mlsanityData to 7 Oct 2026

Beyond Local vs. External: A Game-Theoretic Framework for Trustworthy Knowledge Acquisition

arXiv 2604.23413 ↗First posted 25 Apr 2026cs.CLcited by 0

On the map: LLM reasoning → Reasoning for agents, planning and code

Abstract (arXiv)

Cloud-hosted Large Language Models (LLMs) offer unmatched reasoning capabilities and dynamic knowledge, yet submitting raw queries to these external services risks exposing sensitive user intent. Conversely, relying exclusively on trusted local models preserves privacy but often compromises answer quality due to limited parameter scale and knowledge. To resolve this dilemma, we propose Game-theoretic Trustworthy Knowledge Acquisition (GTKA), a framework that formulates the trade-off between knowledge utility and privacy as a strategic game. GTKA consists of three components: (i) a privacy-aware sub-query generator that decomposes sensitive intent into generalized, low-risk fragments; (ii) an adversarial reconstruction attacker that attempts to infer the original query from these fragments, providing adaptive leakage signals; and (iii) a trusted local integrator that synthesizes external responses within a secure boundary. By training the generator and attacker in an alternating adversarial manner, GTKA optimizes the sub-query generation policy to maximize knowledge acquisition accuracy while minimizing the reconstructability of the original sensitive intent. To validate our approach, we construct two sensitive-domain benchmarks in the biomedical and legal fields. Extensive experiments demonstrate that GTKA significantly reduces intent leakage compared to state-of-the-art baselines while maintaining high-fidelity answer quality.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning. Builds on 15 of its 39 references that are in the map. First posted on arXiv 25 Apr 2026.

Builds on

15 of 39 references.

References · 15

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,160
  2. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  3. 2014Explaining and Harnessing Adversarial Examplescited by 7,287
  4. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  5. 2016Deep Learning with Differential Privacycited by 2,682
  6. 2020Threats to Federated Learning: A Surveycited by 171
  7. 2021PPFL: Privacy-preserving Federated Learning with Trusted Execution Environmentscited by 93
  8. 2023Locally Differentially Private Document Generation Using Zero Shot Promptingcited by 63
  9. 2023Protecting User Privacy in Remote Conversational Systems: A Privacy-Preserving framework based on text sanitizationcited by 23
  10. 2020End-to-End Learning and Intervention in Gamescited by 19
  11. 2024Meta Stackelberg Game: Robust Federated Learning against Adaptive and Mixed Poisoning Attackscited by 8
  12. 2024Privacy-Preserving Language Model Inference with Instance Obfuscationcited by 6
3 more
  1. 2023Automating Knowledge Acquisition for Content-Centric Cognitive Agents Using LLMscited by 4
  2. 2016Machine Learning Techniques for Stackelberg Security Games: a Surveycited by 3
  3. 2024Learning from Response not Preference: A Stackelberg Approach for LLM Detoxification using Non-parallel Datacited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026P2Skill: Privacy Preserving Skill Distillation for Cloud-Local LLM Inference Systems
  2. 2026From Leaky Thoughts to Private Reasoning: Controlling What LRMs Say to Themselves
  3. 2025Ensemble Privacy Defense for Knowledge-Intensive LLMs against Membership Inference Attacks
  4. 2026LLMs as Linguistic Chameleons: Decoupling Semantics and Structure for Privacy-Preserving Communication
  5. 2025Fine-Grained Privacy Extraction from Retrieval-Augmented Generation Systems via Knowledge Asymmetry Exploitation
  6. 2026Security and Privacy in Retrieval-Augmented Generation: Architectures, Threats, Defenses, and Future Directions for Building Trustworthy Systems
  7. 2026Learning to Refer: Client-Resolved Generation for Privacy-Aware Language Models
  8. 2023Can LLMs Keep a Secret? Testing Privacy Implications of Language Models via Contextual Integrity Theory