Pith. sign in

REVIEW 4 major objections 6 minor 66 references

The Game-Theoretic Symbiosis of Trust and AI in Networked Systems

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper reframes network trust as a dynamic, controllable interaction that AI can manage in real time.

desk verdict A readable, derivative book chapter that repackages the authors' prior game-theoretic trust work without new results or addressing the estimation problem at the core of its own formalization. read the letter →

arxiv 2411.12859 v1 pith:UGTVGO4S submitted 2024-11-19 cs.AI

Yunfei Ge , Quanyan Zhu This is my paper · ORCID
classification cs.AI
keywords trustmanagementgametheorycybersecurityzeroBayesianinferenceadversarialmachinelearningcyberdeceptionAIgovernance
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This chapter argues that trust in networked systems is not a fixed score to be estimated but a live interaction that attackers, defenders, and AI systems can deliberately shape. The authors' central proposal is to treat trust as a controllable variable, updated by Bayesian reasoning from prior knowledge, side evidence, and observed strategies, and to model the adversarial struggle around it with game theory. If this framing holds, trust management becomes a design problem: systems can continuously tune trust levels to support zero-trust access, expose cyber deception, counter misinformation, and harden AI against attacks. Because the same game-theoretic apparatus also models adversarial training and red-team exercises, AI both improves trust management and becomes more worthy of trust, forming a positive feedback loop.

What carries the argument

The load-bearing mechanism is the Bayesian trust update (Definition 2): $$TS_{t+1}(i) = \Pr(\$theta^{{t+1}}$_i \in \Theta_T \mid a^t, e^t, \pi^t) = \frac{h(e^t \mid a^t, \$\theta$^t_i \in \Theta_T)\,\$\sigma$(a^t \mid \$\theta$^t_i \in \Theta_T)\,\pi^t(\$\theta$^t_i \in \Theta_T)}{\sum_{\hat{\$\theta$}_i \in \Theta} h(e^t \mid a^t, \hat{\$\theta$}_i)\,\$\sigma$(a^t \mid \hat{\$\theta$}_i)\,\pi^t(\hat{\$\theta$}_i)}.$$ This recursion turns trust into a quantity that can be recomputed after every action and every security alert, with the prior feeding the posterior and the posterior becoming the next prior. Around this update, the paper assembles the game-theoretic machinery of Bayesian and signaling games for asymmetric information, Stackelberg games for sequential attack-defense moves, and min-max formulations for adversarial training. That combination is what makes trust simultaneously a mathematical object and a strategic weapon.

What would settle it

Simulate or log a real enterprise network with user actions and security alerts, feed them through the Bayesian trust update, and compare access decisions against a static-risk baseline; if the updated scores cannot be computed without hand-supplied $h$ and $\sigma$, or if they do not detect compromised accounts sooner than the baseline, the practical controllability claim is falsified.

Watch

Extended reading notes

Core claim

The central claim is that a network's trust score $TS_t(i)$, defined as the probability that entity $i$ is non-adversarial at time $t$, should be an endogenous, strategically updated quantity rather than an exogenous, static attribute. The paper's formal engine is a recursive Bayesian update that computes the next trust score from a prior $\pi^t$, the entity's observed action $a^t$, third-party side evidence $e^t$, and the entity's observed strategy $\sigma(a^t \mid \theta^t)$. Because the updating rule makes trust depend on inputs the system chooses (priors, evidence channels, incentives) and inputs the adversary chooses (actions, deceptive signals), trust becomes controllable from both sides of the interaction. Game theory then provides the equilibrium concepts — Bayesian Nash equilibrium, perfect Bayesian equilibrium, and Stackelberg games — for predicting how these strategic pushes and pulls settle. On the AI side, adversarial training is cast as a min-max game and red/blue teaming as a repeated strategic interaction, so the same formal toolkit that manages trust also strengthens AI systems, closing the loop between AI-enabled trust and trust in AI.

Load-bearing premise

The framework's Bayesian update requires the system to already know the likelihood functions $h$ (how side evidence arises from actions and types) and $\sigma$ (how each type behaves), but real networks would have to estimate those, and the paper gives no procedure or empirical check.

