{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:7G4LM3UMGR5ZQ2RC4R5ALYRN77","short_pith_number":"pith:7G4LM3UM","schema_version":"1.0","canonical_sha256":"f9b8b66e8c347b986a22e47a05e22dffe3005aaa99f9326698d69a0cbb5f6583","source":{"kind":"arxiv","id":"2512.12930","version":2},"attestation_state":"computed","paper":{"title":"SeVeDo: A Heterogeneous Transformer Accelerator for Low-Bit Inference via Hierarchical Group Quantization and SVD-Guided Mixed Precision","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AR"],"primary_cat":"cs.LG","authors_text":"Byeongcheol Kim, Hoi-Jun Yoo, Jungjun Oh, Sangjin Kim, Yuseon Choi","submitted_at":"2025-12-15T02:29:08Z","abstract_excerpt":"Low-bit quantization is a promising technique for efficient transformer inference by reducing computational and memory overhead. However, aggressive bitwidth reduction remains challenging due to activation outliers, leading to accuracy degradation. Existing methods, such as outlier-handling and group quantization, achieve high accuracy but incur substantial energy consumption. To address this, we propose SeVeDo, an energy-efficient SVD-based heterogeneous accelerator that structurally separates outlier-sensitive components into a high-precision low-rank path, while the remaining computations a"},"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":"2512.12930","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-12-15T02:29:08Z","cross_cats_sorted":["cs.AR"],"title_canon_sha256":"3b3ec925f312ccb8c233faf7baa7d2576f29a8c235df61bde523fde199e1f3b9","abstract_canon_sha256":"5951c036bd10210e9d330a6cdd8f3100097295dcc545c818c6f48baa76e4cd38"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T03:13:52.420929Z","signature_b64":"+ibeLBy9lXMIUDAAQGLh/ws3v5AZHhgASD3RDClvtdwAvp2+V331HkzAY4BhK2IqYigs57q1ukU9vwx5bGsBDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f9b8b66e8c347b986a22e47a05e22dffe3005aaa99f9326698d69a0cbb5f6583","last_reissued_at":"2026-06-23T03:13:52.420445Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T03:13:52.420445Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SeVeDo: A Heterogeneous Transformer Accelerator for Low-Bit Inference via Hierarchical Group Quantization and SVD-Guided Mixed Precision","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AR"],"primary_cat":"cs.LG","authors_text":"Byeongcheol Kim, Hoi-Jun Yoo, Jungjun Oh, Sangjin Kim, Yuseon Choi","submitted_at":"2025-12-15T02:29:08Z","abstract_excerpt":"Low-bit quantization is a promising technique for efficient transformer inference by reducing computational and memory overhead. However, aggressive bitwidth reduction remains challenging due to activation outliers, leading to accuracy degradation. Existing methods, such as outlier-handling and group quantization, achieve high accuracy but incur substantial energy consumption. To address this, we propose SeVeDo, an energy-efficient SVD-based heterogeneous accelerator that structurally separates outlier-sensitive components into a high-precision low-rank path, while the remaining computations a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2512.12930","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/2512.12930/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":"2512.12930","created_at":"2026-06-23T03:13:52.420504+00:00"},{"alias_kind":"arxiv_version","alias_value":"2512.12930v2","created_at":"2026-06-23T03:13:52.420504+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2512.12930","created_at":"2026-06-23T03:13:52.420504+00:00"},{"alias_kind":"pith_short_12","alias_value":"7G4LM3UMGR5Z","created_at":"2026-06-23T03:13:52.420504+00:00"},{"alias_kind":"pith_short_16","alias_value":"7G4LM3UMGR5ZQ2RC","created_at":"2026-06-23T03:13:52.420504+00:00"},{"alias_kind":"pith_short_8","alias_value":"7G4LM3UM","created_at":"2026-06-23T03:13:52.420504+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2604.14626","citing_title":"ELMoE-3D: Leveraging Intrinsic Elasticity of MoE for Hybrid-Bonding-Enabled Self-Speculative Decoding in On-Premises Serving","ref_index":7,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7G4LM3UMGR5ZQ2RC4R5ALYRN77","json":"https://pith.science/pith/7G4LM3UMGR5ZQ2RC4R5ALYRN77.json","graph_json":"https://pith.science/api/pith-number/7G4LM3UMGR5ZQ2RC4R5ALYRN77/graph.json","events_json":"https://pith.science/api/pith-number/7G4LM3UMGR5ZQ2RC4R5ALYRN77/events.json","paper":"https://pith.science/paper/7G4LM3UM"},"agent_actions":{"view_html":"https://pith.science/pith/7G4LM3UMGR5ZQ2RC4R5ALYRN77","download_json":"https://pith.science/pith/7G4LM3UMGR5ZQ2RC4R5ALYRN77.json","view_paper":"https://pith.science/paper/7G4LM3UM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2512.12930&json=true","fetch_graph":"https://pith.science/api/pith-number/7G4LM3UMGR5ZQ2RC4R5ALYRN77/graph.json","fetch_events":"https://pith.science/api/pith-number/7G4LM3UMGR5ZQ2RC4R5ALYRN77/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7G4LM3UMGR5ZQ2RC4R5ALYRN77/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7G4LM3UMGR5ZQ2RC4R5ALYRN77/action/storage_attestation","attest_author":"https://pith.science/pith/7G4LM3UMGR5ZQ2RC4R5ALYRN77/action/author_attestation","sign_citation":"https://pith.science/pith/7G4LM3UMGR5ZQ2RC4R5ALYRN77/action/citation_signature","submit_replication":"https://pith.science/pith/7G4LM3UMGR5ZQ2RC4R5ALYRN77/action/replication_record"}},"created_at":"2026-06-23T03:13:52.420504+00:00","updated_at":"2026-06-23T03:13:52.420504+00:00"}