arXiv subjects

Runyu Zhang

Publications and source records attributed to Runyu Zhang.

At least 19 recordsLinked to original sources

arXiv2026

Auditable and Deployable Multi-Agent Motion Planning via Posetal Games from Legal Guidelines

When robots are deployed in the real-world, they must interact with other agents and accomplish their own goals while adhering to legal constraints. These potentially conflicting objectives are often only partially prioritized, making scalarized objectives or fully specified hierarchies unsuitable. We present a framework for multi-agent motion planning under legal requirements based on games with partially ordered preferences (posetal games). Scenario-independent legal guidelines order abstract legal concerns, from which we automatically derive agent preferences and a legal game for any concrete scenario. Agents may refine their legal preferences with private objectives, and regulators may refine the guideline as laws evolve. We prove that such customizations can only remove equilibria and thus preserve legality. To coordinate agents without explicitly communicating private preferences, we introduce belief play, which provably recovers the true equilibria from conservative local games, and a distributed iterative best-response protocol requiring less information. We apply the framework to maritime navigation, systematically lifting pairwise traffic rules to multi-vessel encounters. On seven encounters with up to four vessels, every rule violation of the committed equilibria is certified as forced by the legal hierarchy or unsatisfiable, and defectors are identified from public information in all tested cases. These results indicate that guideline-based posetal games offer an auditable route from partially specified laws to multi-agent motion planning.

cs.MA↗

arXiv2026

Metric-Based Equilibrium Selection in Noncooperative Differentiable Games

Metric conditioning can change which equilibrium attracts learning dynamics without moving the equilibria of a differentiable game. We study how to use state-dependent symmetric positive-definite (SPD) metrics to select among known equilibria by keeping a designated equilibrium attracting while making a rival unstable. Building on classical results on matrix stability under positive-definite multiplication, we characterize when such destabilization is possible and show how restricting the metric to act independently on each player limits the stability changes that can be achieved at differential Nash equilibria. We then combine a stabilizing metric at the selected equilibrium with a destabilizing metric at the rival through smooth interpolation, yielding a single metric field that preserves every equilibrium of the game while assigning the desired local stability types. We also derive a step-size condition for implementing the method in discrete time. The approach is validated on a continuous-commitment stag hunt and an entropy-regularized iterated prisoner's dilemma, where it selects the desired equilibrium in both settings.

cs.GT↗

arXiv2026

CONTRAMEM: Learning Self-Evolving Procedural Memory from Contrasting Multi-Model Trajectories

Autonomous computer-use agents are increasingly applied to long-horizon tasks requiring coordinated application calls, persistent state tracking, and verifier-sensitive writes, yet they remain prone to procedural failures: misreading application state, tool semantics, or task progress. Procedural memory promises more consistent decisions and less redundant exploration, but constructing high-quality memory without model training remains challenging. We introduce CONTRAMEM, a source-flexible, training-free framework for self-evolving procedural memory that treats same-task outcome variation as supervision: differences in correctness, efficiency, recovery, and failure modes expose outcome-relevant procedural distinctions, distilled into a compact bank of app-level Function Cards and task-level Skill Cards that evolves through localized curation rather than append-only accumulation or whole-bank rewriting. On held-out GAIA2/ARE computer-use tasks, CONTRAMEM more than doubles the success rate across the three source-model targets (26.2% to 55.3%), with consistent per-model gains (GPT-5.5: 27.5 to 61.0; Claude Sonnet 4.6: 28.0 to 52.5; DeepSeek V4 Pro: 23.0 to 52.5). The same bank transfers unchanged to the unseen Qwen3.7 Plus (18.5 to 35.5), indicating transferable procedural knowledge rather than model-specific behavior. The same construction carries over unchanged to AppWorld, beating both no memory and its own single-source self-memory variant for all three mid-tier agents on both public test splits. Under a matched trajectory budget, heterogeneous multi-model trajectories yield stronger memory than self- or same-model multi-rollout memory: the margin comes from contrastive behavioral diversity, not stronger source agents or more sampling.

