mlsanityData to 14 Oct 2026

PRISM: Privacy-Aware Routing for Adaptive Cloud-Edge LLM Inference via Semantic Sketch Collaboration

arXiv 2511.22788 ↗First posted 27 Nov 2025cs.CR cs.CLcited by 10

Not on the map. Why?

Abstract (arXiv)

Large Language Models (LLMs) demonstrate impressive capabilities in natural language understanding and generation, but incur high communication overhead and privacy risks in cloud deployments, while facing compute and memory constraints when confined to edge devices. Cloud-edge inference has emerged as a promising paradigm for improving privacy in LLM services by retaining sensitive computations on local devices. However, existing cloud-edge inference approaches apply uniform privacy protection without considering input sensitivity, resulting in unnecessary perturbation and degraded utility even for non-sensitive tokens. To address this limitation, we propose Privacy-aware Routing for Inference with Semantic Modulation (PRISM), a context-aware framework that dynamically balances privacy and inference quality. PRISM executes in four stages: (1) the edge device profiles entity-level sensitivity; (2) a soft gating module on the edge selects an execution mode - cloud, edge, or collaboration; (3) for collaborative paths, the edge applies adaptive two-layer local differential privacy based on entity risks; and (4) the cloud LLM generates a semantic sketch from the perturbed prompt, which is then refined by the edge-side small language model (SLM) using local context. Our results show that PRISM consistently achieves superior privacy-utility trade-offs across various scenarios, reducing energy consumption and latency to 40-50% of baseline methods such as Uniform and Selective LDP, while maintaining high output quality under strong privacy constraints. These findings are validated through comprehensive evaluations involving realistic prompts, actual energy measurements, and heterogeneous cloud-edge model deployments.

Key facts. Cited by 10 papers in the map (data through 14 Oct 2026). Builds on 16 of its 29 references that are in the map. First posted on arXiv 27 Nov 2025.

Builds on

16 of 29 references.

References · 16

  1. 2024The Llama 3 Herd of Modelscited by 14,715
  2. 2025Qwen3 Technical Reportcited by 8,820
  3. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  4. 2024Mobile Edge Intelligence for Large Language Models: A Contemporary Surveycited by 128
  5. 2023Towards Efficient Generative Large Language Model Serving: A Survey from Algorithms to Systemscited by 102
  6. 2021A Neural Network Solves, Explains, and Generates University Math Problems by Program Synthesis and Few-Shot Learning at Human Levelcited by 96
  7. 2023Efficient Parallel Split Learning over Resource-constrained Wireless Edge Networkscited by 95
  8. 2023DP-Forward: Fine-tuning and Inference on Language Models with Differential Privacy in Forward Passcited by 67
  9. 2021Selective Differential Privacy for Language Modelingcited by 62
  10. 2024Hierarchical Split Federated Learning: Convergence Analysis and System Optimizationcited by 38
  11. 2024EHRNoteQA: An LLM Benchmark for Real-World Clinical Practice Using Discharge Summariescited by 35
  12. 2024PrivacyRestore: Privacy-Preserving Inference in Large Language Models via Privacy Removal and Restorationcited by 28
4 more
  1. 2019Context-Aware Local Differential Privacycited by 18
  2. 2024EmojiPrompt: Generative Prompt Obfuscation for Privacy-Preserving Communication with Cloud-based LLMscited by 9
  3. 2025PICE: A Semantic-Driven Progressive Inference System for LLM Serving in Cloud-Edge Networkscited by 6
  4. 2025Graph Representation-based Model Poisoning on the Heterogeneous Internet of Agentscited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 10.

10 papers.

  1. 2026MemPrivacy: Privacy-Preserving Personalized Memory Management for Edge-Cloud Agentscited by 7
  2. 2026Conflict-Aware Client Selection for Multi-Server Federated Learningcited by 5
  3. 2026Lever: Speculative LLM Inference on Smartphonescited by 2
  4. 2026Seeing is Free, Speaking is Not: Uncovering the True Energy Bottleneck in Edge VLM Inferencecited by 2
  5. 2026Towards Privacy-Preserving Large Language Model: Text-free Inference Through Alignment and Adaptationcited by 1
  6. 2026Graph Representation Learning Augmented Model Manipulation on Federated Fine-Tuning of LLMscited by 1
  7. 2026P2Skill: Privacy Preserving Skill Distillation for Cloud-Local LLM Inference Systemscited by 0
  8. 2026Trains but Doesn't Learn: A Post-Training Delivery Benchmark for LLM Agents as Forward-Deployed Engineerscited by 0
  9. 2026Ask Without Telling: Local SLMs Consult Cloud LLMs Without Revealing Task Intentcited by 0
  10. 2026Learning to Refer: Client-Resolved Generation for Privacy-Aware Language Modelscited by 0