paper-with-me

홈 › Papers

Budget-Adaptive Routing: Skipping the Weak When the Strong Answers Anyway

2026-06-29 · Wei Geng, Nitinder Mohan, Jörg Ott arxiv

Edge-cloud inference collaborations are often designed with a routing estimator that decides whether to offload each frame from weak models at the edge to stronger models in the cloud. Existing systems place the routing estimator after the weak detector, so the weak forward pass still runs even on frames that are later offloaded. In this paper, we argue that this weak-conditioned design can be suboptimal when the offload budget varies. First, we present a competitive weak-skipping estimator (0.153 GFLOPs, about 29x lighter than the weak detector at 4.49 GFLOPs) that extracts routing signal from raw pixels, outperforming the common after-weak placement weak-conditioned baselines. Second, we show that neither weak-skipping nor weak-conditioned placement dominates across the full operating curve, and we propose budget-adaptive routing, which selects between them by offload budget via two offline-tuned thresholds. On PASCAL VOC, our budget-adaptive router traces the upper accuracy envelope of both fixed placements across the operating range. Our method reduces per-frame latency by up to 19.1 ms (about 30% lower at rho = 0.9). Besides outperforming SOTA methods, it is surprisingly stronger than the strong model (+1.7 pp over the strong model's peak mAP) at some operating points with far less compute. Artifacts are available at https://github.com/ViGeng/bgt-ada

📄 PDF Abstract BibTeX arXiv:2606.30919

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

BUDDY: BUdget-Driven DYnamic Depth Routing for Adaptive Large Language Model Inference

2026-06-08 · Yuhua Zhou, Shaoqi Yu, Shichao Weng, Changhai Zhou 외 arxiv

Large language models (LLMs) incur high inference cost due to their depth and parameter scale. Depth pruning can reduce latency by skipping redundant Transformer blocks, but existing methods (i) provide limited control u…

ACE: Adaptive Calibration-Free Expert Skipping for MoE-based LLMs

2026-09-04 · Zukang Xu, Zhixiong Zhao, Xing Hu, Jiangyong Yu 외 arxiv

Mixture-of-Experts (MoE) architectures provide an efficient paradigm for scaling large language models (LLMs), yet fixed top-k routing activates the same number of expert slots for every token, causing substantial redund…

Fractional Skipping: Towards Finer-Grained Dynamic CNN Inference

2020-01-03 · Jianghao Shen, Yonggan Fu, Yue Wang, Pengfei Xu 외

While increasingly deep networks are still in general desired for achieving state-of-the-art performance, for many specific inputs a simpler network might already suffice. Existing works exploited this observation by lea…

Quantization

LiteStage: Latency-aware Layer Skipping for Multi-stage Reasoning

2025-10-16 · Beomseok Kang, Jiwon Song, Jae-Joon Kim arxiv

Multi-stage reasoning has emerged as an effective strategy for enhancing the reasoning capability of small language models by decomposing complex problems into sequential sub-stages. However, this comes at the cost of in…

R2-Router: A New Paradigm for LLM Routing with Reasoning

2026-02-02 · Jiaqi Xue, Qian Lou, Jiarong Xing, Heng Huang arxiv

As LLMs proliferate with diverse capabilities and costs, LLM routing has emerged by learning to predict each LLM's quality and cost for a given query, then selecting the one with high quality and low cost. However, exist…