cs.AI↗

arXiv2026

MatToolBench: Benchmarking Multimodal Agents in Real-World Materials Science Workflows

Multimodal GUI agents have achieved impressive results on general software benchmarks, yet their ability to operate professional scientific software remains largely unexplored. In materials science, sparse domain-specific web data, specialized interfaces, and tacit workflow conventions create blind spots that general-purpose pretraining cannot readily bridge. We present MatToolBench, the first real-environment benchmark for evaluating multimodal GUI agents on professional materials science software, comprising 204 tasks across 10 tools in three modalities: GUI operation, OriginPro scripting, and code-based database queries, all executed inside a Windows 11 VM. Each task is decomposed into fine-grained sub-criteria by domain experts, enabling interpretable partial-credit scoring; the GUI component of our multi-level evaluation pipeline achieves an average F1 of 0.98. For OriginPro figure-generation tasks, we further conduct a human-LLM agreement study to validate the use of a multimodal judge for secondary aesthetic assessment. Our experiments show that strong performance on general benchmarks does not transfer to professional scientific workflows, and that this gap is not a visual-grounding problem alone: failures arise from domain-specific operational knowledge, sparse pretraining coverage of scientific software, weak cross-tool artifact handoff, and critical states exposed only visually. Even the best model reaches only 25% success rate on GUI tasks and 45% on code tasks. MatToolBench therefore serves as a challenging diagnostic benchmark and real-environment testbed for data-scarce, knowledge-intensive scientific workflows.

cs.AI↗

arXiv2026

Co-Investment with Payoff-Sharing Mechanism for Cooperative Decision-Making in Network Design Games

Network-based systems are inherently interconnected, with the design and performance of subnetworks being interdependent. However, the decisions of self-interested operators may lead to suboptimal outcomes for users and the overall system. This paper explores cooperative mechanisms that can simultaneously benefit both operators and users. We address this challenge using a game-theoretical framework that integrates both non-cooperative and cooperative game theory. In the non-cooperative stage, we propose a network design game in which subnetwork decision-makers strategically design local infrastructures. In the cooperative stage, co-investment with payoff-sharing mechanism is developed to enlarge collective benefits and fairly distribute them. To demonstrate the effectiveness of our framework, we conduct case studies on the Sioux Falls network and real-world public transport networks in Zurich and Winterthur, Switzerland. Our evaluation considers impacts on environmental sustainability, social welfare, and economic efficiency. The proposed framework provides a foundation for improving interdependent networked systems by enabling strategic cooperation among self-interested operators.

eess.SY↗

arXiv2026

Denoising as Projection: Constrained Optimization with Gradient-Guided Diffusion

Diffusion models are increasingly used not only for sampling from learned data distributions, but also for generating samples that optimize task-specific objectives. A common approach is to guide the reverse diffusion process using gradients of an external objective. However, when the data distribution is supported on a structured feasible set, such as a manifold or a constraint set, gradient guidance can move samples away from the learned data geometry. In this paper, we study a simple projected-gradient-guided diffusion update based on the observation that the Stein denoising operator can act as an approximate projection onto the data geometry. The proposed update incorporates the objective gradient inside the denoising step, yielding an inference-time method that uses only a pretrained denoiser and gradient evaluations. We analyze this update as an inexact projected-gradient method for constrained optimization over learned feasible geometries. Our theory covers three settings: linear manifolds, compact convex feasible sets, and compact Riemannian submanifolds. In all these settings, we prove descent and finite-time convergence guarantees. Numerical experiments support the theoretical interpretation and illustrate how the proposed update balances objective descent with preservation of the learned geometry.

cs.LG↗

arXiv2026

PersonaEval: Persona-Based User Simulation for Evaluating Interactive Applications

Real user studies are important for understanding how people interact with systems under test or already deployed. In practice, however, they are often costly, time-consuming, and difficult to scale. To address these challenges, we introduce PersonaEval, a persona-based user simulation framework that approximates real-user behavior across diverse interactive settings. PersonaEval connects simulated users drawn from existing persona datasets to task-specific application interfaces and collects the interaction trajectories and outcomes. PersonaEval provides a plug-and-play evaluation workflow in which the application being evaluated can be easily changed. In this demo, we present PersonaEval on three forms of interactive applications: surveys, chatbots, and web applications. Together, these examples show that PersonaEval can support repeatable, parallelizable, and scalable evaluation across different interaction settings, while producing user-oriented feedback and task-specific behavior.

cs.HC↗

arXiv2026

MicroVerse: An Instrument for Measuring Self-Authored Identity Drift in Long-Horizon Multi-Agent Language-Model Simulations

Long-horizon, multi-agent language model (LM) simulations are widely proposed for studying social behavior, yet instruments to measure whether persona-conditioned agents maintain identity fidelity under sustained pressure are lacking. We present MicroVerse, a behavioral-science instrument that measures identity drift in generative agents. Agents carry an immutable "soul file" (core values, moral boundaries, personality, goals) and inhabit a resource-scarce 50 x 50 environment where water is a non-respawning survival constraint. Scarcity is operationalized via a per-tick existence-cost gradient. The eight-verb action space maps directly to moral boundaries (trade, talk, attack, scavenge). Using a three-layer memory architecture, agents periodically revise a mutable current identity against their immutable original soul via importance-triggered reflection. To mitigate survivor bias, MicroVerse decouples measurement from behavior using uniform longitudinal engine snapshots every N ticks alongside a forced-end snapshot of all living and dead agents. Identity drift is scored offline using a paraphrase-aware, value-anchored, multi-register diff rather than raw cosine similarity. We evaluate the instrument via a controlled seed run (n = 25) and a reflection-threshold sweep (thresholds {40, 80, 150}) to determine if drift dynamics are gate artifacts or threshold-robust properties. We report two primary findings: (1) Anti-self-deception emerges unprompted as the single largest semantic category of identity modification (27 of 111 added boundaries, 24%). (2) The system is threshold-robust; lower gates accelerate and increase revision frequency but preserve drift direction. All empirical results are strictly preliminary existence proofs and effect shapes (one model, one seed per arm, n = 25) rather than statistical significance claims.

cs.CL↗

arXiv2026

MatrAIx: Simulating the World with 8.3 Billion Persona Agents

Human evaluation of AI systems and digital products is costly, slow, and difficult to scale. Offline evaluations are more scalable but often abstract away human diversity and interactive behavior. We therefore introduce MatrAIx, a population-scale simulated-user evaluation infrastructure for testing AI systems and digital products with heterogeneous users. MatrAIx has three core components: First, Persona 8B contains 8.3 billion persona records represented by 1,290 categorical dimensions. Records are either sampled from a dependency graph that preserves correlated attributes or derived from human-authored profiles. We release a quality-filtered coreset of approximately 1 million personas, comprising 599,847 human-grounded and 400,000 synthetic records. Second, the MatrAIx Playground provides four environments in which diverse users evaluate and interact with digital products: Survey, AI Chatbot, Web, and App. Third, MatrAIx provides 1,010 application tasks spanning more than 25 domains, including Commerce, Software, Finance, and Healthcare. We conducted 18,189 evaluation trials across eight representative tasks. Persona agents were powered by three LLMs: Claude Opus 4.8, GPT 5.5, and Claude Haiku 4.5. The resulting feedback captures how decisions and preferences vary across persona backgrounds, including hesitation after a price increase, willingness to continue after an AI assistant fails, and latency tolerance. We conducted two main validation studies: First, a 400-trial controlled study evaluated persona adherence across ten behavioral attributes and all four environments. The declared behavior was expressed or correctly suppressed in 366 trials (91.5%). Second, human and LLM judges evaluated the extraction quality of human-grounded personas. Overall, MatrAIx provides an end-to-end infrastructure for evaluating AI systems and digital products with diverse simulated human users.

cs.AI↗

arXiv2026

Adaptive-Horizon Conflict-Based Search for Closed-Loop Multi-Agent Path Finding

Multi-Agent Path Finding (MAPF) is a core coordination problem for large robot fleets in automated warehouses and logistics. Existing approaches are typically either open-loop planners, which must compute complete trajectories before execution and therefore may incur substantial planning latency before actions can be taken, or closed-loop heuristics without reliable performance guarantees, limiting their use in safety-critical deployments. This paper presents Anytime Closed-Loop Conflict-Based Search (ACCBS), a closed-loop algorithm built on a finite-horizon variant of Conflict-Based Search (CBS) with a horizon-changing mechanism inspired by iterative horizon-deepening in Model Predictive Control (MPC). ACCBS dynamically adjusts the planning horizon based on the available computational budget, and reuses a single constraint tree to enable seamless transitions between horizons. As a result, it produces high-quality feasible solutions quickly while being asymptotically optimal as the budget increases, exhibiting anytime behavior. Extensive case studies demonstrate that ACCBS achieves a favorable balance between computational efficiency, solution quality, and execution flexibility, while naturally accommodating online disturbances through its closed-loop formulation.

cs.RO↗

arXiv2026

Implementation-Based Incentive Design for Autonomous Mobility-on-Demand and Transit Systems

Achieving a socially desirable operating point for a multimodal transportation system is challenging when Autonomous Mobility-on-Demand (AMoD) and Public Transit (PT) operators pursue selfish objectives alongside endogenous passenger choices. Existing equilibrium-based regulation models typically search over municipal policies to predict the induced operator equilibrium, creating strong behavioral assumptions, equilibrium-selection issues, and difficult bilevel optimization problems. This paper proposes an implementation-based alternative. Rather than asking which municipal action induces the best equilibrium, we ask: given a target operating profile, what minimum realized transfer makes unilateral deviation unattractive for each operator? Using k-implementation theory, this payment decomposes into two unilateral deviation gains: one for the AMoD operator and one for PT. Calculating this payment requires computing three distinct objects: the social target, the AMoD best response, and the PT best response. This is nontrivial because each represents a large-scale network optimization problem complicated by endogenous mode choice and congestion. To address this, we develop tailored mathematical formulations and algorithms for each oracle. For the social target, we derive a decomposition and entropy-regularized mixed-integer convex formulation balancing social optimality and implementability. For the AMoD oracle, we derive an exact reformulation, a convex relaxation providing a global upper bound, and a sequential convex approximation for feasible lower bounds. For the PT oracle, we develop a mixed-integer convex relaxation and characterize its exactness condition. A NYC case study shows the framework computes tight implementation-payment bounds and reveals how the dominant source of incentive misalignment shifts with congestion.

math.OC↗

arXiv2026

Muon-OGD: Muon-based Spectral Orthogonal Gradient Projection for LLM Continual Learning