Editorial extensions

If this is right

  • Zero-trust access control becomes a sequential Bayesian decision problem: every login, action, and alert revises the trust score, so access is continuously re-evaluated instead of granted once.
  • Deceptive defenses such as honeypots can be designed as signaling games, where the defender selects signals to manipulate the attacker's updated beliefs about what is genuine.
  • Adversarial training of AI models is formally a two-player zero-sum game, so its convergence, equilibrium, and robustness properties can be analyzed with game-theoretic tools.
  • Governance of AI becomes a meta-game: policies that make AI more transparent and accountable raise trust in AI, which accelerates AI adoption and thereby improves AI-driven trust management.
  • If the positive feedback loop holds, organizations can reach a stable equilibrium where AI-supported network defense and user confidence in AI grow together rather than one lagging the other.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The Bayesian update assumes known likelihood functions $h$ and $\sigma$; a natural next step the paper does not take is to estimate these from network logs, turning trust management into a learning problem rather than a purely analytical one.
  • The trust-score definition suggests a calibration test: if trust scores were well calibrated, the probability an entity is non-adversarial would match the observed fraction of non-adversarial entities at each score, a property that could be checked on real access-control data.
  • The same strategic-trust framing could be extended to human-AI teams, where a user's trust in an AI recommendation is updated by observed AI actions and side evidence, making trust a measurable input to human-machine decision making.
  • A concrete testable prediction follows: systems that explicitly maintain Bayesian trust scores with side evidence should detect compromised accounts or insider threats faster than static-risk systems; a simulated or historical dataset of alerts and access logs could test this directly.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. This manuscript (arXiv:2411.12859) is a book chapter that proposes reframing trust in networked systems as a dynamic, controllable, game-theoretic interaction rather than a static variable. It defines a trust score and a Bayesian trust update (Definition 2), reviews policy-based and reputation-based trust management, connects trust evaluation to Bayesian games and signaling games, and discusses adversarial training and red/blue teaming for AI trustworthiness. A case study of an AI-driven traffic management system is used to illustrate these ideas. The chapter also argues for a symbiotic, mutually reinforcing relationship between AI and trust and calls for governance frameworks to sustain a positive equilibrium.

Significance. If the central reframing is made rigorous and validated, it could provide a useful unifying perspective: trust management becomes a game-theoretic control problem that AI can optimize, and AI trustworthiness becomes a strategic variable. The chapter gives a competent survey of relevant background and correctly states standard Bayesian update and equilibrium formulas. However, the novel parts are not supported by formal derivation or empirical evidence. The Bayesian trust update is the only substantial formalization of the central claim, and it presumes knowledge of likelihood functions that are typically unknown in adversarial settings. The case study is explicitly hypothetical and does not instantiate the proposed update or any equilibrium concept. The manuscript therefore reads as a position chapter whose central claims remain untested, though the underlying ideas are plausible and potentially valuable.

major comments (4)
  1. [Section 3.1.3, Definition 2 (Eq. 2)] The Bayesian trust update requires the system to know the evidence-generating function h(e^t|a^t,θ) and the opponent's observed strategy σ(a^t|θ), but the chapter provides no estimation procedure, no identifiability conditions, and no analysis of the consequences of misspecifying these quantities. In adversarial networks these functions are strategically manipulated, so the update cannot be applied as stated. This is load-bearing because Definition 2 is the only formalization of the paper's central claim that trust is dynamically controllable.
  2. [Section 4.2.1] The traffic management case study is a hypothetical, simulated scenario that never instantiates Definition 2 or any game-theoretic solution concept (BNE, PBE, etc.). It asserts that red and blue teams optimize their strategies through game theory and that foundation models detect anomalies, but no data, no attack success metrics, no baseline comparisons, and no security improvement measure are provided. The case study is therefore illustrative only and cannot support the claim that the proposed framework improves security outcomes.
  3. [Section 2 and Figure 2] The claimed 'symbiotic relationship' or 'meta-game' between AI and trust is described only verbally; no game model, payoff functions, action spaces, or equilibrium concept are specified for this meta-game. As a result, the positive feedback loop in Figure 2 is an assumption rather than a derived result. The chapter needs either a formal model of this meta-game or an explicit statement that the symbiosis is a conceptual framing rather than a technical contribution.
  4. [Section 3.2] The exposition of Bayesian games, BNE, signaling games, and PBE reproduces standard textbook formulas, but it does not connect these equilibrium concepts to the proposed Bayesian trust update in Definition 2. The link between the trust score update and the players' equilibrium strategies is asserted rather than demonstrated, so the paper falls short of delivering a new game-theoretic trust evaluation framework as promised.
minor comments (6)
  1. [Section 1.2] The sentence 'we introduce the of trust as a strategic interaction' is missing a noun; it should be 'the notion of trust' or similar.
  2. [Section 3.1.1] In 'we expands the target of trust', the verb should agree with the subject: 'we expand'.
  3. [Section 3.1.3] The sentence ending 'more vulnerable to false positives or false negatives. evaluation.' contains a stray period and an incomplete phrase; it should be rewritten.
  4. [Section 3.1.4] The phrase 'In this thesis' should be 'In this chapter', since the work is presented as a book chapter.
  5. [Section 4.1.3] There is an extra space before the period in 'guarantee robust solutions .'
  6. [Section 3.2] The phrase 'the of strategic cyber risk' is missing a noun; it should be 'the concept of strategic cyber risk' or similar.

