arXiv subjects

Jesse Milzman

Publications and source records attributed to Jesse Milzman.

At least 19 recordsLinked to original sources

arXiv2026

On Type Deception in Linear-Quadratic Differential Games

We consider two-player linear-quadratic differential games of incomplete information, in which one player has a private type initially unknown to the other. The typed player has incentive to conceal their type, while the uninformed player has the potential to infer it during play. Any ex-ante equilibrium in this setting will decompose into a deceptive, pooling phase, and a complete-information, revelatory phase. We demonstrate how to solve both phases via nested Riccati equations. Candidate equilibria are then found by maximizing the game value over a scalar revelation time, for which we provide a gradient in the case of time-homogeneous system matrices. We conclude by demonstrating our framework in a pursuit-evasion game with time-varying control advantages, finding interior optimal revelation times that confirm deception has quantifiable ex-ante value.

cs.GT↗

arXiv2026

SACHI: Structured Agent Coordination via Holistic Information Integration in Multi-Agent Reinforcement Learning

Cooperative multi-agent reinforcement learning agents that act on partial local observations face a fundamental information bottleneck: the knowledge needed to select jointly optimal actions is scattered across the team, yet each agent must commit to a decision without access to its teammates' observations, intentions, or chosen actions. Existing methods either ignore this bottleneck, compress it into a scalar mixing signal, or route around it with learned communication channels. Framing action coordination as a problem of structured information integration among agents, we propose \textit{structured agent coordination via holistic information integration}, or SACHI, in which graph transformer convolutions over an inter-agent coordination graph enrich each agent's representation with receiver-sensitive, content-dependent signals from teammates prior to action selection. We evaluate SACHI across five cooperative tasks spanning spatial, communicative, and adversarial coordination challenges against twelve baselines. SACHI consistently matches or outperforms the best baseline on every task, and rigorous aggregate statistical analyses, including normalized metrics with bootstrap confidence intervals, Friedman ranking, and performance profiling, confirm that this advantage is statistically significant, robust across environments, and not attributable to increased model capacity. Parameter-matched ablations further trace the source of the gains to a single architectural property: the degree of content-dependence in the message-passing operator.

cs.LG↗

arXiv2026

NeuroMesh: A Unified Neural Inference Framework for Decentralized Multi-Robot Collaboration

Deploying learned multi-robot models on heterogeneous robots remains challenging due to hardware heterogeneity, communication constraints, and the lack of a unified execution stack. This paper presents NeuroMesh, a multi-domain, cross-platform, and modular decentralized neural inference framework that standardizes observation encoding, message passing, aggregation, and task decoding in a unified pipeline. NeuroMesh combines a dual-aggregation paradigm for reduction- and broadcast-based information fusion with a parallelized architecture that decouples cycle time from end-to-end latency. Our high-performance C++ implementation leverages Zenoh for inter-robot communication and supports hybrid GPU/CPU inference. We validate NeuroMesh on a heterogeneous team of aerial and ground robots across collaborative perception, decentralized control, and task assignment, demonstrating robust operation across diverse task structures and payload sizes. We plan to release NeuroMesh as an open-source framework to the community.

cs.RO↗

arXiv2026

Linear-Quadratic Gaussian Games with Distributed Sparse Estimation

Linear-quadratic Gaussian games provide a framework for modeling strategic interactions in multi-agent systems, where agents must estimate system states from noisy observations while also making decisions to optimize a quadratic cost. However, these formulations usually require agents to utilize the full set of available observations when forming their state estimates, which can be unrealistic in large-scale or resource-constrained settings. In this paper, we consider linear-quadratic Gaussian games with sparse interagent observations. To enforce sparsity in the estimation stage, we design a distributed estimator that balances estimation effectiveness with interagent measurement sparsity via a group lasso problem, while agents implement feedback Nash strategies based on their state estimates. We provide sufficient conditions under which the sparse estimator is guaranteed to trigger a corrective reset to the optimal estimation gain, ensuring that estimation quality does not degrade beyond a level determined by the regularization parameters. Simulations on a formation game show that the proposed approach yields a significant reduction in communication resources consumed while only minimally affecting the nominal equilibrium trajectories.

