PAAC: Privacy-Aware Agentic Device-Cloud Collaboration

• URL: http://arxiv.org/abs/2605.08646v1
• Date: Sat, 09 May 2026 03:29:20 GMT
• Title: PAAC: Privacy-Aware Agentic Device-Cloud Collaboration
• Abstract summary: We develop PAAC, a privacy-aware agentic framework that aligns planner--executor decomposition with the device-cloud boundary.<n>We find consistent improvements on 17 benchmarks spanning 10 domains, including math, science, and finance.
• Score: 27.486312939995145
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language model (LLM) agents face a structural tension: cloud agents provide strong reasoning but expose user data, while on-device agents preserve privacy at the cost of overall capability. Existing device-cloud designs treat this boundary as a compute split rather than a trust boundary suited to agentic workloads, and existing sanitizers force a choice between policy flexibility and the structural fidelity tool calls require. In this work, we develop PAAC, a privacy-aware agentic framework that aligns planner--executor decomposition with the device-cloud boundary so that role specialization itself becomes the privacy mechanism. The cloud agent reasons over typed placeholder tokens that preserve each sensitive value's reasoning role while discarding its content, while the on-device agent identifies sensitive spans and distills each step's execution outcome into compact key findings. Sanitization confines the on-device LLM to proposing which spans to mask, while a deterministic registry performs all substitution and reversal, keeping actions directly executable on device. On three agentic benchmarks under strict privacy settings, PAAC dominates the Pareto frontier of privacy and accuracy, improving average accuracy by 15-36\% and reducing average leakage by 2-6$\times$ over state-of-the-art device-cloud baselines, with the largest margins on privacy targets outside fixed entity taxonomies. We find consistent improvements on 17 additional benchmarks spanning 10 domains, including math, science, and finance.