Circularity Check

2 steps flagged · score 6.0 of 10

The formal trust-update equation restates the definition of trust score by Bayes' rule, and the central reframing of trust leans on the first author's own thesis; both moves are self-referential.

  1. self definitional [Section 3.1.3, Definition 2 (Eq. 2)]
    "Definition 2 (Bayesian Trust Update). The Trust Score (TS) of the entity i at time t+1 is the probability that the entity is non-adversarial (θ^{t+1}_i ∈ Θ_T) based on the prior knowledge, side evidence, and observed strategies of the entity: TS^{t+1}(i) = Pr(θ^{t+1}_i ∈ Θ_T | a^t, e^t, π^t) = ... (2) ... Note that the relationship between TS and π is: TS^t(i) = π^t(θ^t_i ∈ Θ_T)."

    Definition 1 already defines TS^t(i) := Pr(θ^t_i ∈ Θ_T). Equation (2) then sets TS^{t+1}(i) to the posterior probability of exactly the same event, so the proposed 'Bayesian trust update' is the definition of TS together with Bayes' theorem, not an independent model-derived prediction. The likelihood h and observed strategy σ are assumed inputs with no estimation procedure, so the posterior is determined by construction from those inputs. The formal core of the chapter's trust model therefore reduces to a definitional identity rather than an independently derived result.

  2. self citation load bearing [Section 1.1, sentence on deception and trust]
    "Trust lies at the heart of deception. To fully understand deception in networked systems, it is essential to examine trust not just as a static or random variable but as a strategic interaction that can be shaped, manipulated, and exploited [8]."

    This sentence is the seed of the chapter's central contribution, stated in Section 1.2 as 'We propose reframing trust as a dynamic and controllable interaction, rather than a static variable.' The only support offered for this reframing is reference [8], which is the first author's own PhD thesis: 'Yunfei Ge. The Symbiosis of Trust and AI: Scientific Foundations...' No external proof, benchmark, or independent argument is given. The central premise thus rests on a self-citation that is itself unverified in the chapter, making the reframing load-bearing on the authors' prior work rather than on independent evidence.

full rationale

The chapter is a position/review rather than a quantitative derivation, so much of its content consists of literature review and citations to the authors' earlier frameworks (GAZETA, MUFAZA, ADAPT, PenHeal, trust-threshold policy). Those citations are frequent but they are mostly used as examples rather than as mathematical premises. The two genuinely circular moves are the formal trust-update equation and the load-bearing citation for the central reframing. Definition 1 defines the trust score as a probability, and Definition 2's update is exactly the conditional probability of that same event under Bayes' rule; no estimation of h or σ is supplied, so the only formal 'model' in the chapter is equivalent to its own definition by construction. Separately, the chapter's strongest claim—trust should be viewed as a dynamic, controllable strategic interaction—is introduced with the sole citation [8], the first author's PhD thesis, so the central premise leans on the authors' own unverified prior claim. I do not score this as 8 or 10 because the chapter also engages external material (MITRE ATLAS, adversarial training literature, OWASP/CVE references), and because the self-cited prior frameworks are not used to force a concrete numerical prediction in this text. The hypothetical traffic-management case study adds no new empirical evidence. The score of 6 reflects partial circularity: the formal core reduces by construction and the central reframing is supported by a self-citation chain, while the surrounding review content retains independent content.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The central claim rests on several unverified modeling choices. The trust score definition is a probability, the Bayesian update requires known likelihood functions, and the existence of a virtuous cycle between AI and trust is assumed rather than demonstrated.

assumptions (4)
  • domain assumption Trust score is defined as the probability that an entity's attributes lie in a non-adversarial set (Definition 1).
    This operationalization assumes trust can be captured by a single probability, which is a strong simplification.
  • domain assumption The Bayesian update (Equation 2) assumes known likelihood functions h and sigma for side evidence and observed strategies.
    In practical settings these likelihoods are rarely known and are not estimated in the chapter.
  • domain assumption Entities are strategic and rational, as in game-theoretic models (Section 3.2).
    The chapter assumes that both attackers and defenders optimize expected utility, which may not reflect bounded rationality in real networks.
  • ad hoc to paper A positive feedback loop between AI and trust exists (Section 2, Figure 2).
    This is the central thesis of the chapter and is asserted rather than demonstrated with empirical evidence.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The Game-Theoretic Symbiosis of Trust and AI in Networked Systems." pith.science (2026). https://pith.science/paper/UGTVGO4S

@misc{pith2026241112859,
  author       = {Pith},
  title        = {Pith review of: The Game-Theoretic Symbiosis of Trust and AI in Networked Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UGTVGO4S}},
  note         = {Machine review of arXiv:2411.12859}
}
read the original abstract

This chapter explores the symbiotic relationship between Artificial Intelligence (AI) and trust in networked systems, focusing on how these two elements reinforce each other in strategic cybersecurity contexts. AI's capabilities in data processing, learning, and real-time response offer unprecedented support for managing trust in dynamic, complex networks. However, the successful integration of AI also hinges on the trustworthiness of AI systems themselves. Using a game-theoretic framework, this chapter presents approaches to trust evaluation, the strategic role of AI in cybersecurity, and governance frameworks that ensure responsible AI deployment. We investigate how trust, when dynamically managed through AI, can form a resilient security ecosystem. By examining trust as both an AI output and an AI requirement, this chapter sets the foundation for a positive feedback loop where AI enhances network security and the trust placed in AI systems fosters their adoption.

Figures

Figures reproduced from arXiv: 2411.12859 by the authors.

Figure 1
Figure 1. Trust is integral to all stages of a networked system. A trustworthy network [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The symbiotic relationship between AI and Trust forms a cyber trust ecosystem, [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

66 extracted references · 59 canonical work pages

  1. [1]

    The role of machine learning in cybersecurity

    Giovanni Apruzzese, Pavel Laskov, Edgardo Montes de Oca, Wissam Mallouli, Luis Brdalo Rapa, Athanasios Vasileios Grammatopoulos, and Fabio Di Franco. The role of machine learning in cybersecurity. Digital Threats: Research and Practice, 4(1):1–38, 2023

    work page 2023

  2. [2]

    An integration of reputation-based and policy-based trust management

    Piero Bonatti, Claudiu Duma, Daniel Olmedilla, and Nahid Shahmehri. An integration of reputation-based and policy-based trust management. In W9: The Semantic Web and Policy Workshop (SWPW), page 136. Citeseer, 2007

    work page 2007

  3. [3]

    Stackelberg games for adversarial prediction problems

    Michael Br¨ uckner and Tobias Scheffer. Stackelberg games for adversarial prediction problems. In Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining, pages 547–555, 2011

    work page 2011

  4. [4]

    A linear quadratic differential game approach to dynamic contract design for systemic cyber risk management under asymmetric information

    Juntao Chen and Quanyan Zhu. A linear quadratic differential game approach to dynamic contract design for systemic cyber risk management under asymmetric information. In 2018 56th Annual Allerton Conference on Communication, Control, and Computing (Allerton), pages 575–582. IEEE, 2018

    work page 2018

  5. [5]

    A survey on trust modeling

    Jin-Hee Cho, Kevin Chan, and Sibel Adali. A survey on trust modeling. ACM Computing Surveys (CSUR), 48(2):1–40, 2015

    work page 2015

  6. [6]

    Network slicing in 5g: Survey and challenges

    Xenofon Foukas, Georgios Patounas, Ahmed Elmokashfi, and Mahesh K Marina. Network slicing in 5g: Survey and challenges. IEEE communications magazine, 55(5):94–100, 2017

    work page 2017

  7. [7]

    Facid: A trust-based collaborative decision framework for intrusion detection networks

    Carol J Fung and Quanyan Zhu. Facid: A trust-based collaborative decision framework for intrusion detection networks. Ad Hoc Networks, 53:17–31, 2016

    work page 2016

  8. [8]

    The Symbiosis of Trust and AI: Scientific Foundations for Strategic Network Security, Autonomous Resilience, and Prescriptive Governance

