About

David Fridovich-Keil is an Assistant Professor in the Department of Aerospace Engineering and Engineering Mechanics at the University of Texas at Austin. He joined the department as an assistant professor in Fall 2021. His research spans optimal control, dynamic game theory, learning for control, and robot safety. His work investigates the role of dynamic game theory in multi-agent interactive settings such as traffic, and focuses on the interplay between machine learning and classical ideas from robust, adaptive, and geometric control theory. Dr. Fridovich-Keil received his doctorate from the University of California, Berkeley, where he developed some of the first efficient techniques for solving noncooperative, game-theoretic motion planning problems. During his graduate studies, he briefly worked at the self-driving car company Nuro. His postdoctoral research focused on exploiting computational parallelism in stochastic optimal control problems. His research interests include optimal control theory, dynamic game theory, motion planning, multi-agent decision making, and machine learning and control. He was awarded the National Science Foundation CAREER Award in 2024.

Research topics

• Computer Science
• Artificial Intelligence
• Mathematical optimization
• Mathematics
• Engineering
• Algorithm
• Control engineering

Selected publications

• Bayesian Inverse Games with High-Dimensional Multi-Modal Observations

  ArXiv.org · 2026-01-02

  articleOpen access

  Many multi-agent interaction scenarios can be naturally modeled as noncooperative games, where each agent's decisions depend on others' future actions. However, deploying game-theoretic planners for autonomous decision-making requires a specification of all agents' objectives. To circumvent this practical difficulty, recent work develops maximum likelihood techniques for solving inverse games that can identify unknown agent objectives from interaction data. Unfortunately, these methods only infer point estimates and do not quantify estimator uncertainty; correspondingly, downstream planning decisions can overconfidently commit to unsafe actions. We present an approximate Bayesian inference approach for solving the inverse game problem, which can incorporate observation data from multiple modalities and be used to generate samples from the Bayesian posterior over the hidden agent objectives given limited sensor observations in real time. Concretely, the proposed Bayesian inverse game framework trains a structured variational autoencoder with an embedded differentiable Nash game solver on interaction datasets and does not require labels of agents' true objectives. Extensive experiments show that our framework successfully learns prior and posterior distributions, improves inference quality over maximum likelihood estimation-based inverse game approaches, and enables safer downstream decision-making without sacrificing efficiency. When trajectory information is uninformative or unavailable, multimodal inference further reduces uncertainty by exploiting additional observation modalities.

  PublisherOA PDF

• Bayesian Inverse Games with High-Dimensional Multi-Modal Observations

  arXiv (Cornell University) · 2026-01-02

  preprintOpen access

  Many multi-agent interaction scenarios can be naturally modeled as noncooperative games, where each agent's decisions depend on others' future actions. However, deploying game-theoretic planners for autonomous decision-making requires a specification of all agents' objectives. To circumvent this practical difficulty, recent work develops maximum likelihood techniques for solving inverse games that can identify unknown agent objectives from interaction data. Unfortunately, these methods only infer point estimates and do not quantify estimator uncertainty; correspondingly, downstream planning decisions can overconfidently commit to unsafe actions. We present an approximate Bayesian inference approach for solving the inverse game problem, which can incorporate observation data from multiple modalities and be used to generate samples from the Bayesian posterior over the hidden agent objectives given limited sensor observations in real time. Concretely, the proposed Bayesian inverse game framework trains a structured variational autoencoder with an embedded differentiable Nash game solver on interaction datasets and does not require labels of agents' true objectives. Extensive experiments show that our framework successfully learns prior and posterior distributions, improves inference quality over maximum likelihood estimation-based inverse game approaches, and enables safer downstream decision-making without sacrificing efficiency. When trajectory information is uninformative or unavailable, multimodal inference further reduces uncertainty by exploiting additional observation modalities.

  PublisherDOI

• Decomposing Control Lyapunov Functions for Efficient Reinforcement Learning

  2025-07-08

  articleSenior author

  Recent methods using Reinforcement Learning (RL) have proven to be successful for training intelligent agents in unknown environments. However, current state-of-the-art RL methods require large amounts of data to learn a specific task, leading to unreasonable costs when deploying the agent to collect data in real-world applications. In this paper, we take a control-theoretic approach to improve sample efficiency in RL. Our approach has two key components. First, we build upon recent work which shows that the inclusion of a Control Lyapunov Function (CLF) in the reward function can enable RL algorithms to learn with substantially lower discount factors, and therefore require less data to train. To do this “reward shaping” requires access to a CLF, however, and computational methods to compute a CLF do not scale well to high-dimensional systems. Therefore, the primary contribution of our work is the development of a new variety of CLF-like functions, which we term Decomposed Control Lyapunov Functions (DCLFs). We show that these functions can be used in the same way as CLFs for RL reward shaping, yet are more readily computable in higher-dimensional cases via a system decomposition technique. Through multiple examples, we demonstrate the effectiveness of this approach, where our method finds a policy to successfully land a quadcopter with less than half the amount of real-world data required by two state-of-the-art RL algorithms.

  PublisherDOI

• You Can't Always Get What You Want: Games of Ordered Preference

  IEEE Robotics and Automation Letters · 2025-05-30 · 1 citations

  articleSenior author

  We study noncooperative games, in which each player's objective is composed of a sequence of ordered—and potentially conflicting—preferences. Problems of this type naturally model a wide variety of scenarios: for example, drivers at a busy intersection must balance the desire to make forward progress with the risk of collision. Mathematically, these problems possess a nested structure, and to behave properly players must prioritize their most important preference, and only consider less important preferences to the extent that they do not compromise performance on more important ones. We consider multi-agent, noncooperative variants of these problems, and seek generalized Nash equilibria in which each player's decision reflects both its hierarchy of preferences <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">and</i> other players' actions. We make two key contributions. First, we develop a recursive approach for deriving the first-order optimality conditions of each player's nested problem. Second, we propose a sequence of increasingly tight relaxations, each of which can be transcribed as a mixed complementarity problem and solved via existing methods. Experimental results demonstrate that our approach reliably converges to equilibrium solutions that strictly reflect players' individual ordered preferences.

  PublisherDOI

• Learning responsibility allocations for multi-agent interactions: A differentiable optimization approach with control barrier functions

  2025-07-08

  article

  From autonomous driving to package delivery, ensuring safe yet efficient multi-agent interaction is challenging as the interaction dynamics are influenced by hard-to-model factors such as social norms and contextual cues. Understanding these influences can aid in the design and evaluation of socially- aware autonomous agents whose behaviors are aligned with human values. In this work, we seek to codify factors governing safe multi-agent interactions via the lens of responsibility, i.e., an agent's willingness to deviate from their desired control to accommodate safe interaction with others. Specifically, we propose a data-driven modeling approach based on control barrier functions and differentiable optimization that efficiently learns agents' responsibility allocation from data. We demonstrate on synthetic and real-world datasets that we can obtain an interpretable and quantitative understanding of how much agents adjust their behavior to ensure the safety of others given their current environment.

  PublisherDOI

• UNCAP: Uncertainty-Guided Neurosymbolic Planning Using Natural Language Communication for Cooperative Autonomous Vehicles

  arXiv (Cornell University) · 2025-10-14

  preprintOpen access

  Safe large-scale coordination of multiple cooperative connected autonomous vehicles (CAVs) hinges on communication that is both efficient and interpretable. Existing approaches either rely on transmitting high-bandwidth raw sensor data streams or neglect perception and planning uncertainties inherent in shared data, resulting in systems that are neither scalable nor safe. To address these limitations, we propose Uncertainty-Guided Natural Language Cooperative Autonomous Planning (UNCAP), a vision-language model-based planning approach that enables CAVs to communicate via lightweight natural language messages while explicitly accounting for perception uncertainty in decision-making. UNCAP features a two-stage communication protocol: (i) an ego CAV first identifies the subset of vehicles most relevant for information exchange, and (ii) the selected CAVs then transmit messages that quantitatively express their perception uncertainty. By selectively fusing messages that maximize mutual information, this strategy allows the ego vehicle to integrate only the most relevant signals into its decision-making, improving both the scalability and reliability of cooperative planning. Experiments across diverse driving scenarios show a 63% reduction in communication bandwidth with a 31% increase in driving safety score, a 61% reduction in decision uncertainty, and a four-fold increase in collision distance margin during near-miss events. Project website: https://uncap-project.github.io/

  PublisherOA PDFDOI

• SHIELD: Safety on Humanoids via CBFs In Expectation on Learned Dynamics

  ArXiv.org · 2025-05-16

  preprintOpen access

  Robot learning has produced remarkably effective ``black-box'' controllers for complex tasks such as dynamic locomotion on humanoids. Yet ensuring dynamic safety, i.e., constraint satisfaction, remains challenging for such policies. Reinforcement learning (RL) embeds constraints heuristically through reward engineering, and adding or modifying constraints requires retraining. Model-based approaches, like control barrier functions (CBFs), enable runtime constraint specification with formal guarantees but require accurate dynamics models. This paper presents SHIELD, a layered safety framework that bridges this gap by: (1) training a generative, stochastic dynamics residual model using real-world data from hardware rollouts of the nominal controller, capturing system behavior and uncertainties; and (2) adding a safety layer on top of the nominal (learned locomotion) controller that leverages this model via a stochastic discrete-time CBF formulation enforcing safety constraints in probability. The result is a minimally-invasive safety layer that can be added to the existing autonomy stack to give probabilistic guarantees of safety that balance risk and performance. In hardware experiments on an Unitree G1 humanoid, SHIELD enables safe navigation (obstacle avoidance) through varied indoor and outdoor environments using a nominal (unknown) RL controller and onboard perception.