Related papers

• OCELOT: Inference-Leakage Budgets for Privacy-Preserving LLM Agents [32.5466552510287]
  Leakage is cumulative, as individually innocuous releases accumulate across honest-but-curious or colluding sinks.<n>We present OCELOT, a runtime mediator that budgets how much an adversary's belief about a secret may improve.<n>OCELOT attains significantly lower leakage at higher task utility, resists adaptive injection, jailbreak, cumulative inference, and sink collusion.
  arXiv  Detail & Related papers  (2026-06-10T17:13:35Z)
• PRAG: End-to-End Privacy-Preserving Retrieval-Augmented Generation [15.502387220719813]
  Existing privacy-preserving solutions often sacrifice retrieval quality due to noise injection or only provide partial encryption.<n>We propose PRAG, an end-to-end privacy-preserving RAG system that achieves end-to-end confidentiality for both documents and queries.<n>Experiments on large-scale datasets demonstrate that PRAG achieves competitive recall (72.45%-74.45%), practical retrieval latency, and strong resilience against graph reconstruction attacks.
  arXiv  Detail & Related papers  (2026-04-29T10:46:45Z)
• SpecSteer: Synergizing Local Context and Global Reasoning for Efficient Personalized Generation [79.75755802397312]
  SpecSteer is an asymmetric collaborative inference framework that synergizes private on-device context with cloud-scale reasoning.<n>Experiments demonstrate that SpecSteer successfully closes the reasoning gap and achieves superior personalized generation performance.
  arXiv  Detail & Related papers  (2026-03-17T07:51:29Z)
• STAMP: Selective Task-Aware Mechanism for Text Privacy [5.453215632053115]
  STAMP is a new framework for task-aware text privatization that achieves an improved privacy-utility trade-off.<n>We introduce the polar mechanism, which perturbs only the direction of embeddings on the unit sphere while preserving their magnitude.<n> Experimental evaluations on SQuAD, Yelp, and AG News datasets demonstrate that STAMP, when combined with the normalized polar mechanism, consistently achieves superior privacy-utility trade-offs.
  arXiv  Detail & Related papers  (2026-03-12T17:55:07Z)
• SplitAgent: A Privacy-Preserving Distributed Architecture for Enterprise-Cloud Agent Collaboration [0.04061135251278187]
  Current agent frameworks assume complete data sharing, making them unsuitable for enterprise environments with confidential information.<n>We present SplitAgent, a novel distributed architecture that enables privacy-preserving collaboration between enterprise-side privacy agents and cloud-side reasoning agents.<n>Our key innovation is context-aware dynamic sanitization that adapts privacy protection based on task semantics.
  arXiv  Detail & Related papers  (2026-03-09T10:51:31Z)
• NeuroFilter: Privacy Guardrails for Conversational LLM Agents [50.75206727081996]
  This work addresses the computational challenge of enforcing privacy for agentic Large Language Models (LLMs)<n>NeuroFilter is a guardrail framework that operationalizes contextual integrity by mapping norm violations to simple directions in the model's activation space.<n>A comprehensive evaluation across over 150,000 interactions, covering models from 7B to 70B parameters, illustrates the strong performance of NeuroFilter.
  arXiv  Detail & Related papers  (2026-01-21T05:16:50Z)
• Efficient Privacy-Preserving Retrieval Augmented Generation with Distance-Preserving Encryption [25.87368479678027]
  RAG has emerged as a key technique for enhancing response quality of LLMs without high computational cost.<n>In traditional architectures, RAG services are provided by a single entity that hosts the dataset within a trusted local environment.<n>This dependence on untrusted third-party services introduces privacy risks.<n>We propose an efficient privacy-preserving RAG framework (ppRAG) tailored for untrusted cloud environments.
  arXiv  Detail & Related papers  (2026-01-18T09:29:50Z)
• PRISM: Privacy-Aware Routing for Adaptive Cloud-Edge LLM Inference via Semantic Sketch Collaboration [8.776463501718737]
  We propose a context-aware framework that dynamically balances privacy and inference quality.<n>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.
  arXiv  Detail & Related papers  (2025-11-27T22:32:33Z)
• DRIFT: Dynamic Rule-Based Defense with Injection Isolation for Securing LLM Agents [52.92354372596197]
  Large Language Models (LLMs) are increasingly central to agentic systems due to their strong reasoning and planning capabilities.<n>This interaction also introduces the risk of prompt injection attacks, where malicious inputs from external sources can mislead the agent's behavior.<n>We propose a Dynamic Rule-based Isolation Framework for Trustworthy agentic systems, which enforces both control and data-level constraints.
  arXiv  Detail & Related papers  (2025-06-13T05:01:09Z)
• PWC-MoE: Privacy-Aware Wireless Collaborative Mixture of Experts [59.5243730853157]
  Large language models (LLMs) hosted on cloud servers alleviate the computational and storage burdens on local devices but raise privacy concerns.<n>Small language models (SLMs) running locally enhance privacy but suffer from limited performance on complex tasks.<n>We propose a privacy-aware wireless collaborative mixture of experts (PWC-MoE) framework to balance computational cost, performance, and privacy protection under bandwidth constraints.
  arXiv  Detail & Related papers  (2025-05-13T16:27:07Z)
• Optimizing Cross-Client Domain Coverage for Federated Instruction Tuning of Large Language Models [87.49293964617128]
  Federated domain-specific instruction tuning (FedDIT) for large language models (LLMs) aims to enhance performance in specialized domains using distributed private and limited data.<n>We empirically establish that cross-client domain coverage, rather than data heterogeneity, is the pivotal factor.<n>We introduce FedDCA, an algorithm that explicitly maximizes this coverage through diversity-oriented client center selection and retrieval-based augmentation.
  arXiv  Detail & Related papers  (2024-09-30T09:34:31Z)
• Unified Mechanism-Specific Amplification by Subsampling and Group Privacy Amplification [54.1447806347273]
  Amplification by subsampling is one of the main primitives in machine learning with differential privacy. We propose the first general framework for deriving mechanism-specific guarantees. We analyze how subsampling affects the privacy of groups of multiple users.
  arXiv  Detail & Related papers  (2024-03-07T19:36:05Z)
• THE-X: Privacy-Preserving Transformer Inference with Homomorphic Encryption [112.02441503951297]
  Privacy-preserving inference of transformer models is on the demand of cloud service users. We introduce $textitTHE-X$, an approximation approach for transformers, which enables privacy-preserving inference of pre-trained models.
  arXiv  Detail & Related papers  (2022-06-01T03:49:18Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.