Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Game Theory Meets LLM and Agentic AI: Reimagining Cybersecurity for the Age of Intelligent Threats

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read This chapter argues that the arrival of large-language-model agents changes the natural locus of strategy in cybersecurity games: instead of choosing actions, an agent chooses a reasoning prompt, and the model's output distribution over…

desk verdict Readable survey with a flawed core: the LLM-Nash example is arithmetically wrong, and the 'new' equilibrium concepts reduce to classical ones under relabeled prompts. read the letter →

arxiv 2507.10621 v1 pith:DHERVGWQ submitted 2025-07-14 cs.CR cs.AIcs.CYcs.GT

Quanyan Zhu This is my paper · ORCID
classification cs.CRcs.AIcs.CYcs.GT MSC 91A8091A10
keywords gametheorylargelanguagemodelsagenticAIcybersecurityNashequilibriumStackelbergpromptengineeringmulti-agentsystems
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

The chapter argues that the arrival of large-language-model agents changes the natural locus of strategy in cybersecurity games: instead of choosing actions, an agent chooses a reasoning prompt, and the model's output distribution over actions is the resulting mixed strategy. It introduces two equilibrium concepts over prompt spaces—the LLM-Nash equilibrium and the LLM-Stackelberg equilibrium—and claims these capture how bounded, language-conditioned agents would settle into stable strategic patterns. The paper also proposes that multi-agent workflows built from LLM agents can be modeled as layered games-in-games, with a 'gestalt' equilibrium characterizing the whole system. If this view is right, cyber defense becomes a matter of designing reasoning environments (prompts, retrieval, fine-tuning) rather than just patching actions.

What carries the argument

The load-bearing object is the reasoning prompt as a metastrategy: a structured natural-language string that conditions an LLM's generative policy and thereby induces a probability distribution over the agent's actions. Equilibrium is defined at the level of prompt pairs, with the LLM-Nash definition (Definition 1) requiring expected-utility best responses in prompt space, and the LLM-Stackelberg definition adding a leader's anticipatory maximization over the follower's prompt-based best response. The third mechanism is the 'gestalt' game: a games-in-games composition in which local cooperative and adversarial games among workflow agents are nested so that the whole system exhibits an emergent strategic equilibrium that is not decomposable into isolated pairwise games.

What would settle it

Recompute the 5×5 expected-utility table in the Rock-Paper-Scissors example from the stated action distributions and zero-sum payoff matrix, and check whether (x5,y3) is indeed a mutual best response; the paper states U53=0.02 and U35=0.04, but direct computation yields different values if the arithmetic is in error. A second check is to instrument an actual LLM with the five x-prompts and five y-prompts, measure the induced action frequencies, and test whether the best responses match the equilibrium conditions.

Watch

Extended reading notes

Core claim

The central claim is that equilibrium analysis in cybersecurity should move from the action space to the reasoning space. For a two-player LLM game, each agent's mindset includes private information, a prompt space, and an LLM whose parameters generate a probability distribution over actions for each prompt. Definition 1 defines an LLM-Nash equilibrium as a pair of prompts from which neither player can raise expected utility by unilaterally choosing a different prompt. The LLM-Stackelberg equilibrium extends this to a leader-follower setting: the leader chooses a prompt anticipating that the follower will best-respond in its own prompt space for every received message. The paper illustrates the Nash concept with an extended Rock-Paper-Scissors game in which the claimed equilibrium (x5,y3) induces behavioral distributions that differ from the classical uniform random equilibrium; it concludes that stable equilibria under LLM constraints may be rational within the agents' reasoning models yet suboptimal in the classical sense. For multi-agent workflows, the chapter combines these local games into a 'gestalt' game-theoretic description based on the games-in-games framework.

Load-bearing premise

The LLM-Nash equilibrium definition assumes that each prompt reliably induces a known probability distribution over actions and that both players know these distributions well enough to optimize against them; if an LLM's output distribution is unstable, unobservable, or unknown to the other player, the equilibrium concept does not apply.

Editorial extensions

If this is right

  • Defenders can be engineered at the reasoning level: changing prompts, retrieval context, and fine-tuning budgets becomes the main control knob, rather than hard-coded defense policies.
  • Equilibrium outcomes under LLM constraints may be suboptimal compared with classical Nash equilibria, so defenders should expect and plan for bounded, language-conditioned adversaries that stabilize on non-classical patterns.
  • The LLM-Stackelberg model gives a formal language for studying deception, phishing, and persuasion in natural language, connecting game theory to message-level interaction rather than symbolic signals.
  • Multi-agent LLM workflows (chain, star, parallel, feedback) can be analyzed as layered games, so questions about robustness, oscillation, and coordination become equilibrium-selection questions.
  • Alignment of LLMs to human preferences can be reframed as computing Nash or Pareto equilibria among competing preference models, which is exactly the direction the paper points to via Nash learning from human feedback.

