at":"2026-08-05T01:37:30.964118+00:00"}