    Yunfei Ge. The Symbiosis of Trust and AI: Scientific Foundations for Strategic Network Security, Autonomous Resilience, and Prescriptive Governance. PhD thesis, New York University Tandon School of Engineering, 2024

    work page 2024

Show all 66 references
  1. [9]

    Accountability and insurance in iot supply chain

    Yunfei Ge and Quanyan Zhu. Accountability and insurance in iot supply chain. arXiv preprint arXiv:2201.11855, 2022. The Game-Theoretic Symbiosis of Trust and AI in Networked Systems 25

    2022 arXiv

  2. [10]

    Mufaza: Multi-source fast and autonomous zero- trust authentication for 5g networks

    Yunfei Ge and Quanyan Zhu. Mufaza: Multi-source fast and autonomous zero- trust authentication for 5g networks. In MILCOM 2022-2022 IEEE Military Communications Conference (MILCOM), pages 571–576. IEEE, 2022

    2022

  3. [11]

    Trust threshold policy for explainable and adaptive zero-trust defense in enterprise networks

    Yunfei Ge and Quanyan Zhu. Trust threshold policy for explainable and adaptive zero-trust defense in enterprise networks. In 2022 IEEE Conference on Communications and Network Security (CNS), pages 359–364. IEEE, 2022

    2022

  4. [12]

    Ai liability insurance with an example in ai-powered e-diagnosis system

    Yunfei Ge and Quanyan Zhu. Ai liability insurance with an example in ai-powered e-diagnosis system. arXiv preprint arXiv:2306.01149, 2023

    2023 arXiv

  5. [13]

    Gazeta: Game-theoretic zero-trust authentication for defense against lateral movement in 5g iot networks

    Yunfei Ge and Quanyan Zhu. Gazeta: Game-theoretic zero-trust authentication for defense against lateral movement in 5g iot networks. IEEE Transactions on Information Forensics and Security, 2023

    2023

  6. [14]

    Zero trust for cyber resilience

    Yunfei Ge and Quanyan Zhu. Zero trust for cyber resilience. arXiv preprint arXiv:2312.02882, 2023

    2023 arXiv

  7. [15]

    Attributing responsibility in ai-induced incidents: A computational reflective equilibrium framework for accountability

    Yunfei Ge and Quanyan Zhu. Attributing responsibility in ai-induced incidents: A computational reflective equilibrium framework for accountability. arXiv preprint arXiv:2404.16957, 2024

    2024

  8. [16]

    Mega-pt: A meta-game framework for agile penetration testing

    Yunfei Ge and Quanyan Zhu. Mega-pt: A meta-game framework for agile penetration testing. In International Conference on Decision and Game Theory for Security, pages 24–44. Springer, 2024

    2024

  9. [17]

    Scenario-agnostic zero-trust defense with explainable threshold policy: A meta-learning approach

    Yunfei Ge, Tao Li, and Quanyan Zhu. Scenario-agnostic zero-trust defense with explainable threshold policy: A meta-learning approach. arXiv preprint arXiv:2303.03349, 2023

    2023 arXiv

  10. [18]

    Industry 4.0, digitization, and opportunities for sustain- ability

    Morteza Ghobakhloo. Industry 4.0, digitization, and opportunities for sustain- ability. Journal of cleaner production, 252:119869, 2020

    2020

  11. [19]

    Penheal: A two-stage llm framework for automated pentesting and optimal remediation.arXiv preprint arXiv:2407.17788, 2024

    Junjie Huang and Quanyan Zhu. Penheal: A two-stage llm framework for automated pentesting and optimal remediation.arXiv preprint arXiv:2407.17788, 2024

    2024 arXiv

  12. [20]

    Adaptive strategic cyber defense for advanced persistent threats in critical infrastructure networks

    Linan Huang and Quanyan Zhu. Adaptive strategic cyber defense for advanced persistent threats in critical infrastructure networks. ACM SIGMETRICS Performance Evaluation Review, 46(2):52–56, 2019

    2019

  13. [21]

    Dynamic bayesian games for adversarial and defensive cyber deception

    Linan Huang and Quanyan Zhu. Dynamic bayesian games for adversarial and defensive cyber deception. Autonomous Cyber Deception: Reasoning, Adaptive Planning, and Evaluation of HoneyThings, pages 75–97, 2019

    2019

  14. [22]

    A dynamic games approach to proactive defense strategies against advanced persistent threats in cyber-physical systems