A central challenge in continual learning for large language models (LLMs) is catastrophic forgetting, where adapting to new tasks can substantially degrade performance on previously learned ones. Existing projection-based methods mitigate such interference by restricting parameter updates to subspaces that are orthogonal to directions associated with past tasks. However, these methods are typically formulated under Euclidean parameter geometry, with update magnitudes and projections governed by the Frobenius norm. The recent empirical success of the Muon optimizer, which applies orthogonalized matrix updates and admits a spectral-norm interpretation, suggests that Frobenius geometry may not be the most effective choice for matrix-valued LLM parameters. Motivated by this observation, we propose Muon-OGD, a spectral-norm-aware continual learning framework that integrates Muon-style operator-norm geometry with orthogonal projection constraints. Our method formulates each update as a spectral-norm-constrained optimization problem with linear non-interference constraints, and solves it efficiently through dual iterations and Newton--Schulz matrix-sign approximations. By applying orthogonalized momentum updates that avoid protected directions associated with prior tasks, Muon-OGD aims to improve the stability--plasticity trade-off in sequential LLM adaptation. We evaluate the proposed method on standard continual learning benchmarks, TRACE, and domain-specific Coding--Math--Medical curricula using both encoder--decoder and decoder-only architectures. Empirically, Muon-OGD consistently improves over sequential fine-tuning and competitive orthogonal-gradient baselines, while remaining computationally scalable. These results suggest that spectral-norm-aware update geometry provides a practical and effective alternative to Frobenius-norm projection for continual learning in LLMs.

cs.LG↗

arXiv2026

AdamFLIP: Adaptive Momentum Feedback Linearization Optimization for Hard Constrained PINN Training

Physics-informed neural networks (PINNs) provide a flexible framework for solving forward and inverse problems governed by partial differential equations (PDEs), but standard PINN training typically relies on soft penalty formulations that combine PDE residuals, data mismatch, and initial/boundary conditions using manually chosen weights. This often leads to ill-conditioning, sensitivity to loss weights, and poor constraint satisfaction. In this work, we reformulate PINN training as an equality-constrained optimization problem and propose a novel Adaptive Momentum Feedback Linearization Optimization for Hard Constrained PINN (AdamFLIP). The key idea is to view the constraint residuals as the output of a controlled dynamical system and to compute the Lagrange multiplier as a feedback input that locally drives these residuals toward stable linear contraction dynamics. AdamFLIP then applies Adam-style first- and second-moment adaptation to the resulting feedback-linearized Lagrangian gradient, combining principled constraint handling with the scalability and robustness of adaptive neural-network optimization. We test AdamFLIP on a range of benchmark forward and inverse PDE problem, and it consistently outperforms both the standard soft-constrained PINN and state-of-the-art constrained optimizers. Specifically, on the Navier--Stokes equations benchmark, AdamFLIP \textbf{reduces relative $L_2$ error by more than two thirds} for the predicted solution compared to the next best method. Our AdamFLIP framework provides an effective and computationally scalable hard constraint optimization method for PINN training.

cs.LG↗

arXiv2026

Random-Subspace Sequential Quadratic Programming for Constrained Zeroth-Order Optimization

We study nonlinear constrained optimization problems in which only function evaluations of the objective and constraints are available. Existing zeroth-order methods rely on noisy gradient and Jacobian surrogates in high dimensions, making it difficult to simultaneously achieve computational efficiency and accurate constraint satisfaction. We propose a zeroth-order random-subspace sequential quadratic programming method (ZO-RS-SQP) that combines two-point directional estimation with low-dimensional SQP updates. At each iteration, the method samples a random low-dimensional subspace, estimates the projected objective gradient and constraint Jacobians using two-point evaluations, and solves a reduced quadratic program to compute the step. As a result, the per-iteration evaluation cost scales with the subspace dimension rather than the ambient dimension, while retaining the structured linearized-constraint treatment of SQP. We also consider an Armijo line-search variant that improves robustness in practice. Under standard smoothness and regularity assumptions, we establish convergence to first-order KKT points with high probability. Numerical experiments illustrate the effectiveness of the proposed approach on nonlinear constrained problems.

math.OC↗

arXiv2026

Certificate-Driven Closed-Loop Multi-Agent Path Finding with Inheritable Factorization

