AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

PipeSD: An Efficient Cloud-Edge Collaborative Pipeline Inference Framework with Speculative Decoding

Yunhe Han, Yunqi Gao, Bing Hu, Mahdi Boloursaz Mashhadi, Yitong Duan, Pei Xiao, Yanfeng Zhang

ICML 2026 · Architectures and efficiency › RNNs and normalization · approach: Distributed LLM inference systems · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add token-batch pipeline scheduling (overlap draft generation and communication) (ablation)PipeSD · HumanEval, Scenario 1time per accepted token (ms): 147 -> 129Improved
use dual-threshold NAV triggering instead of fixed-length triggering (ablation)PipeSD · HumanEval, Scenario 1time per accepted token (ms): 164 -> 129Improved
use dual-threshold instead of token-level confidence only triggering (ablation)PipeSD · HumanEval, Scenario 1time per accepted token (ms): 137 -> 129Improved
use dual-threshold instead of sequence-level confidence only triggering (ablation)PipeSD · HumanEval, Scenario 1time per accepted token (ms): 139 -> 129Improved
tune thresholds with BO autotuner instead of grid search (ablation)PipeSD · HumanEval, Scenario 1time per accepted token (ms): 139 -> 129Improved
tune thresholds with BO autotuner instead of random search (ablation)PipeSD · GSM8K, Scenario 1time per accepted token (ms): 162 -> 145Improved
use BO autotuner instead of fixed (R1, R2) thresholds (ablation)PipeSD · HumanEval, Scenario 1time per accepted token (ms): 138 (best fixed 0.9, 0.9) -> 129Improved
use PipeSD instead of Vanilla, HSL and EdgeLLM on a weak edge device (Scenario 3)PipeSD · Scenario 3, GSM8Kspeedup over Vanilla: 2.16xImproved
use PipeSD versus baselines at high uplink bandwidth (80 Mbps)PipeSD · HumanEval, Scenario 1, uplink 80 Mbpsspeedup over Vanilla: 1.47x at 20 Mbps -> 1.34x at 80 MbpsMixed
use PipeSD instead of baselines for cloud energyPipeSD · HumanEval, Scenario 1cloud energy reduction (%): 17.6%, 21.1%, 25.3% vs Vanilla, HSL, EdgeLLMImproved
use dual-threshold triggering instead of HSL and EdgeLLM triggeringPipeSD · HumanEval, Scenario 1acceptance rate: 0.9148 (HSL) / 0.8917 (EdgeLLM) -> 0.9616Improved
add DP token batching versus heuristic pipelined policies (ablation)PipeSD · HumanEval, Scenario 1time per accepted tokenImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy