mlsanityData to 14 Oct 2026

CR^2: Cost-Aware Risk-Controlled Routing for Wireless Device-Edge LLM Inference

arXiv 2605.12001 ↗First posted 12 May 2026cs.IT cs.AI math.ITcited by 1

Not on the map. Why?

Abstract (arXiv)

As large language models (LLMs) move from centralized clouds to mobile edge environments, efficient serving must balance latency, energy consumption, and accuracy under constrained device-edge resources. Query-level routing between lightweight on-device models and stronger edge models provides a flexible mechanism to navigate this trade-off. However, existing routers are designed for centralized cloud settings and optimize token-level costs, failing to capture the dynamic latency and energy overheads in wireless edge deployments. In this paper, we formulate mobile edge LLM routing as a deployment-constrained, cost-aware decision problem, and propose CR^2, a two-stage device-edge routing framework. CR^2 decouples a lightweight on-device margin gate from an edge-side utility selector for deferred queries. The margin gate operates on frozen query embeddings and a user-specified cost weight to predict whether local execution is utility-optimal relative to the best edge alternative under the target operating point. We further introduce a conformal risk control (CRC) calibration procedure that maps each operating point to an acceptance threshold, enabling explicit control of the marginal false-acceptance risk under the full-information utility reference. Experiments on the routing task show that CR^2 closely matches a full-information reference router using only device-side signals before deferral. Compared with strong query-level baselines, CR^2 consistently improves the deployable accuracy-cost Pareto frontier and reduces normalized deployment cost by up to 16.9% at matched accuracy.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 40 of its 49 references that are in the map. First posted on arXiv 12 May 2026.

Builds on

40 of 49 references.

References · 40

  1. 2023GPT-4 Technical Reportcited by 18,396
  2. 2025Qwen3 Technical Reportcited by 8,820
  3. 2020Measuring Massive Multitask Language Understandingcited by 7,963
  4. 2021Program Synthesis with Large Language Modelscited by 3,637
  5. 2021Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsitycited by 3,073
  6. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,879
  7. 2022Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Themcited by 1,807
  8. 2022Fast Inference from Transformers via Speculative Decodingcited by 1,653
  9. 2023FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performancecited by 853
  10. 2024RouteLLM: Learning to Route LLMs with Preference Datacited by 682
  11. 2022Mixture-of-Experts with Expert Choice Routingcited by 623
  12. 2024EAGLE: Speculative Sampling Requires Rethinking Feature Uncertaintycited by 578
28 more
  1. 2024Hybrid LLM: Cost-Efficient and Quality-Aware Query Routingcited by 372
  2. 2022Conformal Risk Controlcited by 299
  3. 2021Learn then Test: Calibrating Predictive Algorithms to Achieve Risk Controlcited by 271
  4. 2024RouterBench: A Benchmark for Multi-LLM Routing Systemcited by 238
  5. 2024GraphRouter: A Graph-based Router for LLM Selectionscited by 165
  6. 2023Routing to the Expert: Efficient Reward-guided Ensemble of Large Language Modelscited by 145
  7. 2024Mobile Edge Intelligence for Large Language Models: A Contemporary Surveycited by 128
  8. 2020SPINN: Synergistic Progressive Inference of Neural Networks over Device and Cloudcited by 112
  9. 2024Sequoia: Scalable, Robust, and Hardware-aware Speculative Decodingcited by 103
  10. 2024RouterDC: Query-Based Router by Dual Contrastive Learning for Assembling Large Language Modelscited by 88
  11. 2020EdgeBERT: Sentence-Level Energy Optimizations for Latency-Aware Multi-Task NLP Inferencecited by 70
  12. 2025BEST-Route: Adaptive LLM Routing with Test-Time Optimal Computecited by 67
  13. 2024TensorOpera Router: A Multi-Model Router for Efficient LLM Inferencecited by 60
  14. 2018Edge Intelligence: On-Demand Deep Learning Model Co-Inference with Device-Edge Synergycited by 51
  15. 2024Smoothie: Label Free Language Model Routingcited by 31
  16. 2024CE-CoLLM: Efficient and Adaptive Large Language Models Through Cloud-Edge Collaborationcited by 25
  17. 2024WDMoE: Wireless Distributed Mixture of Experts for Large Language Modelscited by 12
  18. 2025R2-T2: Re-Routing in Test-Time for Multimodal Mixture-of-Expertscited by 12
  19. 2025RadialRouter: Structured Representation for Efficient and Robust Large Language Models Routingcited by 11
  20. 2025Serving Long-Context LLMs at the Mobile Edge: Test-Time Reinforcement Learning-based Model Caching and Inference Offloadingcited by 10
  21. 2025SlimCaching: Edge Caching of Mixture-of-Experts for Distributed Inferencecited by 7
  22. 2025Quality-of-Service Aware LLM Routing for Edge Computing with Multiple Expertscited by 7
  23. 2025Dynamic Quality-Latency Aware Routing for LLM Inference in Wireless Edge-Device Networkscited by 4
  24. 2025LLMRank: Understanding LLM Strengths for Model Routingcited by 4
  25. 2025CSGO: Generalized Optimization for Cold Start in Wireless Collaborative Edge LLM Systemscited by 2
  26. 2025Stable-MoE: Lyapunov-based Token Routing for Distributed Mixture-of-Experts Training over Edge Networkscited by 1
  27. 2026TriSpec: Ternary Speculative Decoding via Lightweight Proxy Verificationcited by 1
  28. 2026WISV: Wireless-Informed Semantic Verification for Distributed Speculative Decoding in Device-Edge LLM Inferencecited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026CopRoute: Dependence-Aware Conformal Routing for Language Modelscited by 0