{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:TGZQ4H32FWS5UDE5KI2QM46UXW","short_pith_number":"pith:TGZQ4H32","schema_version":"1.0","canonical_sha256":"99b30e1f7a2da5da0c9d52350673d4bd839124f1116c152e98cdb373a5c17bc8","source":{"kind":"arxiv","id":"2605.17997","version":1},"attestation_state":"computed","paper":{"title":"MARR: Module-Adaptive Residual Reconstruction for Low-Bit Post-Training Quantization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.LG","authors_text":"Le Su, Xing Luo, Zhi Jin","submitted_at":"2026-05-18T07:51:49Z","abstract_excerpt":"Recently, residual reconstruction-based model quantization methods have achieved promising performance in low-bit post-training quantization (PTQ) by introducing cross-layer residuals to reduce error accumulated from previous layers.However, these residuals may also introduce additional bias arising from the Hessian-approximation (HA) assumption underlying reconstruction-based PTQ, leading to suboptimal quantization performance.In this work, we analyze that multiplying the residual term by a scaling coefficient provides a direct way to mitigate the HA bias associated with residual strength, wh"},"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":"2605.17997","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-18T07:51:49Z","cross_cats_sorted":["cs.AI","cs.CV"],"title_canon_sha256":"d5583d045d28317a9e268d92836775a33846c8fa1129731f5c8695761dbcc055","abstract_canon_sha256":"cf8a8b14aae658a14f08883dc1891b8bf300d6e5b77f299018910aa576e6ca21"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:05:10.353077Z","signature_b64":"/UBstmBIX+ZElzhinZtxaOVzVOAiTMSOO8p6n+sXCzSW1mIBD1pgMXjZFeV80EEF/5WO/tEHlMKsfeP+a/nFBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"99b30e1f7a2da5da0c9d52350673d4bd839124f1116c152e98cdb373a5c17bc8","last_reissued_at":"2026-05-20T00:05:10.352269Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:05:10.352269Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MARR: Module-Adaptive Residual Reconstruction for Low-Bit Post-Training Quantization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.LG","authors_text":"Le Su, Xing Luo, Zhi Jin","submitted_at":"2026-05-18T07:51:49Z","abstract_excerpt":"Recently, residual reconstruction-based model quantization methods have achieved promising performance in low-bit post-training quantization (PTQ) by introducing cross-layer residuals to reduce error accumulated from previous layers.However, these residuals may also introduce additional bias arising from the Hessian-approximation (HA) assumption underlying reconstruction-based PTQ, leading to suboptimal quantization performance.In this work, we analyze that multiplying the residual term by a scaling coefficient provides a direct way to mitigate the HA bias associated with residual strength, wh"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.17997","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/2605.17997/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-19T23:33:35.545908Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"84550f013e02dd3683b14dbfd3f3219e666f5c987d92999342b2a6dda130c978"},"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":"2605.17997","created_at":"2026-05-20T00:05:10.352419+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.17997v1","created_at":"2026-05-20T00:05:10.352419+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17997","created_at":"2026-05-20T00:05:10.352419+00:00"},{"alias_kind":"pith_short_12","alias_value":"TGZQ4H32FWS5","created_at":"2026-05-20T00:05:10.352419+00:00"},{"alias_kind":"pith_short_16","alias_value":"TGZQ4H32FWS5UDE5","created_at":"2026-05-20T00:05:10.352419+00:00"},{"alias_kind":"pith_short_8","alias_value":"TGZQ4H32","created_at":"2026-05-20T00:05:10.352419+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/TGZQ4H32FWS5UDE5KI2QM46UXW","json":"https://pith.science/pith/TGZQ4H32FWS5UDE5KI2QM46UXW.json","graph_json":"https://pith.science/api/pith-number/TGZQ4H32FWS5UDE5KI2QM46UXW/graph.json","events_json":"https://pith.science/api/pith-number/TGZQ4H32FWS5UDE5KI2QM46UXW/events.json","paper":"https://pith.science/paper/TGZQ4H32"},"agent_actions":{"view_html":"https://pith.science/pith/TGZQ4H32FWS5UDE5KI2QM46UXW","download_json":"https://pith.science/pith/TGZQ4H32FWS5UDE5KI2QM46UXW.json","view_paper":"https://pith.science/paper/TGZQ4H32","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.17997&json=true","fetch_graph":"https://pith.science/api/pith-number/TGZQ4H32FWS5UDE5KI2QM46UXW/graph.json","fetch_events":"https://pith.science/api/pith-number/TGZQ4H32FWS5UDE5KI2QM46UXW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TGZQ4H32FWS5UDE5KI2QM46UXW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TGZQ4H32FWS5UDE5KI2QM46UXW/action/storage_attestation","attest_author":"https://pith.science/pith/TGZQ4H32FWS5UDE5KI2QM46UXW/action/author_attestation","sign_citation":"https://pith.science/pith/TGZQ4H32FWS5UDE5KI2QM46UXW/action/citation_signature","submit_replication":"https://pith.science/pith/TGZQ4H32FWS5UDE5KI2QM46UXW/action/replication_record"}},"created_at":"2026-05-20T00:05:10.352419+00:00","updated_at":"2026-05-20T00:05:10.352419+00:00"}