    Linan Huang and Quanyan Zhu. A dynamic games approach to proactive defense strategies against advanced persistent threats in cyber-physical systems. Computers & Security, 89:101660, 2020

    2020

  15. [23]

    Advert: Defending against reactive attention attacks

    Linan Huang and Quanyan Zhu. Advert: Defending against reactive attention attacks. In Cognitive Security: A System-Scientific Approach , pages 67–83. Springer, 2023

    2023

  16. [24]

    Game theory and machine learning for cyber security

    Charles A Kamhoua, Christopher D Kiekintveld, Fei Fang, and Quanyan Zhu. Game theory and machine learning for cyber security . John Wiley & Sons, 2021

    2021

  17. [25]

    Adversarial self-supervised contrastive learning

    Minseon Kim, Jihoon Tack, and Sung Ju Hwang. Adversarial self-supervised contrastive learning. Advances in neural information processing systems, 33: 2983–2994, 2020. 26 Yunfei Ge and Quanyan Zhu

    2020

  18. [26]

    Human-in-the-loop cyber intrusion detection using active learning

    Yeongwoo Kim, Gy¨orgy D´an, and Quanyan Zhu. Human-in-the-loop cyber intrusion detection using active learning. IEEE Transactions on Information Forensics and Security, 2024

    2024

  19. [27]

    The real story of stuxnet

    David Kushner. The real story of stuxnet. ieee Spectrum, 50(3):48–53, 2013

    2013

  20. [28]

    Commitment with signaling under double-sided information asymmetry

    Tao Li and Quanyan Zhu. Commitment with signaling under double-sided information asymmetry. arXiv preprint arXiv:2212.11446, 2022

    2022 arXiv

  21. [29]

    On the price of transparency: A comparison between overt persuasion and covert signaling

    Tao Li and Quanyan Zhu. On the price of transparency: A comparison between overt persuasion and covert signaling. In 2023 62nd IEEE Conference on Decision and Control (CDC), pages 4267–4272. IEEE, 2023

    2023

  22. [30]

    Symbiotic game and foundation models for cyber de- ception operations in strategic cyber warfare

    Tao Li and Quanyan Zhu. Symbiotic game and foundation models for cyber de- ception operations in strategic cyber warfare. arXiv preprint arXiv:2403.10570, 2024

    2024 arXiv

  23. [31]

    The role of information structures in game-theoretic multi-agent learning

    Tao Li, Yuhan Zhao, and Quanyan Zhu. The role of information structures in game-theoretic multi-agent learning. Annual Reviews in Control, 53:296–314, 2022

    2022

  24. [32]

    Conjectural online learning with first-order beliefs in asymmetric information stochastic games

    Tao Li, Kim Hammar, Rolf Stadler, and Quanyan Zhu. Conjectural online learning with first-order beliefs in asymmetric information stochastic games. arXiv preprint arXiv:2402.18781, 2024

    2024 arXiv

  25. [33]

    Meta stackelberg game: Robust federated learning against adaptive and mixed poisoning attacks

    Tao Li, Henger Li, Yunian Pan, Tianyi Xu, Zizhan Zheng, and Quanyan Zhu. Meta stackelberg game: Robust federated learning against adaptive and mixed poisoning attacks. arXiv preprint arXiv:2410.17431, 2024

    2024 arXiv

  26. [34]

    On the role of risk perceptions in cyber insurance contracts

    Shutian Liu and Quanyan Zhu. On the role of risk perceptions in cyber insurance contracts. In 2022 IEEE Conference on Communications and Network Security (CNS), pages 377–382. IEEE, 2022

    2022

  27. [35]

    Stackelberg risk preference design.Mathematical Programming, pages 1–39, 2024

    Shutian Liu and Quanyan Zhu. Stackelberg risk preference design.Mathematical Programming, pages 1–39, 2024

    2024

  28. [36]

    Game theory meets network security and privacy

    Mohammad Hossein Manshaei, Quanyan Zhu, Tansu Alpcan, Tamer Bacs ¸ar, and Jean-Pierre Hubaux. Game theory meets network security and privacy. ACM Computing Surveys (CSUR), 45(3):1–39, 2013

    2013

  29. [37]

    A survey of cybersecurity certification for the internet of things

    Sara N Matheu, Jose L Hernandez-Ramos, Antonio F Skarmeta, and Gianmarco Baldini. A survey of cybersecurity certification for the internet of things. ACM Computing Surveys (CSUR), 53(6):1–36, 2020

    2020

  30. [38]