Reading between the lines

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

  • The chapter's framing implies that the security of an LLM-based system is the security of its prompt space: adversarial prompt injection becomes an attempt to shift the induced distributions and thereby break the equilibrium, so defense should formally track equilibrium-preserving prompt perturbations.
  • A direct testable extension is to treat soft prompts as continuous and use gradient-based optimization to solve for LLM-Nash equilibria; whether these optima match the discrete prompt examples in the chapter is an open question.
  • The same equilibrium language could govern multi-agent alignment: choose system prompts so that the ensemble of models is at an equilibrium, meaning no single model gains by secret deviation—an implicit but unstated design principle for agent swarms.
  • The gestalt games-in-games view suggests that cross-level coupling (a local agent's prompt change affects the global game) can be analyzed through nested equilibria; this is a natural place to bring in learning dynamics such as regret minimization to test convergence.
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 / 5 minor

Summary. This manuscript is a chapter-length position paper arguing that game theory and LLM-based agentic AI should be jointly used to model and operate cyber defense. Sections 2–3 review classical security games (static, dynamic, Bayesian, and signaling games with Nash, Stackelberg, SPNE, and PBNE concepts) and argue that LLMs relax classical assumptions of full rationality, common knowledge, and static payoffs. Section 4 introduces the principal original contributions: an LLM-Nash reasoning equilibrium (Definition 1, Eqs. (15)–(16)) in which players choose prompts that induce action distributions through LLMs, and an LLM-Stackelberg equilibrium for signaling with prompt-based message and response policies. Section 5 surveys multi-agent workflow architectures (chain, star, parallel, feedback, and hybrid) and invokes a 'gestalt' games-in-games model for the overall system. The central claimed demonstration is the prompt-space RPS example in §4.1, which asserts that (x5, y3) is a reasoning-level equilibrium.

Significance. If the proposed equilibrium concepts were rigorously established, they would provide a useful bridge between the mature security-game literature and the rapidly growing area of LLM-based agents, enabling system designers to reason about strategy at the level of prompts rather than actions. The paper is readable, covers a wide literature, and its workflow taxonomy is a practical contribution. However, the new solution concepts are not backed by any theorem, experiment, or valid numerical example: the single illustrative equilibrium pair is refuted by the paper's own numbers, and the 'empirical evaluations' are not described. The paper ships no code, data, or machine-checked proofs, so those verifiable strengths are absent. The significance is therefore conditional on a corrected or replaced demonstration and on an explicit statement of the epistemic assumptions of Definition 1.

major comments (4)
  1. [§4.1, RPS example] The paper's only numerical demonstration of Definition 1 is invalid. Direct computation from the stated distributions gives U53 = 0, not 0.02, and U35 = 0.03, not 0.04. More importantly, against attacker x5 = (0.3, 0.4, 0.3), the attacker's expected payoff against defender prompts is 0 for y1 through y4 and −0.02 for y5 = (0.3, 0.2, 0.5); since uD = −uA, the defender strictly prefers y5 over y3 = (0.2, 0.6, 0.2), so inequality (16) fails for (x5, y3). The sentence beginning 'Assume empirical evaluations indicate' does not repair this, because no evaluation data are provided and the numerical contradiction is intrinsic to the stated distributions. This is load-bearing: Section 6's claim that the chapter shows strategic behavior emerging from prompt-space equilibria rests entirely on this example.
  2. [§4.1, 'Assume empirical evaluations indicate...'] The empirical assertion is unsupported. The manuscript provides no LLM, no prompt texts, no inference protocol, no sample size, and no results; it simply assumes that x5 and y3 are best responses. Because Definition 1 presupposes that the distributions µA(x_i) and µD(y_j) are known to the players, the claimed best responses are not derived from the model but imported as an unexamined assumption. The example should be replaced by a pair verified arithmetically from the stated distributions, or explicitly labeled as a hypothetical illustration; in the latter case it cannot serve as validation of the new concept.
  3. [§4.1, Definition 1 vs. §3.1] The epistemic assumptions of Definition 1 are not reconciled with the paper's own critique of classical game theory. Equations (15)–(16) are well defined only if each prompt induces a fixed action distribution and both players know these distributions with common knowledge; otherwise the expected utilities and the unilateral-deviation test are not well posed. Yet Section 3.1 motivates LLMs precisely by rejecting common knowledge and perfect rationality. The paper should state explicitly whether LLM-Nash equilibrium is a normative concept over known distributions, a positive description of LLM behavior, or an equilibrium of a game in which the distributions are learned, and it should specify the information available to each player when comparing prompts.
  4. [§4 (overall)] No existence, uniqueness, or tractability result is given for either LLM-Nash or LLM-Stackelberg equilibrium, and no experiment is reported. For a chapter that announces new solution concepts, the absence of any positive formal or empirical result—together with the invalid RPS example—means the central claim is currently unsupported. At minimum, a correct finite-space equilibrium example and a statement of conditions under which an LLM-Nash equilibrium exists (or a counterexample to existence) are needed before the concepts can be used as advertised in Section 6.
minor comments (5)
  1. [§4.2] The phrase 'The LLM-Stackelberg Stackelberg game' contains a duplicated word; also, 'we introduce LLM-Stackelberg game' needs an article.
  2. [§5.1.6] The text refers to 'Figure 5.1.6', but the surrounding figures are numbered Figure 5, so the cross-reference is incorrect.
  3. [§5.4] The phrase 'we call a view Gestalt game-theoretic' is awkward; it should be reworded as 'a gestalt game-theoretic view'.
  4. [References] Several arXiv identifiers are malformed or missing their declaration numbers (e.g., [1] appears as 'arXiv:240518718' and [31] as 'arXiv:250201587'); these should be checked against the journal's reference style.
  5. [§2.1 and §4.1] The notation 'I= {1, 2, . . . , N}' has a spacing issue, and the mindset parameters θ, δ in §4.1 are introduced without a clear domain; both should be cleaned up.

Circularity Check

1 steps flagged · score 4.0 of 10

The paper's sole equilibrium demonstration assumes the best-response conditions of Definition 1 and then restates them as a conclusion; the LLM-Nash definition itself is not circular.

