Preprint · 2025 · Open access

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

Junfei Zhan, Shen, Haoxun, Lin, Zheng, Tengjiao He

AAAI Publications (The Association for the Advancement of Artificial Intelligence (AAAI)) · Published 27 November 2025 · DOI 10.48550/arxiv.2511.22788

Open PDFView at the publisherCite this paper

In brief

“PRISM: Privacy-Aware Routing for Adaptive Cloud-Edge LLM Inference via Semantic Sketch Collaboration” is a 2025 preprint by Junfei Zhan and 3 co-authors, published in AAAI Publications (The Association for the Advancement of Artificial Intelligence (AAAI)). OpenAlex counts 16 citations as of 19 November 2026. It is open access under a CC BY licence: the full text is free to read at arXiv (Cornell University).

Citations
16
Most cited in
2026 (15)
References
29
Authors
4

Abstract

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.

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

Zhan, J., Haoxun, S., Zheng, L., & He, T. (2025). PRISM: Privacy-Aware Routing for Adaptive Cloud-Edge LLM Inference via Semantic Sketch Collaboration. AAAI Publications (The Association for the Advancement of Artificial Intelligence (AAAI)). https://doi.org/10.48550/arxiv.2511.22788

MLA 9

Zhan, Junfei, et al. “PRISM: Privacy-Aware Routing for Adaptive Cloud-Edge LLM Inference via Semantic Sketch Collaboration.” AAAI Publications (The Association for the Advancement of Artificial Intelligence (AAAI)), 2025. https://doi.org/10.48550/arxiv.2511.22788.

Chicago

Zhan, Junfei, Shen, Haoxun, Lin, Zheng, and Tengjiao He. 2025. “PRISM: Privacy-Aware Routing for Adaptive Cloud-Edge LLM Inference via Semantic Sketch Collaboration.” AAAI Publications (The Association for the Advancement of Artificial Intelligence (AAAI)). https://doi.org/10.48550/arxiv.2511.22788.

Harvard

Zhan, J. et al. (2025) ‘PRISM: Privacy-Aware Routing for Adaptive Cloud-Edge LLM Inference via Semantic Sketch Collaboration’, AAAI Publications (The Association for the Advancement of Artificial Intelligence (AAAI)). doi:10.48550/arxiv.2511.22788.

IEEE

J. Zhan, S. Haoxun, L. Zheng, and T. He, “PRISM: Privacy-Aware Routing for Adaptive Cloud-Edge LLM Inference via Semantic Sketch Collaboration,” AAAI Publications (The Association for the Advancement of Artificial Intelligence (AAAI)), 2025, doi: 10.48550/arxiv.2511.22788.

BibTeX

@misc{zhan2025prism,
  title     = {{PRISM: Privacy-Aware Routing for Adaptive Cloud-Edge LLM Inference via Semantic Sketch Collaboration}},
  author    = {Zhan, Junfei and Haoxun, Shen, and Zheng, Lin, and He, Tengjiao},
  howpublished = {AAAI Publications (The Association for the Advancement of Artificial Intelligence (AAAI))},
  year      = {2025},
  doi       = {10.48550/arxiv.2511.22788},
  url       = {https://doi.org/10.48550/arxiv.2511.22788}
}

Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)

Topic and keywords

Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data

  • large language model inference
  • cloud-edge collaboration
  • local differential privacy
  • privacy-utility trade-off
  • context-aware inference
  • soft gating
  • energy efficiency
  • latency reduction

Work with this paper

Most-cited papers in Computer Science

All highly cited Computer Science papers

Questions about this paper

How do I cite this paper in APA style?

Zhan, J., Haoxun, S., Zheng, L., & He, T. (2025). PRISM: Privacy-Aware Routing for Adaptive Cloud-Edge LLM Inference via Semantic Sketch Collaboration. AAAI Publications (The Association for the Advancement of Artificial Intelligence (AAAI)). https://doi.org/10.48550/arxiv.2511.22788

How many times has this paper been cited?

16 times, according to OpenAlex on 19 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.

Is this paper free to read?

Yes. It is open access at arXiv (Cornell University): sign in and open the PDF from this page.

Who wrote this paper?

Junfei Zhan, Shen, Haoxun, Lin, Zheng, Tengjiao He.

What is the DOI of this paper?

10.48550/arxiv.2511.22788, permanently at https://doi.org/10.48550/arxiv.2511.22788. Sign in and open it from this page.

Bibliographic data from OpenAlex (CC0), last checked 19 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.