Multi-agent coordination in automated warehouses and logistics is commonly modeled as the Multi-Agent Path Finding (MAPF) problem. Closed-loop MAPF algorithms improve scalability by planning only the next movement and replanning online, but this finite-horizon viewpoint can be shortsighted and makes it difficult to preserve global guarantees and exploit compositional structure. This issue is especially visible in Anytime Closed-Loop Conflict-Based Search (ACCBS), which applies Conflict-Based Search (CBS) over dynamically extended finite horizons but, under finite computational budgets, may terminate with short active prefixes in dense instances. We introduce certificate trajectories and their associated fleet budget as a general mechanism for filtering closed-loop updates. A certificate provides a conflict-free fallback plan and a monotone upper bound on the remaining cost; accepting only certificate-improving updates yields completeness. The same budget information induces a budget-limited factorization that enables global, inheritable decomposition across timesteps. Instantiating the framework on ACCBS yields Certificate-Driven Conflict-Based Search (CDCBS). Experiments on benchmark maps show that CDCBS achieves more consistent solution quality than ACCBS, particularly in dense settings, while the proposed factorization reduces effective group size.

cs.RO↗

arXiv2026

Equilibrium Selection for Multi-agent Reinforcement Learning: A Unified Framework

While multi-agent reinforcement learning (MARL) has produced numerous algorithms that converge to Nash or related equilibria, such equilibria are often non-unique and can exhibit widely varying efficiency. This raises a fundamental question: how can one design learning dynamics that not only converge to equilibrium but also select equilibria with desirable performance, such as high social welfare? In contrast to the MARL literature, equilibrium selection has been extensively studied in normal-form games, where decentralized dynamics are known to converge to potential-maximizing or Pareto-optimal Nash equilibria (NEs). Motivated by these results, we study equilibrium selection in finite-horizon stochastic games. We propose a unified actor-critic framework in which a critic learns state-action value functions, and an actor applies a classical equilibrium-selection rule state-wise, treating learned values as stage-game payoffs. We show that, under standard stochastic stability assumptions, the stochastically stable policies of the resulting dynamics inherit the equilibrium selection properties of the underlying normal-form learning rule. As consequences, we obtain potential-maximizing policies in Markov potential games and Pareto-optimal (Markov perfect) equilibria in general-sum stochastic games, together with sample-based implementation of the framework.

cs.GT↗

arXiv2026

Constrained Optimization From a Control Perspective via Feedback Linearization

Tools from control and dynamical systems have proven valuable for analyzing and developing optimization methods. In this paper, we establish rigorous theoretical foundations for using feedback linearization (FL) -- a well-established nonlinear control technique -- to solve constrained optimization problems. For equality-constrained optimization, we establish global convergence rates to first-order Karush-Kuhn-Tucker (KKT) points and uncover the close connection between the FL method and the Sequential Quadratic Programming (SQP) algorithm. Building on this relationship, we extend the FL approach to handle inequality-constrained problems. Furthermore, we introduce a momentum-accelerated feedback linearization algorithm and provide a rigorous convergence guarantee.

math.OC↗

arXiv2026

Zeroth-Order Constrained Optimization from a Control Perspective via Feedback Linearization

Safe derivative-free optimization under unknown constraints is a fundamental challenge in modern learning and control. Existing zeroth-order (ZO) methods typically still assume access to a first-order oracle of the constraint functions or restrict attention to convex settings, leaving nonconvex optimization with black-box constraints largely unexplored. We propose the zeroth-order feedback-linearization (ZOFL) algorithm for ZO constrained optimization that enforces feasibility without access to the first-order oracle of the constraint functions and applies to both equality and inequality constraints. The proposed approach relies only on noisy, sample-based gradient estimates obtained via two-point estimators, yet provably guarantees constraint satisfaction under mild regularity conditions. It adopts a control-theoretic perspective on ZO constrained optimization and leverages feedback linearization, a nonlinear control technique, to enforce feasibility. Finite-time bounds on constraint violation and asymptotic global convergence guarantees are established for the ZOFL algorithm. A midpoint discretization variant is further developed to improve feasibility without sacrificing optimality. Empirical results demonstrate that ZOFL consistently outperforms standard ZO baselines, achieving competitive objective values while maintaining feasibility.

math.OC↗