  1. fitted input called prediction [Section 4.1, 'Example: Reasoning Equilibria in Extended Prompt-Space RPS' (after Definition 1, Eqs. 15-16)]
    "Assume empirical evaluations indicate that the attacker achieves the highest average expected utility when using prompt x5, and that the defender’s best response is prompt y3. Then, with µA(x5) = (0.3, 0.4, 0.3), µD(y3) = (0.2, 0.6, 0.2), we obtain U53 = µ⊤A UA µD = 0.02. If neither player can improve their expected utility by unilaterally changing prompts, the pair (x5, y3) constitutes a reasoning-level equilibrium."

    Definition 1 (Eqs. 15–16) defines LLM-Nash equilibrium as mutual best response: no unilateral prompt change can improve expected utility. The example's premise already asserts exactly these two facts: 'empirical evaluations indicate ... best response' and 'neither player can improve'. The conclusion that (x5, y3) is a reasoning-level equilibrium is therefore a restatement of the input assumption, not a derivation from the model. No LLM, prompt text, or data substantiate the 'empirical evaluations'. Direct computation from the stated distributions contradicts the premise: U53 = 0, not 0.02, and against x5 the defender strictly prefers y5 (attacker payoff −0.02) to y3 (payoff 0). The equilibrium is thus forced by the assumed best responses rather than demonstrated by the game structure.

full rationale

The formal material is mostly non-circular: Definition 1 is a proposal, and the LLM-Stackelberg construction in Section 4.2 independently adapts the standard leader-follower definition to prompt spaces. However, the RPS example is the only concrete demonstration of the new solution concept, and its reasoning reduces to the definition's own conditions: the paper assumes empirical evaluations have already identified the best responses and then, by the 'if neither player can improve' clause, declares the pair an equilibrium. That is exactly the equilibrium condition, so the example's output is contained in its input. The arithmetic errors (U53 = 0, not 0.02; U35 = 0.03, not 0.04) and the unsupported 'empirical evaluations' claim mean the example does not even instantiate the assumed conditions. The self-citations in Section 5.4 to the games-in-games framework [4] are not load-bearing for the core equilibrium definitions, so they do not add to the circularity score. Overall, the framework retains independent formal content, but the central illustrative result reduces by construction to an asserted fit, giving a moderate circularity score.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central proposal is not parameter-fitted, but the only worked example depends on hand-assigned action distributions and an unsupported assertion of best responses. The framework assumes LLM outputs behave like known probability distributions, a strong domain assumption.

free parameters (2)
  • Prompt-induced action distributions µA(x_i), µD(y_j) in RPS example = x1..x5 and y1..y5 distributions as listed
    Hand-assigned, not measured from any LLM; the example's equilibrium depends on these exact values.
  • Assumed best responses in RPS example = x5 for attacker, y3 for defender
    Assumed via 'empirical evaluations indicate' with no data; the stated expected utilities contain errors.
assumptions (4)
  • domain assumption Players know the induced distributions µA, µD and optimize expected utility over prompt spaces with full knowledge of the game structure.
    Definition 1 and Eqs. (15)-(16) assume agents can compute expectations over LLM outputs and each other's prompts; real LLM behavior is stochastic and prompts are not common knowledge.
  • domain assumption LLM outputs are stable, well-defined probability distributions over actions for each prompt.
    The framework treats γA, γD as fixed generative mappings; no temperature, context length, or prompt sensitivity is modeled.
  • ad hoc to paper The 'empirical evaluations indicate' claim in the RPS example is treated as evidence.
    No experiments or data are cited; the assertion is used to select the equilibrium.
  • standard math Standard game-theoretic solution concepts (Nash, Stackelberg, PBNE, SPNE) and Bayes' rule.
    Used throughout Sections 2 and 4 without proof.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Game Theory Meets LLM and Agentic AI: Reimagining Cybersecurity for the Age of Intelligent Threats." pith.science (2026). https://pith.science/paper/DHERVGWQ

@misc{pith2026250710621,
  author       = {Pith},
  title        = {Pith review of: Game Theory Meets LLM and Agentic AI: Reimagining Cybersecurity for the Age of Intelligent Threats},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DHERVGWQ}},
  note         = {Machine review of arXiv:2507.10621}
}
read the original abstract

Protecting cyberspace requires not only advanced tools but also a shift in how we reason about threats, trust, and autonomy. Traditional cybersecurity methods rely on manual responses and brittle heuristics. To build proactive and intelligent defense systems, we need integrated theoretical frameworks and software tools. Game theory provides a rigorous foundation for modeling adversarial behavior, designing strategic defenses, and enabling trust in autonomous systems. Meanwhile, software tools process cyber data, visualize attack surfaces, verify compliance, and suggest mitigations. Yet a disconnect remains between theory and practical implementation. The rise of Large Language Models (LLMs) and agentic AI offers a new path to bridge this gap. LLM-powered agents can operationalize abstract strategies into real-world decisions. Conversely, game theory can inform the reasoning and coordination of these agents across complex workflows. LLMs also challenge classical game-theoretic assumptions, such as perfect rationality or static payoffs, prompting new models aligned with cognitive and computational realities. This co-evolution promises richer theoretical foundations and novel solution concepts. Agentic AI also reshapes software design: systems must now be modular, adaptive, and trust-aware from the outset. This chapter explores the intersection of game theory, agentic AI, and cybersecurity. We review key game-theoretic frameworks (e.g., static, dynamic, Bayesian, and signaling games) and solution concepts. We then examine how LLM agents can enhance cyber defense and introduce LLM-driven games that embed reasoning into AI agents. Finally, we explore multi-agent workflows and coordination games, outlining how this convergence fosters secure, intelligent, and adaptive cyber systems.