    Security information and event management (SIEM) implemen- tation

    David R Miller. Security information and event management (SIEM) implemen- tation. McGraw-Hill Higher Education, 2011

    2011

  31. [39]

    Mitre atlas framework, 2024

    MITRE Corporation. Mitre atlas framework, 2024. URL https://atlas. mitre.org. Accessed: 2024-11-17

    2024

  32. [40]

    The convergence of it and ot in critical infrastructure

    Glenn Murray, Michael N Johnstone, and Craig Valli. The convergence of it and ot in critical infrastructure. 2017

    2017

  33. [41]

    Kerberos: An authentication service for computer networks

    B Clifford Neuman and Theodore Ts’o. Kerberos: An authentication service for computer networks. IEEE Communications magazine, 32(9):33–38, 1994

    1994

  34. [42]

    Deception by design: evidence-based signaling games for network defense

    Jeffrey Pawlick and Quanyan Zhu. Deception by design: evidence-based signaling games for network defense. arXiv preprint arXiv:1503.05458, 2015

    2015 arXiv

  35. [43]

    A stackelberg game perspective on the conflict between machine learning and data obfuscation

    Jeffrey Pawlick and Quanyan Zhu. A stackelberg game perspective on the conflict between machine learning and data obfuscation. In 2016 IEEE International The Game-Theoretic Symbiosis of Trust and AI in Networked Systems 27 Workshop on Information Forensics and Security (WIFS) ...

    2016

  36. [44]

    Strategic trust in cloud-enabled cyber- physical systems with an application to glucose control

    Jeffrey Pawlick and Quanyan Zhu. Strategic trust in cloud-enabled cyber- physical systems with an application to glucose control. IEEE Transactions on Information Forensics and Security, 12(12):2906–2919, 2017

    2017

  37. [45]

    Game Theory for Cyber Deception: From Theory to Applications

    Jeffrey Pawlick and Quanyan Zhu. Game Theory for Cyber Deception: From Theory to Applications. Springer Nature, 2021

    2021

  38. [46]

    istrict: An interdependent strategic trust mechanism for the cloud-enabled internet of controlled things

    Jeffrey Pawlick, Juntao Chen, and Quanyan Zhu. istrict: An interdependent strategic trust mechanism for the cloud-enabled internet of controlled things. IEEE Transactions on Information Forensics and Security, 14(6):1654–1669, 2018

    2018

  39. [47]

    A game-theoretic taxon- omy and survey of defensive deception for cybersecurity and privacy

    Jeffrey Pawlick, Edward Colbert, and Quanyan Zhu. A game-theoretic taxon- omy and survey of defensive deception for cybersecurity and privacy. ACM Computing Surveys (CSUR), 52(4):1–28, 2019

    2019

  40. [48]

    Game theory for security and risk management

    Stefan Rass and Stefan Schauer. Game theory for security and risk management. Springer International Publishing. doi, 10:978–3, 2018

    2018

  41. [49]

    Bounded rationality

    Stefan Rass, Stefan Schauer, Sandra K¨onig, Quanyan Zhu, Stefan Rass, Stefan Schauer, Sandra K¨onig, and Quanyan Zhu. Bounded rationality. Cyber-Security in Critical Infrastructures: A Game-Theoretic Approach, pages 99–114, 2020

    2020

  42. [50]

    Nonconvex min-max optimization: Applications, challenges, and recent theoretical advances

    Meisam Razaviyayn, Tianjian Huang, Songtao Lu, Maher Nouiehed, Maziar Sanjabi, and Mingyi Hong. Nonconvex min-max optimization: Applications, challenges, and recent theoretical advances. IEEE Signal Processing Magazine, 37(5):55–66, 2020

    2020

  43. [51]

    Towards a unified game- theoretic view of adversarial perturbations and robustness

    Jie Ren, Die Zhang, Yisen Wang, Lu Chen, Zhanpeng Zhou, Yiting Chen, Xu Cheng, Xin Wang, Meng Zhou, Jie Shi, et al. Towards a unified game- theoretic view of adversarial perturbations and robustness. Advances in Neural Information Processing Systems, 34:3797–3810, 2021

    2021

  44. [52]

    Towards trustworthy internet of things: A survey on trust management applications and schemes

    Avani Sharma, Emmanuel S Pilli, Arka P Mazumdar, and Poonam Gera. Towards trustworthy internet of things: A survey on trust management applications and schemes. Computer Communications, 160:475–493, 2020

