Delay-Adaptive Speculation Control for Low-Latency Edge-Cloud LLM Inference
Speculative decoding accelerates large language model (LLM) inference by using a lightweight draft model to propose tokens and a larger target model to verify them in parallel. In distributed edge-cloud inference, however, draft length must be controlled online: longer drafts amortize communication delay but reduce token acceptance, whereas shorter drafts preserve acceptance but trigger more communication rounds. We formulate this tradeoff as a ratio-type optimal stopping problem and prove that the optimal draft length is a finite delay-monotone threshold. The analysis identifies a critical delay below which single-token speculation is optimal and shows that the optimal length grows only ...
| Authors | Kangkang Sun, Jianhua Li, Xiuzhen Chen et al. |
|---|---|
| Primary category | cs.NI |
| Published | 17 May 2026 |
| Last updated at source | 17 May 2026 |
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.