Figures

Figures reproduced from arXiv: 2507.10621 by the authors.

Figure 1
Figure 1. Chain workflow: a stepwise pipeline of LLM agents in a linear processing sequence. [PITH_FULL_IMAGE:figures/full_fig_p025_1.png] view at source ↗
Figure 2
Figure 2. Star workflow: a central LLM agent coordinates specialized analyses across peripheral agents. [PITH_FULL_IMAGE:figures/full_fig_p026_2.png] view at source ↗
Figure 3
Figure 3. Parallel workflow: independent LLM agents process distributed data streams before aggregating [PITH_FULL_IMAGE:figures/full_fig_p026_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Feedback workflow: cyclic coordination among LLM agents supporting adaptive cybersecurity [PITH_FULL_IMAGE:figures/full_fig_p027_4.png]
Figure 5
Figure 5. Figure 5: Hybrid workflow with a sequential initialization, parallel branching, and a convergence task with [PITH_FULL_IMAGE:figures/full_fig_p028_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Guarding Against Malicious Biased Threats (GAMBiT) Experiments: Revealing Cognitive Bias in Human-Subjects Red-Team Cyber Range Operations

    cs.CR 2025-08 conditional novelty 6.0 of 10

    Three multi-modal datasets capture 59 skilled attackers' full operational traces (keystrokes, shell history, PCAP, surveys) in a simulated enterprise network, with labels designed to reveal cognitive biases.

Reference graph

Works this paper leans on

81 extracted references · 77 canonical work pages · cited by 1 Pith paper

  1. [1]

    arXiv preprint arXiv:240518718

    Bai F, Wang M, Zhang Z, Chen B, Xu Y, Wen Y, Yang Y (2024) Efficient model-agnostic alignment via bayesian persuasion. arXiv preprint arXiv:240518718

    2024

  2. [2]

    In: 2015 IEEE 14th International Conference on Machine Learning and Applications (ICMLA), IEEE, pp 941–947

    Biswas T, Regan K (2015) Measuring level-k reasoning, satisficing, and human error in game-play data. In: 2015 IEEE 14th International Conference on Machine Learning and Applications (ICMLA), IEEE, pp 941–947

    work page 2015

  3. [3]

    In: 2018 52nd Annual Conference on Information Sciences and Systems (CISS), IEEE, pp 1–6

    Chen J, Zhu Q (2018) Security investment under cognitive constraints: A gestalt nash equilibrium approach. In: 2018 52nd Annual Conference on Information Sciences and Systems (CISS), IEEE, pp 1–6

    work page 2018

  4. [4]

    IEEE Transactions on Control of Network Systems 7(3):1056–1068

    Chen J, Zhu Q (2019) Control of multilayer mobile autonomous systems in adversarial environments: A games-in-games approach. IEEE Transactions on Control of Network Systems 7(3):1056–1068

    work page 2019

  5. [5]

    Springer

    Chen J, Zhu Q (2019) A game-and decision-theoretic approach to resilient interdependent network analysis and design. Springer

    work page 2019

  6. [6]

    IEEE Transactions on Control of Network Systems 7(1):398–409, DOI 10.1109/TCNS.2019.2906893

    Chen J, Touati C, Zhu Q (2020) Optimal secure two-layer iot network design. IEEE Transactions on Control of Network Systems 7(1):398–409, DOI 10.1109/TCNS.2019.2906893

    work page arXiv 2020

  7. [7]

    Dynamic Games and Applications 11:294–325

    Chen J, Zhu Q, Ba¸ sar T (2021) Dynamic contract design for systemic cyber risk management of interdependent enterprise networks. Dynamic Games and Applications 11:294–325

    work page 2021

  8. [8]

    Chen JCY, Saha S, Bansal M (2024) Reconcile: Round-table conference improves reasoning via consensus among diverse llms. In: Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024), Association for Computational Linguistics, Bangkok, Thai- land, pp 7066–7085, DOI 10.18653/v1/2024.acl-long.381, URL https://aclanth...

    doi:10.18653/v1/2024.acl-long.381 2024

Show all 81 references
  1. [9]

    In: ICLR 2024 (Poster), URL https://openreview.net/forum?id=EHg5GDnyq1, poster

    Chen W, Su Y, Zuo J, Yang C, Yuan C, Chan CM, Yu H, Lu Y, Hung YH, Qian C, Qin Y, Cong X, Xie R, Liu Z, Sun M, Zhou J (2024) Agentverse: Facilitating multi-agent collaboration and exploring emergent behaviors. In: ICLR 2024 (Poster), URL https://openreview.net/forum?id=EHg5GDn...

