{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,23]],"date-time":"2026-10-23T20:14:53Z","timestamp":1792786493377,"version":"4.5.0"},"reference-count":23,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/legalcode"}],"funder":[{"name":"Flemish Government through the \u201cOnderzoeksprogramma Artifici\u00eble Intelligentie (AI) Vlaanderen\u201d Program","id":[{"id":"https:\/\/ror.org\/04qxsrb28","id-type":"ROR","asserted-by":"crossref"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Access"],"published-print":{"date-parts":[[2026]]},"DOI":"10.1109\/access.2026.3737178","type":"journal-article","created":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T19:07:34Z","timestamp":1790276854000},"page":"151709-151721","source":"Crossref","is-referenced-by-count":0,"title":["Hybrid Speculative Decoding in RAG: When Do the Pros Outweigh the Cons?"],"prefix":"10.1109","volume":"14","author":[{"ORCID":"https:\/\/orcid.org\/0009-0001-4956-1998","authenticated-orcid":false,"given":"Gianni","family":"van de Velde","sequence":"first","affiliation":[{"name":"Ghent University\u2014imec","place":["Ghent, Belgium"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Cedric","family":"Goemaere","sequence":"additional","affiliation":[{"name":"Ghent University\u2014imec","place":["Ghent, Belgium"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4775-6975","authenticated-orcid":false,"given":"Francois","family":"Remy","sequence":"additional","affiliation":[{"name":"Parallia Healthcare","place":["Ghent, Belgium"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2707-4176","authenticated-orcid":false,"given":"Chris","family":"Develder","sequence":"additional","affiliation":[{"name":"Ghent University\u2014imec","place":["Ghent, Belgium"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Thomas","family":"Demeester","sequence":"additional","affiliation":[{"name":"Ghent University\u2014imec","place":["Ghent, Belgium"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","first-page":"19274","article-title":"Fast inference from transformers via speculative decoding","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Leviathan"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.52202\/085713-4562"},{"key":"ref3","article-title":"Medusa: Simple LLM inference acceleration framework with multiple decoding heads","author":"Cai","year":"2024","journal-title":"arXiv:2401.10774"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.emnlp-main.1337"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.acl-long.595"},{"key":"ref6","article-title":"Optimized multi-token joint decoding with auxiliary model for LLM inference","author":"Qin","year":"2024","journal-title":"arXiv:2407.09722"},{"key":"ref7","volume-title":"Prompt Lookup Decoding","author":"Saxena","year":"2023"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.naacl-long.88"},{"key":"ref9","article-title":"Inference with reference: Lossless acceleration of large language models","author":"Yang","year":"2023","journal-title":"arXiv:2304.04487"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.emnlp-main.911"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.52202\/085713-4211"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.findings-acl.320"},{"key":"ref13","article-title":"RAGBench: Explainable benchmark for retrieval-augmented generation systems","author":"Friel","year":"2024","journal-title":"arXiv:2407.11005"},{"key":"ref14","volume-title":"Stanford Alpaca: An Instruction-Following Llama Model","author":"Taori","year":"2023"},{"key":"ref15","article-title":"Training verifiers to solve math word problems","author":"Cobbe","year":"2021","journal-title":"arXiv:2110.14168"},{"key":"ref16","article-title":"Evaluating large language models trained on code","author":"Chen","year":"2021","journal-title":"arXiv:2107.03374"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2020"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1162\/tacl_a_00276"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/K16-1028"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.emnlp-main.422"},{"key":"ref21","article-title":"The Llama 3 herd of models","author":"Grattafiori","year":"2024","journal-title":"arXiv:2407.21783"},{"key":"ref22","article-title":"Qwen3 technical report","volume-title":"arXiv:2505.09388","author":"Yang","year":"2025"},{"key":"ref23","volume-title":"Speculative Decoding in VLLM","year":"2024"}],"container-title":["IEEE Access"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/6287639\/11323511\/11707247.pdf?arnumber=11707247","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,7]],"date-time":"2026-10-07T19:14:16Z","timestamp":1791400456000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11707247\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":23,"URL":"https:\/\/doi.org\/10.1109\/access.2026.3737178","relation":{},"ISSN":["2169-3536"],"issn-type":[{"value":"2169-3536","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026]]}}}