{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:GZTLU2WX55X6X5HSVEB23XAKYM","short_pith_number":"pith:GZTLU2WX","schema_version":"1.0","canonical_sha256":"3666ba6ad7ef6febf4f2a903addc0ac339d3b4ade89ed107b95f6e85418513c3","source":{"kind":"arxiv","id":"2505.22179","version":2},"attestation_state":"computed","paper":{"title":"Speculative Decoding Meets Quantization: Compatibility Evaluation and Hierarchical Framework Design","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Conghui Zhu, Hailong Cao, Tiejun Zhao, Wang Xu, Weilin Zhao, Xu Han, Yudi Zhang","submitted_at":"2025-05-28T09:55:08Z","abstract_excerpt":"Speculative decoding and quantization effectively accelerate memory-bound inference of large language models. Speculative decoding mitigates the memory bandwidth bottleneck by verifying multiple tokens within a single forward pass, which increases computational effort. Quantization achieves this optimization by compressing weights and activations into lower bit-widths and also reduces computations via low-bit matrix multiplications. To further leverage their strengths, we investigate the integration of these two techniques. Surprisingly, experiments applying the advanced speculative decoding m"},"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":"2505.22179","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-28T09:55:08Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"52e1cedaa2cee9b749f49b8a1ceb0453b1735f746d14d489dcb58902dcad4dec","abstract_canon_sha256":"16e6fb4fd0f5b69e00a5b6fbb0bbadaaa47b1f97b2c32ff730e8691bf93ad14c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:11:47.782270Z","signature_b64":"3IhKwiqKQtyMC7ZAOlDZD2+Ygqz8LDmCtT/hbvzEvaPB83pYSCENS/e2ECL2uBT/HJVhaLXciAEQf200orf7CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3666ba6ad7ef6febf4f2a903addc0ac339d3b4ade89ed107b95f6e85418513c3","last_reissued_at":"2026-07-05T11:11:47.781777Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:11:47.781777Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Speculative Decoding Meets Quantization: Compatibility Evaluation and Hierarchical Framework Design","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Conghui Zhu, Hailong Cao, Tiejun Zhao, Wang Xu, Weilin Zhao, Xu Han, Yudi Zhang","submitted_at":"2025-05-28T09:55:08Z","abstract_excerpt":"Speculative decoding and quantization effectively accelerate memory-bound inference of large language models. Speculative decoding mitigates the memory bandwidth bottleneck by verifying multiple tokens within a single forward pass, which increases computational effort. Quantization achieves this optimization by compressing weights and activations into lower bit-widths and also reduces computations via low-bit matrix multiplications. To further leverage their strengths, we investigate the integration of these two techniques. Surprisingly, experiments applying the advanced speculative decoding m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.22179","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/2505.22179/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":"2505.22179","created_at":"2026-07-05T11:11:47.781838+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.22179v2","created_at":"2026-07-05T11:11:47.781838+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.22179","created_at":"2026-07-05T11:11:47.781838+00:00"},{"alias_kind":"pith_short_12","alias_value":"GZTLU2WX55X6","created_at":"2026-07-05T11:11:47.781838+00:00"},{"alias_kind":"pith_short_16","alias_value":"GZTLU2WX55X6X5HS","created_at":"2026-07-05T11:11:47.781838+00:00"},{"alias_kind":"pith_short_8","alias_value":"GZTLU2WX","created_at":"2026-07-05T11:11:47.781838+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/GZTLU2WX55X6X5HSVEB23XAKYM","json":"https://pith.science/pith/GZTLU2WX55X6X5HSVEB23XAKYM.json","graph_json":"https://pith.science/api/pith-number/GZTLU2WX55X6X5HSVEB23XAKYM/graph.json","events_json":"https://pith.science/api/pith-number/GZTLU2WX55X6X5HSVEB23XAKYM/events.json","paper":"https://pith.science/paper/GZTLU2WX"},"agent_actions":{"view_html":"https://pith.science/pith/GZTLU2WX55X6X5HSVEB23XAKYM","download_json":"https://pith.science/pith/GZTLU2WX55X6X5HSVEB23XAKYM.json","view_paper":"https://pith.science/paper/GZTLU2WX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.22179&json=true","fetch_graph":"https://pith.science/api/pith-number/GZTLU2WX55X6X5HSVEB23XAKYM/graph.json","fetch_events":"https://pith.science/api/pith-number/GZTLU2WX55X6X5HSVEB23XAKYM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GZTLU2WX55X6X5HSVEB23XAKYM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GZTLU2WX55X6X5HSVEB23XAKYM/action/storage_attestation","attest_author":"https://pith.science/pith/GZTLU2WX55X6X5HSVEB23XAKYM/action/author_attestation","sign_citation":"https://pith.science/pith/GZTLU2WX55X6X5HSVEB23XAKYM/action/citation_signature","submit_replication":"https://pith.science/pith/GZTLU2WX55X6X5HSVEB23XAKYM/action/replication_record"}},"created_at":"2026-07-05T11:11:47.781838+00:00","updated_at":"2026-07-05T11:11:47.781838+00:00"}