    2024

  2. [10]

    arXiv preprint arXiv:240101335

    Chen Z, Deng Y, Yuan H, Ji K, Gu Q (2024) Self-play fine-tuning converts weak language models to strong language models. arXiv preprint arXiv:240101335

    2024

  3. [11]

    In: Forty-first International Conference on Machine Learning

    Du Y, Li S, Torralba A, Tenenbaum JB, Mordatch I (2023) Improving factuality and reasoning in language models through multiagent debate. In: Forty-first International Conference on Machine Learning

    2023

  4. [12]

    arXiv preprint arXiv:240201306

    Ethayarajh K, Xu W, Muennighoff N, Jurafsky D, Kiela D (2024) Kto: Model alignment as prospect theoretic optimization. arXiv preprint arXiv:240201306

    2024

  5. [13]

    arXiv preprint arXiv:200804893

    Fang S, Zhu Q (2020) Channel leakage, information-theoretic limitations of obfuscation, and optimal privacy mask design for streaming data. arXiv preprint arXiv:200804893

    2020

  6. [14]

    In: 2021 American Control Conference (ACC), IEEE, pp 4574–4579 32

    Fang S, Zhu Q (2021) Fundamental limits of obfuscation for linear gaussian dynamical systems: An information-theoretic approach. In: 2021 American Control Conference (ACC), IEEE, pp 4574–4579 32

    2021

  7. [15]

    In: Decision and Game Theory for Security: 15th International Conference, GameSec 2024, New York City, NY, USA, October 16–18, 2024, Proceedings, Springer Nature, vol 14908, p 124

    Farina G, Kroer C, Iyengar G (2025) Contested logistics: A game-theoretic approach. In: Decision and Game Theory for Security: 15th International Conference, GameSec 2024, New York City, NY, USA, October 16–18, 2024, Proceedings, Springer Nature, vol 14908, p 124

    2025

  8. [16]

    In: 2011 7th International Conference on Network and Service Management, IEEE, pp 1–6

    Fung C, Zhu Q, Boutaba R, Ba¸ sar T (2011) Smurfen: A system framework for rule sharing collabora- tive intrusion detection. In: 2011 7th International Conference on Network and Service Management, IEEE, pp 1–6

    2011

  9. [17]

    In: Decision and Game Theory for Security: 8th International Conference, GameSec 2017, Vienna, Austria, October 23-25, 2017, Proceedings, Springer, pp 273–294

    Hor´ ak K, Zhu Q, Boˇ sansk` y B (2017) Manipulating adversary’s belief: A dynamic game approach to deception by design for proactive network security. In: Decision and Game Theory for Security: 8th International Conference, GameSec 2017, Vienna, Austria, October 23-25, 2017, ...

    2017

  10. [18]

    In: 2022 IEEE 61st Conference on Decision and Control (CDC), IEEE, pp 6111–6117

    Hu Y, Zhu Q (2022) Evasion-aware neyman-pearson detectors: A game-theoretic approach. In: 2022 IEEE 61st Conference on Decision and Control (CDC), IEEE, pp 6111–6117

    2022

  11. [19]

    In: 2023 IEEE Statistical Signal Processing Workshop (SSP), IEEE, pp 36–40

    Hu Y, Zhu Q (2023) Detection in human-sensor systems under quantum prospect theory using bayesian persuasion frameworks. In: 2023 IEEE Statistical Signal Processing Workshop (SSP), IEEE, pp 36–40

    2023

  12. [20]

    IEEE Transactions on Information Forensics and Security

    Hu Y, Chen J, Zhu Q (2024) Game-theoretic neyman-pearson detection to combat strategic evasion. IEEE Transactions on Information Forensics and Security

    2024

  13. [21]

    Computers & Security 89:101660

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

    2020

  14. [22]

    Springer Science & Business Media

    Jajodia S, Ghosh AK, Swarup V, Wang C, Wang XS (2011) Moving target defense: creating asym- metric uncertainty for cyber threats, vol 54. Springer Science & Business Media

    2011

  15. [23]

    In: Handbook of the fundamentals of financial decision making: Part I, World Scientific, pp 99–127

    Kahneman D, Tversky A (2013) Prospect theory: An analysis of decision under risk. In: Handbook of the fundamentals of financial decision making: Part I, World Scientific, pp 99–127

    2013

  16. [24]

    Springer

    Kieras T, Farooq J, Zhu Q (2022) IoT Supply Chain Security Risk Analysis and Mitigation: Mod- eling, Computations, and Software Tools. Springer

    2022

  17. [25]

    In: Globerson A, Mackey L, Belgrave D, Fan A, Paquet U, Tomczak J, Zhang C (eds) Advances in Neural Information Processing Systems, Curran Associates, Inc., vol 37, pp 79410–79452

    Kim Y, Park C, Jeong H, Chan YS, Xu X, McDuff D, Lee H, Ghassemi M, Breazeal C, Park HW (2024) Mdagents: An adaptive collaboration of llms for medical decision-making. In: Globerson A, Mackey L, Belgrave D, Fan A, Paquet U, Tomczak J, Zhang C (eds) Advances in Neural Informati...

    2024

  18. [26]

    Springer

    Kott A, Linkov I (2019) Cyber resilience of systems and networks, vol 1. Springer

    2019

  19. [27]

    In: International Conference on Decision and Game Theory for Security, Springer, pp 168–187

    Kulkarni AN, Fu J, Luo H, Kamhoua CA, Leslie NO (2020) Decoy allocation games on graphs with temporal logic objectives. In: International Conference on Decision and Game Theory for Security, Springer, pp 168–187

    2020

  20. [28]

    In: Proceedings of the ACM Web Conference 2024, pp 66–76 33

    Laufer B, Kleinberg J, Heidari H (2024) Fine-tuning games: Bargaining and adaptation for general- purpose models. In: Proceedings of the ACM Web Conference 2024, pp 66–76 33

    2024

  21. [29]

    In: Oh A, Naumann T, Globerson A, Saenko K, Hardt M, Levine S (eds) Advances in Neural Information Processing Systems, Curran Associates, Inc., vol 36, pp 51991–52008

    Li G, Hammoud H, Itani H, Khizbullin D, Ghanem B (2023) Camel: Communicative agents for ”mind” exploration of large language model society. In: Oh A, Naumann T, Globerson A, Saenko K, Hardt M, Levine S (eds) Advances in Neural Information Processing Systems, Curran Associates,...

