Abstract:Speculative decoding is widely used in accelerating large language model (LLM) inference. In this work, we focus on the online draft model selection problem in speculative decoding. We design an algorithm that provably competes with the best draft model in hindsight for each query in terms of either the token acceptance probability or expected acceptance length. In particular, we show that we can accurately evaluate all draft models, instead of only the chosen model without incurring additional queries to the target model, which allows us to improve exponentially over the existing bandit-based approach as the number of draft models increases. Our approach is generically applicable with any speculative decoding methods (single draft, multi-drafts and draft-trees). Moreover, we design system-efficient versions of online learners and demonstrate that the overhead in computation and latency can be substantially reduced. We conduct extensive experiments on open-source LLMs and diverse datasets, demonstrating that our methods substantially outperform the state-of-the-art EAGLE3 and the BanditSpec baseline in a variety of domains where specialized domain-expert drafters are available, especially when long reasoning chains are required.
| Comments: | ICLR'26 |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2510.20064 [cs.LG] |
| (or arXiv:2510.20064v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2510.20064 arXiv-issued DOI via DataCite |
From: Hongyi Liu [view email]
[v1]
Wed, 22 Oct 2025 22:32:26 UTC (2,103 KB)
[v2]
Wed, 22 Apr 2026 23:56:14 UTC (2,119 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。