{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:RBBEHN7DBMDZOFWINC45VNEAHC","short_pith_number":"pith:RBBEHN7D","schema_version":"1.0","canonical_sha256":"884243b7e30b079716c868b9dab4803890614e70f825747d4d674952f8a3cd7e","source":{"kind":"arxiv","id":"2203.16487","version":6},"attestation_state":"computed","paper":{"title":"Speculative Decoding: Exploiting Speculative Execution for Accelerating Seq2seq Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Furu Wei, Heming Xia, Peiyi Wang, Si-Qing Chen, Tao Ge, Zhifang Sui","submitted_at":"2022-03-30T17:27:09Z","abstract_excerpt":"We propose Speculative Decoding (SpecDec), for the first time ever, to formally study exploiting the idea of speculative execution to accelerate autoregressive (AR) decoding. Speculative Decoding has two innovations: Spec-Drafter -- an independent model specially optimized for efficient and accurate drafting -- and Spec-Verification -- a reliable method for verifying the drafted tokens efficiently in the decoding paradigm. Experimental results on various seq2seq tasks including machine translation and abstractive summarization show our approach can achieve around $5\\times$ speedup for the popu"},"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":"2203.16487","kind":"arxiv","version":6},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-03-30T17:27:09Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"9bfa70fc29cfe8405c939bf4d481b0047603b27eb9917c83643b7e1fb7583809","abstract_canon_sha256":"4adf4b84b61e72544c87a43e5817673763c7ea01bb434d150bc4adddd38874d8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:06:18.427645Z","signature_b64":"hvbWV+jqD67kp4pEqnDT2IE9X0i7PQm6Vs2YYXGxk44dT2p/Gsm+Co6ebo6Z8dedfOkII5jXG5AQt0P3AwzkAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"884243b7e30b079716c868b9dab4803890614e70f825747d4d674952f8a3cd7e","last_reissued_at":"2026-07-05T07:06:18.427101Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:06:18.427101Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Speculative Decoding: Exploiting Speculative Execution for Accelerating Seq2seq Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Furu Wei, Heming Xia, Peiyi Wang, Si-Qing Chen, Tao Ge, Zhifang Sui","submitted_at":"2022-03-30T17:27:09Z","abstract_excerpt":"We propose Speculative Decoding (SpecDec), for the first time ever, to formally study exploiting the idea of speculative execution to accelerate autoregressive (AR) decoding. Speculative Decoding has two innovations: Spec-Drafter -- an independent model specially optimized for efficient and accurate drafting -- and Spec-Verification -- a reliable method for verifying the drafted tokens efficiently in the decoding paradigm. Experimental results on various seq2seq tasks including machine translation and abstractive summarization show our approach can achieve around $5\\times$ speedup for the popu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.16487","kind":"arxiv","version":6},"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/2203.16487/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":"2203.16487","created_at":"2026-07-05T07:06:18.427161+00:00"},{"alias_kind":"arxiv_version","alias_value":"2203.16487v6","created_at":"2026-07-05T07:06:18.427161+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.16487","created_at":"2026-07-05T07:06:18.427161+00:00"},{"alias_kind":"pith_short_12","alias_value":"RBBEHN7DBMDZ","created_at":"2026-07-05T07:06:18.427161+00:00"},{"alias_kind":"pith_short_16","alias_value":"RBBEHN7DBMDZOFWI","created_at":"2026-07-05T07:06:18.427161+00:00"},{"alias_kind":"pith_short_8","alias_value":"RBBEHN7D","created_at":"2026-07-05T07:06:18.427161+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2211.15089","citing_title":"Continuous diffusion for categorical data","ref_index":96,"is_internal_anchor":false},{"citing_arxiv_id":"2312.08935","citing_title":"Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotations","ref_index":89,"is_internal_anchor":false},{"citing_arxiv_id":"2604.27747","citing_title":"Position-Aware Drafting for Inference Acceleration in LLM-Based Generative List-Wise Recommendation","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07307","citing_title":"Rethinking Dense Sequential Chains: Reasoning Language Models Can Extract Answers from Sparse, Order-Shuffling Chain-of-Thoughts","ref_index":12,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RBBEHN7DBMDZOFWINC45VNEAHC","json":"https://pith.science/pith/RBBEHN7DBMDZOFWINC45VNEAHC.json","graph_json":"https://pith.science/api/pith-number/RBBEHN7DBMDZOFWINC45VNEAHC/graph.json","events_json":"https://pith.science/api/pith-number/RBBEHN7DBMDZOFWINC45VNEAHC/events.json","paper":"https://pith.science/paper/RBBEHN7D"},"agent_actions":{"view_html":"https://pith.science/pith/RBBEHN7DBMDZOFWINC45VNEAHC","download_json":"https://pith.science/pith/RBBEHN7DBMDZOFWINC45VNEAHC.json","view_paper":"https://pith.science/paper/RBBEHN7D","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2203.16487&json=true","fetch_graph":"https://pith.science/api/pith-number/RBBEHN7DBMDZOFWINC45VNEAHC/graph.json","fetch_events":"https://pith.science/api/pith-number/RBBEHN7DBMDZOFWINC45VNEAHC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RBBEHN7DBMDZOFWINC45VNEAHC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RBBEHN7DBMDZOFWINC45VNEAHC/action/storage_attestation","attest_author":"https://pith.science/pith/RBBEHN7DBMDZOFWINC45VNEAHC/action/author_attestation","sign_citation":"https://pith.science/pith/RBBEHN7DBMDZOFWINC45VNEAHC/action/citation_signature","submit_replication":"https://pith.science/pith/RBBEHN7DBMDZOFWINC45VNEAHC/action/replication_record"}},"created_at":"2026-07-05T07:06:18.427161+00:00","updated_at":"2026-07-05T07:06:18.427161+00:00"}