    2023

  22. [30]

    IEEE Control Sys- tems 42(4):35–67, DOI 10.1109/mcs.2022.3171478

    Li T, Peng G, Zhu Q, Baar T (2022) The Confluence of Networks, Games, and Learning a Game-Theoretic Framework for Multiagent Decision Making Over Networks. IEEE Control Sys- tems 42(4):35–67, DOI 10.1109/mcs.2022.3171478

    2022

  23. [31]

    arXiv preprint arXiv:250201587

    Li W, Lin Y, Wang X, Jin B, Zha H, Wang B (2025) Verbalized bayesian persuasion. arXiv preprint arXiv:250201587

    2025

  24. [32]

    arXiv preprint arXiv:250310990

    Liu K, Long Q, Shi Z, Su WJ, Xiao J (2025) Statistical impossibility and possibility of align- ing llms with human preferences: From condorcet paradox to nash equilibrium. arXiv preprint arXiv:250310990

    2025

  25. [33]

    Advances in Neural Information Processing Systems 37:15497–15525

    Ma H, Hu T, Pu Z, Boyin L, Ai X, Liang Y, Chen M (2024) Coevolving with the other you: Fine-tuning llm with sequential cooperative multi-agent reinforcement learning. Advances in Neural Information Processing Systems 37:15497–15525

    2024

  26. [34]

    ACM Computing Surveys (CSUR) 45(3):1–39

    Manshaei MH, Zhu Q, Alpcan T, Bac¸ sar T, Hubaux JP (2013) Game theory meets network security and privacy. ACM Computing Surveys (CSUR) 45(3):1–39

    2013

  27. [35]

    arXiv preprint arXiv:231200886 18

    Munos R, Valko M, Calandriello D, Azar MG, Rowland M, Guo ZD, Tang Y, Geist M, Mesnard T, Michi A, et al (2023) Nash learning from human feedback. arXiv preprint arXiv:231200886 18

    2023

  28. [36]

    Computation, cryptography, and network security pp 381– 398

    Nagurney A, Nagurney LS, Shukla S (2015) A supply chain game theory framework for cybersecurity investments under network vulnerability. Computation, cryptography, and network security pp 381– 398

    2015

  29. [37]

    In: Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security, pp 1727–1738

    Nasr M, Houmansadr A (2016) Game of decoys: Optimal decoy routing through game theory. In: Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security, pp 1727–1738

    2016

  30. [38]

    In: 2019 American Control Conference (ACC), IEEE, pp 2060–2065

    Nugraha Y, Hayakawa T, Cetinkaya A, Ishii H, Zhu Q (2019) Subgame perfect equilibrium analysis for jamming attacks on resilient graphs. In: 2019 American Control Conference (ACC), IEEE, pp 2060–2065

    2019

  31. [39]

    IEEE Transactions on Control of Network Systems 8(1):246– 259

    Nugraha Y, Cetinkaya A, Hayakawa T, Ishii H, Zhu Q (2020) Dynamic resilient network games with applications to multiagent consensus. IEEE Transactions on Control of Network Systems 8(1):246– 259

    2020

  32. [40]

    Automatica 175:112137

    Nugraha YE, Cetinkaya A, Hayakawa T, Ishii H, Zhu Q (2025) A rolling horizon game considering network effect in cluster forming for dynamic resilient multiagent systems. Automatica 175:112137

    2025

  33. [41]

    In: 2016 IEEE International Workshop on Information Forensics and Security (WIFS), IEEE, pp 1–6

    Pawlick J, Zhu Q (2016) A stackelberg game perspective on the conflict between machine learning and data obfuscation. In: 2016 IEEE International Workshop on Information Forensics and Security (WIFS), IEEE, pp 1–6

    2016

  34. [42]

    IEEE Transactions on Information Forensics and Security 12(12):2906–2919 34

    Pawlick J, Zhu Q (2017) 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 34

    2017

  35. [43]

    In: International Conference on Decision and Game Theory for Security (GameSec), Springer, pp 289–308, DOI 10.1007/978-3-319-25594-1 17

    Pawlick J, Farhang S, Zhu Q (2015) Flip the cloud: Cyber-physical signaling games in the presence of advanced persistent threats. In: International Conference on Decision and Game Theory for Security (GameSec), Springer, pp 289–308, DOI 10.1007/978-3-319-25594-1 17

    2015 doi

  36. [44]

    IEEE Transactions on Information Forensics and Security 14(6):1654–1669

    Pawlick J, Chen J, Zhu Q (2018) 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

  37. [45]

    IEEE Transactions on Information Forensics and Security 14(7):1871–1886

    Pawlick J, Colbert E, Zhu Q (2018) Modeling and analysis of leaky deception using signaling games with evidence. IEEE Transactions on Information Forensics and Security 14(7):1871–1886

    2018

  38. [46]

    ACM Computing Surveys (CSUR) 52(4):1–28

    Pawlick J, Colbert E, Zhu Q (2019) A game-theoretic taxonomy and survey of defensive deception for cybersecurity and privacy. ACM Computing Surveys (CSUR) 52(4):1–28

    2019

  39. [47]

    Springer

    Pawlick J, Zhu Q, et al (2021) Game theory for cyber deception. Springer

    2021

  40. [48]

    Qian C, Liu W, Liu H, Chen N, Dang Y, Li J, Yang C, Chen W, Su Y, Cong X, Xu J, Li D, Liu Z, Sun M (2024) Chatdev: Communicative agents for software development. In: Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)...