  PublisherOA PDFDOI

• Cooperative Bargaining Games Without Utilities: Mediated Solutions from Direction Oracles

  ArXiv.org · 2025-05-20

  preprintOpen accessSenior author

  Cooperative bargaining games are widely used to model resource allocation and conflict resolution. Traditional solutions assume the mediator can access agents utility function values and gradients. However, there is an increasing number of settings, such as human AI interactions, where utility values may be inaccessible or incomparable due to unknown, nonaffine transformations. To model such settings, we consider that the mediator has access only to agents most preferred directions, i.e., normalized utility gradients in the decision space. To this end, we propose a cooperative bargaining algorithm where a mediator has access to only the direction oracle of each agent. We prove that unlike popular approaches such as the Nash and Kalai Smorodinsky bargaining solutions, our approach is invariant to monotonic nonaffine transformations, and that under strong convexity and smoothness assumptions, this approach enjoys global asymptotic convergence to Pareto stationary solutions. Moreover, we show that the bargaining solutions found by our algorithm also satisfy the axioms of symmetry and (under slightly stronger conditions) independence of irrelevant alternatives, which are popular in the literature. Finally, we conduct experiments in two domains, multi agent formation assignment and mediated stock portfolio allocation, which validate these theoretic results. All code for our experiments can be found at https://github.com/suryakmurthy/dibs_bargaining.

  PublisherOA PDFDOI

• Game-theoretic Decentralized Coordination for Airspace Sector Overload Mitigation

  ArXiv.org · 2025-11-14

  preprintOpen access

  Decentralized air traffic management systems offer a scalable alternative to centralized control, but often assume high levels of cooperation. In practice, such assumptions frequently break down since airspace sectors operate independently and prioritize local objectives. We address the problem of sector overload in decentralized air traffic management by proposing a mechanism that models self-interested behaviors based on best response dynamics. Each sector adjusts the departure times of flights under its control to reduce its own congestion, without any shared decision making. A tunable cooperativeness factor models the degree to which each sector is willing to reduce overload in other sectors. We prove that the proposed mechanism satisfies a potential game structure, ensuring that best response dynamics converge to a pure Nash equilibrium, under a mild restriction. In addition, we identify a sufficient condition under which an overload-free solution corresponds to a global minimizer of the potential function. Numerical experiments using 24 hours of European flight data demonstrate that the proposed algorithm substantially reduces overload even with only minimal cooperation between sectors, while maintaining scalability and matching the solution quality of centralized solvers.

  PublisherOA PDFDOI

• SCOPED: Score-Curvature Out-of-distribution Proximity Evaluator for Diffusion

  ArXiv.org · 2025-10-01

  preprintOpen accessSenior author

  Out-of-distribution (OOD) detection is essential for reliable deployment of machine learning systems in vision, robotics, reinforcement learning, and beyond. We introduce Score-Curvature Out-of-distribution Proximity Evaluator for Diffusion (SCOPED), a fast and general-purpose OOD detection method for diffusion models that reduces the number of forward passes on the trained model by an order of magnitude compared to prior methods, outperforming most diffusion-based baselines and closely approaching the accuracy of the strongest ones. SCOPED is computed from a single diffusion model trained once on a diverse dataset, and combines the Jacobian trace and squared norm of the model's score function into a single test statistic. Rather than thresholding on a fixed value, we estimate the in-distribution density of SCOPED scores using kernel density estimation, enabling a flexible, unsupervised test that, in the simplest case, only requires a single forward pass and one Jacobian-vector product (JVP), made efficient by Hutchinson's trace estimator. On four vision benchmarks, SCOPED achieves competitive or state-of-the-art precision-recall scores despite its low computational cost. The same method generalizes to robotic control tasks with shared state and action spaces, identifying distribution shifts across reward functions and training regimes. These results position SCOPED as a practical foundation for fast and reliable OOD detection in real-world domains, including perceptual artifacts in vision, outlier detection in autoregressive models, exploration in reinforcement learning, and dataset curation for unsupervised training.

  PublisherOA PDFDOI

Recent grants

• CAREER: Game Theoretic Models for Robust Cyber-Physical Interactions: Inference and Design under Uncertainty

  NSF · $616k · 2024–2028

• Collaborative Research: Interaction-aware Planning and Control for Robotic Navigation in the Crowd

  NSF · $447k · 2022–2026

Frequent coauthors

• Claire J. Tomlin

  69 shared

• Ufuk Topcu

  The University of Texas at Austin

  26 shared

• Lasse Peters

  Delft University of Technology

  25 shared

• Jaime F. Fisac

  22 shared

• Chih-Yuan Chiu

  15 shared

• Sylvia Herbert

  University of California, San Diego

  14 shared

• Anca D. Dragan

  University of California, Berkeley

  13 shared

• Yue Yu

  11 shared

Labs

• Fridovich-Keil LabPI

  1-2 sentence research focus

Education

• Ph.D.

  University of California, Berkeley

Awards & honors

• National Science Foundation CAREER Award (2024)