eess.SY↗

arXiv2026

Action-Graph Policies: Learning Action Co-dependencies in Multi-Agent Reinforcement Learning

Coordinating actions is the most fundamental form of cooperation in multi-agent reinforcement learning (MARL). Successful decentralized decision-making often depends not only on good individual actions, but on selecting compatible actions across agents to synchronize behavior, avoid conflicts, and satisfy global constraints. In this paper, we propose Action Graph Policies (AGP), that model dependencies among agents' available action choices. It constructs, what we call, \textit{coordination contexts}, that enable agents to condition their decisions on global action dependencies. Theoretically, we show that AGPs induce a strictly more expressive joint policy compared to fully independent policies and can realize coordinated joint actions that are provably more optimal than greedy execution even from centralized value-decomposition methods. Empirically, we show that AGP achieves 80-95\% success on canonical coordination tasks with partial observability and anti-coordination penalties, where other MARL methods reach only 10-25\%. We further demonstrate that AGP consistently outperforms these baselines in diverse multi-agent environments.

cs.LG↗

arXiv2026

Efficiently Solving Mixed-Hierarchy Games with Quasi-Policy Approximations

Multi-robot coordination often exhibits hierarchical structure, with some robots' decisions depending on the planned behaviors of others. While game theory provides a principled framework for such interactions, existing solvers struggle to handle mixed information structures that combine simultaneous (Nash) and hierarchical (Stackelberg) decision-making. We study N-robot forest-structured mixed-hierarchy games, in which each robot acts as a Stackelberg leader over its subtree while robots in different branches interact via Nash equilibria. We derive the Karush-Kuhn-Tucker (KKT) first-order optimality conditions for this class of games and show that they involve increasingly high-order derivatives of robots' best-response policies as the hierarchy depth grows, rendering a direct solution intractable. To overcome this challenge, we introduce a quasi-policy approximation that removes higher-order policy derivatives and develop an inexact Newton method for efficiently solving the resulting approximated KKT systems. We prove local exponential convergence of the proposed algorithm for games with non-quadratic objectives and nonlinear constraints. The approach is implemented in a highly optimized Julia library (MixedHierarchyGames.jl) and evaluated in hardware and simulated multi-agent experiments, demonstrating real-time convergence for complex mixed-hierarchy information structures.

cs.GT↗

arXiv2026

Generalized Information Gathering Under Dynamics Uncertainty

An agent operating in an unknown dynamical system must learn its dynamics from observations. Active information gathering accelerates this learning, but existing methods derive bespoke costs for specific modeling choices: dynamics models, belief update procedures, observation models, and planners. We present a unifying framework that decouples these choices from the information-gathering cost by explicitly exposing the causal dependencies between parameters, beliefs, and controls. Using this framework, we derive a general information-gathering cost based on Massey's directed information that assumes only Markov dynamics with additive noise and is otherwise agnostic to modeling choices. We prove that the mutual information cost used in existing literature is a special case of our cost. Then, we leverage our framework to establish an explicit connection between the mutual information cost and information gain in linearized Bayesian estimation, thereby providing theoretical justification for mutual information-based active learning approaches. Finally, we illustrate the practical utility of our framework through experiments spanning linear, nonlinear, and multi-agent systems.

cs.LG↗

arXiv2025

TIGER-MARL: Enhancing Multi-Agent Reinforcement Learning with Temporal Information through Graph-based Embeddings and Representations

In this paper, we propose capturing and utilizing \textit{Temporal Information through Graph-based Embeddings and Representations} or \textbf{TIGER} to enhance multi-agent reinforcement learning (MARL). We explicitly model how inter-agent coordination structures evolve over time. While most MARL approaches rely on static or per-step relational graphs, they overlook the temporal evolution of interactions that naturally arise as agents adapt, move, or reorganize cooperation strategies. Capturing such evolving dependencies is key to achieving robust and adaptive coordination. To this end, TIGER constructs dynamic temporal graphs of MARL agents, connecting their current and historical interactions. It then employs a temporal attention-based encoder to aggregate information across these structural and temporal neighborhoods, yielding time-aware agent embeddings that guide cooperative policy learning. Through extensive experiments on two coordination-intensive benchmarks, we show that TIGER consistently outperforms diverse value-decomposition and graph-based MARL baselines in task performance and sample efficiency. Furthermore, we conduct comprehensive ablation studies to isolate the impact of key design parameters in TIGER, revealing how structural and temporal factors can jointly shape effective policy learning in MARL. All codes can be found here: https://github.com/Nikunj-Gupta/tiger-marl.

cs.LG↗

arXiv2025

Deceptive Exploration in Multi-armed Bandits

We consider a multi-armed bandit setting in which each arm has a public and a private reward distribution. An observer expects an agent to follow Thompson Sampling according to the public rewards, however, the deceptive agent aims to quickly identify the best private arm without being noticed. The observer can observe the public rewards and the pulled arms, but not the private rewards. The agent, on the other hand, observes both the public and private rewards. We formalize detectability as a stepwise Kullback-Leibler (KL) divergence constraint between the actual pull probabilities used by the agent and the anticipated pull probabilities by the observer. We model successful pulling of public suboptimal arms as a % Bernoulli process where the success probability decreases with each successful pull, and show these pulls can happen at most at a $\Theta(\sqrt{T}) $ rate under the KL constraint. We then formulate a maximin problem based on public and private means, whose solution characterizes the optimal error exponent for best private arm identification. We finally propose an algorithm inspired by top-two algorithms. This algorithm naturally adapts its exploration according to the hardness of pulling arms based on the public suboptimality gaps. We provide numerical examples illustrating the $\Theta(\sqrt{T}) $ rate and the behavior of the proposed algorithm.

cs.LG↗

arXiv2025

Deceptive Planning Exploiting Inattention Blindness

We study decision-making with rational inattention in settings where agents have perception constraints. In such settings, inaccurate prior beliefs or models of others may lead to inattention blindness, where an agent is unaware of its incorrect beliefs. We model this phenomenon in two-player zero-sum stochastic games, where Player 1 has perception constraints and Player 2 deceptively deviates from its security policy presumed by Player 1 to gain an advantage. We formulate the perception constraints as an online sensor selection problem, develop a value-weighted objective function for sensor selection capturing rational inattention, and propose the greedy algorithm for selection under this monotone objective function. When Player 2 does not deviate from the presumed policy, this objective function provides an upper bound on the expected value loss compared to the security value where Player 1 has perfect information of the state. We then propose a myopic decision-making algorithm for Player 2 to exploit Player 1's beliefs by deviating from the presumed policy and, thereby, improve upon the security value. Numerical examples illustrate how Player 1 persistently chooses sensors that are consistent with its priors, allowing Player 2 to systematically exploit its inattention.

cs.GT↗

arXiv2025

Smooth Games of Configuration in the Linear-Quadratic Setting

Dynamic game theory offers a toolbox for formalizing and solving for both cooperative and non-cooperative strategies in multi-agent scenarios. However, the optimal configuration of such games remains largely unexplored. While there is existing literature on the parametrization of dynamic games, little research examines this parametrization from a strategic perspective where each agent's configuration choice is influenced by the decisions of others. In this work, we introduce the concept of a game of configuration, providing a framework for the strategic fine-tuning of differential games. We define a game of configuration as a two-stage game within the setting of finite-horizon, affine-quadratic, AQ, differential games. In the first stage, each player chooses their corresponding configuration parameter, which will impact their dynamics and costs in the second stage. We provide the subgame perfect solution concept and a method for computing first stage cost gradients over the configuration space. This then allows us to formulate a gradient-based method for searching for local solutions to the configuration game, as well as provide necessary conditions for equilibrium configurations over their downstream (second stage) trajectories. We conclude by demonstrating the effectiveness of our approach in example AQ systems, both zero-sum and general-sum.

cs.MA↗

arXiv2025

MEL: Multi-level Ensemble Learning for Resource-Constrained Environments

AI inference at the edge is becoming increasingly common for low-latency services. However, edge environments are power- and resource-constrained, and susceptible to failures. Conventional failure resilience approaches, such as cloud failover or compressed backups, often compromise latency or accuracy, limiting their effectiveness for critical edge inference services. In this paper, we propose Multi-Level Ensemble Learning (MEL), a new framework for resilient edge inference that simultaneously trains multiple lightweight backup models capable of operating collaboratively, refining each other when multiple servers are available, and independently under failures while maintaining good accuracy. Specifically, we formulate our approach as a multi-objective optimization problem with a loss formulation that inherently encourages diversity among individual models to promote mutually refining representations, while ensuring each model maintains good standalone performance. Empirical evaluations across vision, language, and audio datasets show that MEL provides performance comparable to original architectures while also providing fault tolerance and deployment flexibility across edge platforms. Our results show that our ensemble model, sized at 40\% of the original model, achieves similar performance, while preserving 95.6\% of ensemble accuracy in the case of failures when trained using MEL.

cs.LG↗

arXiv2025

FailLite: Failure-Resilient Model Serving for Resource-Constrained Edge Environments

Model serving systems have become popular for deploying deep learning models for various latency-sensitive inference tasks. While traditional replication-based methods have been used for failure-resilient model serving in the cloud, such methods are often infeasible in edge environments due to significant resource constraints that preclude full replication. To address this problem, this paper presents FailLite, a failure-resilient model serving system that employs (i) a heterogeneous replication where failover models are smaller variants of the original model, (ii) an intelligent approach that uses warm replicas to ensure quick failover for critical applications while using cold replicas, and (iii) progressive failover to provide low mean time to recovery (MTTR) for the remaining applications. We implement a full prototype of our system and demonstrate its efficacy on an experimental edge testbed. Our results using 27 models show that FailLite can recover all failed applications with 175.5ms MTTR and only a 0.6% reduction in accuracy.

cs.DC↗

arXiv2025

Value of Information-based Deceptive Path Planning Under Adversarial Interventions

Existing methods for deceptive path planning (DPP) address the problem of designing paths that conceal their true goal from a passive, external observer. Such methods do not apply to problems where the observer has the ability to perform adversarial interventions to impede the path planning agent. In this paper, we propose a novel Markov decision process (MDP)-based model for the DPP problem under adversarial interventions and develop new value of information (VoI) objectives to guide the design of DPP policies. Using the VoI objectives we propose, path planning agents deceive the adversarial observer into choosing suboptimal interventions by selecting trajectories that are of low informational value to the observer. Leveraging connections to the linear programming theory for MDPs, we derive computationally efficient solution methods for synthesizing policies for performing DPP under adversarial interventions. In our experiments, we illustrate the effectiveness of the proposed solution method in achieving deceptiveness under adversarial interventions and demonstrate the superior performance of our approach to both existing DPP methods and conservative path planning approaches on illustrative gridworld problems.

cs.LG↗

arXiv2025

Deep Meta Coordination Graphs for Multi-agent Reinforcement Learning

This paper presents deep meta coordination graphs (DMCG) for learning cooperative policies in multi-agent reinforcement learning (MARL). Coordination graph formulations encode local interactions and accordingly factorize the joint value function of all agents to improve efficiency in MARL. Through DMCG, we dynamically compose what we refer to as \textit{meta coordination graphs}, to learn a more expressive representation of agent interactions and use them to integrate agent information through graph convolutional networks. The goal is to enable an evolving coordination graph to guide effective coordination in cooperative MARL tasks. The graphs are jointly optimized with agents' value functions to learn to implicitly reason about joint actions, facilitating the end-to-end learning of interaction representations and coordinated policies. We demonstrate that DMCG consistently achieves state-of-the-art coordination performance and sample efficiency on challenging cooperative tasks, outperforming several prior graph-based and non-graph-based MARL baselines. Through several ablations, we also isolate the impact of individual components in DMCG, showing that the observed improvements are due to the meaningful design choices in this approach. We also include an analysis of its computational complexity to discuss its practicality in real-world applications. All codes can be found here: {\color{blue}{https://github.com/Nikunj-Gupta/dmcg-marl}.

cs.LG↗

arXiv2024

Approximate Feedback Nash Equilibria with Sparse Inter-Agent Dependencies

Feedback Nash equilibrium strategies in multi-agent dynamic games require availability of all players' state information to compute control actions. However, in real-world scenarios, sensing and communication limitations between agents make full state feedback expensive or impractical, and such strategies can become fragile when state information from other agents is inaccurate. To this end, we propose a regularized dynamic programming approach for finding sparse feedback policies that selectively depend on the states of a subset of agents in dynamic games. The proposed approach solves convex adaptive group Lasso problems to compute sparse policies approximating Nash equilibrium solutions. We prove the regularized solutions' asymptotic convergence to a neighborhood of Nash equilibrium policies in linear-quadratic (LQ) games. Further, we extend the proposed approach to general non-LQ games via an iterative algorithm. Simulation results in multi-robot interaction scenarios show that the proposed approach effectively computes feedback policies with varying sparsity levels. When agents have noisy observations of other agents' states, simulation results indicate that the proposed regularized policies consistently achieve lower costs than standard Nash equilibrium policies by up to 77% for all interacting agents whose costs are coupled with other agents' states.

cs.GT↗

arXiv2024

Sensing Resource Allocation Against Data-Poisoning Attacks in Traffic Routing

Data-poisoning attacks can disrupt the efficient operations of transportation systems by misdirecting traffic flows via falsified data. One challenge in countering these attacks is to reduce the uncertainties on the types of attacks, such as the distribution of their targets and intensities. We introduce a resource allocation method in transportation networks to detect and distinguish different types of attacks and facilitate efficient traffic routing. The idea is to first cluster different types of attacks based on the corresponding optimal routing strategies, then allocate sensing resources to a subset of network links to distinguish attacks from different clusters via lexicographical mixed-integer programming. We illustrate the application of the proposed method using the Anaheim network, a benchmark model in traffic routing that contains more than 400 nodes and 900 links.

math.OC↗

arXiv2024

Measuring the Redundancy of Information from a Source Failure Perspective

In this paper, we define a new measure of the redundancy of information from a fault tolerance perspective. The partial information decomposition (PID) emerged last decade as a framework for decomposing the multi-source mutual information $I(T;X_1, ..., X_n)$ into atoms of redundant, synergistic, and unique information. It built upon the notion of redundancy/synergy from McGill's interaction information (McGill 1954). Separately, the redundancy of system components has served as a principle of fault tolerant engineering, for sensing, routing, and control applications. Here, redundancy is understood as the level of duplication necessary for the fault tolerant performance of a system. With these two perspectives in mind, we propose a new PID-based measure of redundancy $I_{\text{ft}}$, based upon the presupposition that redundant information is robust to individual source failures. We demonstrate that this new measure satisfies the common PID axioms from (Williams 2010). In order to do so, we establish an order-reversing correspondence between collections of source-fallible instantiations of a system, on the one hand, and the PID lattice from (Williams 2010), on the other.

cs.IT↗