    2024 doi

  41. [49]

    In: 2020 IEEE Conference on Control Technology and Applications (CCTA), IEEE, pp 1–7

    Qian Z, Fu J, Zhu Q (2020) A receding-horizon mdp approach for performance evaluation of moving target defense in networks. In: 2020 IEEE Conference on Control Technology and Applications (CCTA), IEEE, pp 1–7

    2020

  42. [50]

    Cyber-Security in Critical Infrastructures: A Game-Theoretic Approach pp 99–114

    Rass S, Schauer S, K¨ onig S, Zhu Q, Rass S, Schauer S, K¨ onig S, Zhu Q (2020) Bounded rationality. Cyber-Security in Critical Infrastructures: A Game-Theoretic Approach pp 99–114

    2020

  43. [51]

    arXiv preprint arXiv:240403715

    Rosset C, Cheng CA, Mitra A, Santacroce M, Awadallah A, Xie T (2024) Direct nash optimization: Teaching language models to self-improve with general preferences. arXiv preprint arXiv:240403715

    2024

  44. [52]

    Addison-Wesley Longman Publishing Co., Inc

    Spitzner L (2002) Honeypots: tracking hackers. Addison-Wesley Longman Publishing Co., Inc

    2002

  45. [53]

    arXiv preprint arXiv:240516276

    Sun H, Chen Y, Wang S, Chen W, Deng X (2024) Mechanism design for llm fine-tuning with multiple reward models. arXiv preprint arXiv:240516276

    2024

  46. [54]

    URL https://arxiv.org/abs/2401.04056, 2401.04056

    Swamy G, Dann C, Kidambi R, Wu ZS, Agarwal A (2024) A minimaximalist approach to reinforce- ment learning from human feedback. URL https://arxiv.org/abs/2401.04056, 2401.04056

    2024 arXiv

  47. [55]

    In: Find- ings of the Association for Computational Linguistics: ACL 2024, Bangkok, Thailand, URL https://aclanthology.org/2024.findings-acl.33/

    Tang X, Zou A, Zhang Z, Li Z, Zhao Y, Zhang X, Cohan A, Gerstein M (2024) Meda- gents: Large language models as collaborators for zero-shot medical reasoning. In: Find- ings of the Association for Computational Linguistics: ACL 2024, Bangkok, Thailand, URL https://aclanthology...

    2024

  48. [56]

    In: Globerson A, Mackey L, Belgrave D, Fan A, Paquet U, Tomczak J, Zhang C (eds) Advances in Neural Information Processing Systems, Curran Associates, Inc., vol 37, pp 51963–51993

    Tao W, Zhou Y, Wang Y, Zhang W, Zhang H, Cheng Y (2024) Magis: Llm-based multi-agent framework for github issue resolution. In: Globerson A, Mackey L, Belgrave D, Fan A, Paquet U, Tomczak J, Zhang C (eds) Advances in Neural Information Processing Systems, Curran Associates, In...

    2024

  49. [57]

    arXiv preprint arXiv:240500675 35

    Wu Y, Sun Z, Yuan H, Ji K, Yang Y, Gu Q (2024) Self-play preference optimization for language model alignment. arXiv preprint arXiv:240500675 35

    2024

  50. [58]

    In: Proceedings of the 1st International Workshop on Safe Control of Connected and Autonomous Vehicles, pp 27–34

    Xu Z, Zhu Q (2017) A game-theoretic approach to secure control of communication-based train control systems under jamming attacks. In: Proceedings of the 1st International Workshop on Safe Control of Connected and Autonomous Vehicles, pp 27–34

    2017

  51. [59]

    arXiv preprint arXiv:250521244

    Yang YT, Zhu Q (2025) When to deceive: A cross-layer stackelberg game framework for strategic timing of cyber deception. arXiv preprint arXiv:250521244

    2025

  52. [60]

    In: International Conference on Decision and Game Theory for Security, Springer, pp 349–368

    Yang YT, Zhang T, Zhu Q (2023) A game-theoretic analysis of auditing differentially private al- gorithms with epistemically disparate herd. In: International Conference on Decision and Game Theory for Security, Springer, pp 349–368

    2023

  53. [61]

    Advances in Neural Information Processing Systems 37:81773–81807

    Ye C, Xiong W, Zhang Y, Dong H, Jiang N, Zhang T (2024) Online iterative reinforcement learning from human feedback with general preference model. Advances in Neural Information Processing Systems 37:81773–81807

    2024

  54. [62]

    arXiv e-prints pp arXiv–2402

    Ye C, Xiong W, Zhang Y, Jiang N, Zhang T (2024) A theoretical analysis of nash learning from human feedback under general kl-regularized preference. arXiv e-prints pp arXiv–2402

    2024

  55. [63]

    arXiv preprint arXiv:241100062

    Ye Z, Agarwal R, Liu T, Joshi R, Velury S, Le QV, Tan Q, Liu Y (2024) Evolving alignment via asymmetric self-play. arXiv preprint arXiv:241100062

    2024

  56. [64]