    2020

  45. [53]

    Zero trust architecture.NIST Special Publication, 800:207, 2020

    V A Stafford. Zero trust architecture.NIST Special Publication, 800:207, 2020

    2020

  46. [54]

    Trust modeling and management: from social trust to digital trust

    Zheng Yan and Silke Holtmanns. Trust modeling and management: from social trust to digital trust. In Computer security, privacy and politics: current issues, challenges and solutions, pages 290–323. IGI Global, 2008

    2008

  47. [55]

    Designing policies for truth: Com- bating misinformation with transparency and information design

    Ya-Ting Yang, Tao Li, and Quanyan Zhu. Designing policies for truth: Com- bating misinformation with transparency and information design. In 2023 21st International Symposium on Modeling and Optimization in Mobile, Ad Hoc, and Wireless Networks (WiOpt), pages 127–134. IEEE, 2023

    2023

  48. [56]

    Faster authenticated key agreement with perfect forward secrecy for industrial internet-of-things

    Zheng Yang, Jun He, Yangguang Tian, and Jianying Zhou. Faster authenticated key agreement with perfect forward secrecy for industrial internet-of-things. IEEE Transactions on Industrial Informatics, 16(10):6584–6596, 2019

    2019

  49. [57]

    Towards a theoretical framework of out-of-distribution generalization

    Haotian Ye, Chuanlong Xie, Tianle Cai, Ruichen Li, Zhenguo Li, and Liwei Wang. Towards a theoretical framework of out-of-distribution generalization. Advances in Neural Information Processing Systems, 34:23519–23531, 2021. 28 Yunfei Ge and Quanyan Zhu

    2021

  50. [58]

    Security of distributed machine learning: A game-theoretic approach to design secure dsvm

    Rui Zhang and Quanyan Zhu. Security of distributed machine learning: A game-theoretic approach to design secure dsvm. Adversary-Aware Learning Techniques and Trends in Cybersecurity, pages 17–36, 2021

    2021

  51. [59]

    A dual perturbation approach for differential private admm-based distributed empirical risk minimization

    Tao Zhang and Quanyan Zhu. A dual perturbation approach for differential private admm-based distributed empirical risk minimization. In Proceedings of the 2016 ACM Workshop on Artificial Intelligence and Security, pages 129–137, 2016

    2016

  52. [60]

    Dynamic differential privacy for admm-based distributed classification learning

    Tao Zhang and Quanyan Zhu. Dynamic differential privacy for admm-based distributed classification learning. IEEE Transactions on Information Forensics and Security, 12(1):172–187, 2016

    2016

  53. [61]

    Hypothesis testing game for cyber deception

    Tao Zhang and Quanyan Zhu. Hypothesis testing game for cyber deception. In Decision and Game Theory for Security: 9th International Conference, GameSec 2018, Seattle, WA, USA, October 29–31, 2018, Proceedings 9, pages 540–555. Springer, 2018

    2018

  54. [62]

    Informational design of dynamic multi-agent system

    Tao Zhang and Quanyan Zhu. Informational design of dynamic multi-agent system. arXiv preprint arXiv:2105.03052, 2021

    2021 arXiv

  55. [63]

    Game-theoretic analysis of cyber deception: Evidence-based strategies and dynamic risk mit- igation

    Tao Zhang, Linan Huang, Jeffrey Pawlick, and Quanyan Zhu. Game-theoretic analysis of cyber deception: Evidence-based strategies and dynamic risk mit- igation. Modeling and Design of secure Internet of Things , pages 27–58, 2020

    2020

  56. [64]

    Foundations of cyber resilience: The confluence of game, control, and learning theories

    Quanyan Zhu. Foundations of cyber resilience: The confluence of game, control, and learning theories. arXiv preprint arXiv:2404.01205, 2024

    2024 arXiv

  57. [65]

    Game theory meets network security: A tutorial

    Quanyan Zhu and Stefan Rass. Game theory meets network security: A tutorial. In Proceedings of the 2018 ACM SIGSAC conference on computer and communications security, pages 2163–2165, 2018

    2018

  58. [66]

    Guidex: A game-theoretic incentive-based mechanism for intrusion detection networks

    Quanyan Zhu, Carol Fung, Raouf Boutaba, and Tamer Basar. Guidex: A game-theoretic incentive-based mechanism for intrusion detection networks. IEEE Journal on Selected Areas in Communications, 30(11):2220–2230, 2012

    2012

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.