    In: 2015 18th International Conference on Information Fusion (Fusion), IEEE, pp 644–651

    Zhang R, Zhu Q (2015) Secure and resilient distributed machine learning under adversarial environ- ments. In: 2015 18th International Conference on Information Fusion (Fusion), IEEE, pp 644–651

    2015

  57. [65]

    IEEE transactions on neural networks and learning systems 29(11):5512–5527

    Zhang R, Zhu Q (2018) A game-theoretic approach to design secure and resilient distributed support vector machines. IEEE transactions on neural networks and learning systems 29(11):5512–5527

    2018

  58. [66]

    Adversary-Aware Learning Techniques and Trends in Cybersecurity p 17

    Zhang R, Zhu Q (2021) Security of distributed machine. Adversary-Aware Learning Techniques and Trends in Cybersecurity p 17

    2021

  59. [67]

    Zhang Y, Sun R, Chen Y, Pfister T, Zhang R, Arı k SO (2024) Chain of agents: Large language models collaborating on long-context tasks. In: Globerson A, Mackey L, Belgrave D, Fan A, Pa- quet U, Tomczak J, Zhang C (eds) Advances in Neural Information Processing Systems, Curran ...

    2024

  60. [68]

    Advances in Neural Information Processing Systems 37:75522–75558

    Zhong Y, Ma C, Zhang X, Yang Z, Chen H, Zhang Q, Qi S, Yang Y (2024) Panacea: Pareto alignment via preference adaptation for llms. Advances in Neural Information Processing Systems 37:75522–75558

    2024

  61. [69]

    arXiv preprint arXiv:240401205

    Zhu Q (2024) Foundations of cyber resilience: The confluence of game, control, and learning theories. arXiv preprint arXiv:240401205

    2024

  62. [70]

    In: Proceedings of the 48h IEEE Conference on Decision and Control (CDC) held jointly with 2009 28th Chinese Control Conference, IEEE, pp 8600–8605

    Zhu Q, Ba¸ sar T (2009) Dynamic policy-based ids configuration. In: Proceedings of the 48h IEEE Conference on Decision and Control (CDC) held jointly with 2009 28th Chinese Control Conference, IEEE, pp 8600–8605

    2009

  63. [71]

    In: International conference on decision and game theory for security, Springer, pp 246–263

    Zhu Q, Ba¸ sar T (2013) Game-theoretic approach to feedback-driven multi-stage moving target defense. In: International conference on decision and game theory for security, Springer, pp 246–263

    2013

  64. [72]

    IEEE Control Systems Magazine 35(1):46–65 36

    Zhu Q, Basar T (2015) Game-theoretic methods for robustness, security, and resilience of cyber- physical control systems: games-in-games principle for optimal cross-layer resilient control systems. IEEE Control Systems Magazine 35(1):46–65 36

    2015

  65. [73]

    arXiv preprint arXiv:240306299

    Zhu Q, Basar T (2024) Disentangling resilience from robustness: Contextual dualism, interactionism, and game-theoretic paradigms. arXiv preprint arXiv:240306299

    2024

  66. [74]

    IEEE Control Systems Letters

    Zhu Q, Ba¸ sar T (2025) Revisiting game-theoretic control in socio-technical networks: Emerging design frameworks and contemporary applications. IEEE Control Systems Letters

    2025

  67. [75]

    IEEE Access 6:13958–13971

    Zhu Q, Rass S (2018) On multi-phase and multi-stage game-theoretic modeling of advanced persis- tent threats. IEEE Access 6:13958–13971

    2018

  68. [76]

    Springer

    Zhu Q, Xu Z (2020) Cross-layer design for secure and resilient cyber-physical systems. Springer

    2020

  69. [77]

    In: 2010 IEEE International Conference on Communications, IEEE, pp 1–6

    Zhu Q, Li H, Han Z, Ba¸ sar T (2010) A stochastic game model for jamming in multi-channel cognitive radio systems. In: 2010 IEEE International Conference on Communications, IEEE, pp 1–6

    2010

  70. [78]

    In: Proceedings of the 2010 American control conference, IEEE, pp 1059–1064

    Zhu Q, Tembine H, Ba¸ sar T (2010) Network security configurations: A nonzero-sum stochastic game approach. In: Proceedings of the 2010 American control conference, IEEE, pp 1059–1064

    2010

  71. [79]

    In: 2011 50th IEEE Conference on Decision and Control and European Control Conference, IEEE, pp 243–248

    Zhu Q, Fung C, Boutaba R, Ba¸ sar T (2011) A game-theoretic approach to rule sharing mechanism in networked intrusion detection systems: Robustness, incentives and security. In: 2011 50th IEEE Conference on Decision and Control and European Control Conference, IEEE, pp 243–248

    2011

  72. [80]

    In: 2011-MILCOM 2011 Military Communications Conference, IEEE, pp 119–124

    Zhu Q, Saad W, Han Z, Poor HV, Ba¸ sar T (2011) Eavesdropping and jamming in next-generation wireless networks: A game-theoretic approach. In: 2011-MILCOM 2011 Military Communications Conference, IEEE, pp 119–124

    2011

  73. [81]

    IEEE Journal on Selected Areas in Communications 30(11):2220– 2230 37

    Zhu Q, Fung C, Boutaba R, Basar T (2012) Guidex: A game-theoretic incentive-based mechanism for intrusion detection networks. IEEE Journal on Selected Areas in Communications 30(11):2220– 2230 37

    2012

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.