About

Roy Fox is an Assistant Professor in the Department of Computer Science at the University of California, Irvine, within the Donald Bren School of Information & Computer Sciences. He is the founder of the Intelligent Dynamics Lab (indylab) and is affiliated with the Center for Machine Learning and Intelligent Systems (CML), the Algorithms, Combinatorics and Optimization Center (ACO), and the Institute for Mathematical Behavioral Sciences (IMBS). Fox's academic background includes a Ph.D. in Computer Science and Engineering from the Hebrew University of Jerusalem, where he worked under Naftali Tishby. He also holds an MSc from the Technion IIT, where he studied under Moshe Tennenholtz, and was a Ph.D. exchange student with Liam Paninski at Columbia University. His previous postdoctoral work was conducted at Berkeley AI Research (BAIR), working with Ion Stoica in the RISELab and with Ken Goldberg in the AUTOLAB at UC Berkeley. His research interests encompass the theory and applications of control learning, reinforcement learning, control theory, information theory, and robotics. His current research focuses on structure, exploration, and optimization in deep control learning of virtual and physical agents.

Research topics

• Natural Language Processing
• Computer Science
• Artificial Intelligence
• Engineering
• Human–computer interaction
• Programming language

Selected publications

• Data-Augmented Game Starts for Accelerating Self-Play Exploration in Imperfect Information Games

  arXiv (Cornell University) · 2026-05-14

  preprintOpen accessSenior author

  Finding approximate equilibria for large-scale imperfect-information competitive games such as StarCraft, Dota, and CounterStrike remains computationally infeasible due to sparse rewards and challenging exploration over long horizons. In this paper, we propose a multi-agent starting-state sampling strategy designed to substantially accelerate online exploration in regularized policy-gradient game methods for two-player zero-sum (2p0s) games. Motivated by an assumption that offline demonstrations from skilled humans can provide good coverage of high-level strategies relevant to equilibrium play, we propose the initialization of reinforcement learning data collection at intermediate states sampled from offline data to facilitate exploration of strategically relevant subgames. Referring to this method as Data-Augmented Game Starts (DAGS), we perform experiments using synthetic datasets and analytically tractable, long-horizon control variants of two-player Kuhn Poker, Goofspiel, and a counterexample game designed to penalize biased beliefs over hidden information. Under fixed computational budgets, DAGS enables regularized policy gradient methods to achieve lower exploitability in games with significantly more challenging exploration. We show that augmenting starting state distributions when solving imperfect information games can lead to biased equilibria, and we provide a straightforward mitigation to this in the form of multi-task observation flags. Finally, we release a new set of benchmark environments that drastically increase exploration challenges and state counts in existing OpenSpiel games while keeping exploitability measurements analytically tractable.

  PublisherDOI

• Data-Augmented Game Starts for Accelerating Self-Play Exploration in Imperfect Information Games

  ArXiv.org · 2026-05-14

  articleOpen accessSenior author

  Finding approximate equilibria for large-scale imperfect-information competitive games such as StarCraft, Dota, and CounterStrike remains computationally infeasible due to sparse rewards and challenging exploration over long horizons. In this paper, we propose a multi-agent starting-state sampling strategy designed to substantially accelerate online exploration in regularized policy-gradient game methods for two-player zero-sum (2p0s) games. Motivated by an assumption that offline demonstrations from skilled humans can provide good coverage of high-level strategies relevant to equilibrium play, we propose the initialization of reinforcement learning data collection at intermediate states sampled from offline data to facilitate exploration of strategically relevant subgames. Referring to this method as Data-Augmented Game Starts (DAGS), we perform experiments using synthetic datasets and analytically tractable, long-horizon control variants of two-player Kuhn Poker, Goofspiel, and a counterexample game designed to penalize biased beliefs over hidden information. Under fixed computational budgets, DAGS enables regularized policy gradient methods to achieve lower exploitability in games with significantly more challenging exploration. We show that augmenting starting state distributions when solving imperfect information games can lead to biased equilibria, and we provide a straightforward mitigation to this in the form of multi-task observation flags. Finally, we release a new set of benchmark environments that drastically increase exploration challenges and state counts in existing OpenSpiel games while keeping exploitability measurements analytically tractable.

  PublisherOA PDF

• Reinforcement Learning from Delayed Observations via World Models

  arXiv (Cornell University) · 2024-03-18 · 3 citations

  preprintOpen accessSenior author

  In standard reinforcement learning settings, agents typically assume immediate feedback about the effects of their actions after taking them. However, in practice, this assumption may not hold true due to physical constraints and can significantly impact the performance of learning algorithms. In this paper, we address observation delays in partially observable environments. We propose leveraging world models, which have shown success in integrating past observations and learning dynamics, to handle observation delays. By reducing delayed POMDPs to delayed MDPs with world models, our methods can effectively handle partial observability, where existing approaches achieve sub-optimal performance or degrade quickly as observability decreases. Experiments suggest that one of our methods can outperform a naive model-based approach by up to 250%. Moreover, we evaluate our methods on visual delayed environments, for the first time showcasing delay-aware reinforcement learning continuous control with visual observations.

  PublisherOA PDFDOI

• Verification-Guided Shielding for Deep Reinforcement Learning

  arXiv (Cornell University) · 2024-06-10 · 1 citations

  preprintOpen accessSenior author

  In recent years, Deep Reinforcement Learning (DRL) has emerged as an effective approach to solving real-world tasks. However, despite their successes, DRL-based policies suffer from poor reliability, which limits their deployment in safety-critical domains. Various methods have been put forth to address this issue by providing formal safety guarantees. Two main approaches include shielding and verification. While shielding ensures the safe behavior of the policy by employing an external online component (i.e., a ``shield'') that overrides potentially dangerous actions, this approach has a significant computational cost as the shield must be invoked at runtime to validate every decision. On the other hand, verification is an offline process that can identify policies that are unsafe, prior to their deployment, yet, without providing alternative actions when such a policy is deemed unsafe. In this work, we present verification-guided shielding -- a novel approach that bridges the DRL reliability gap by integrating these two methods. Our approach combines both formal and probabilistic verification tools to partition the input domain into safe and unsafe regions. In addition, we employ clustering and symbolic representation procedures that compress the unsafe regions into a compact representation. This, in turn, allows to temporarily activate the shield solely in (potentially) unsafe regions, in an efficient manner. Our novel approach allows to significantly reduce runtime overhead while still preserving formal safety guarantees. We extensively evaluate our approach on two benchmarks from the robotic navigation domain, as well as provide an in-depth analysis of its scalability and completeness.

  PublisherOA PDFDOI

• Moonwalk: Inverse-Forward Differentiation

  arXiv (Cornell University) · 2024-02-22

  preprintOpen accessSenior author

  Backpropagation's main limitation is its need to store intermediate activations (residuals) during the forward pass, which restricts the depth of trainable networks. This raises a fundamental question: can we avoid storing these activations? We address this by revisiting the structure of gradient computation. Backpropagation computes gradients through a sequence of vector-Jacobian products, an operation that is generally irreversible. The lost information lies in the cokernel of each layer's Jacobian. We define submersive networks -- networks whose layer Jacobians have trivial cokernels -- in which gradients can be reconstructed exactly in a forward sweep without storing activations. For non-submersive layers, we introduce fragmental gradient checkpointing, which records only the minimal subset of residuals necessary to restore the cotangents erased by the Jacobian. Central to our approach is a novel operator, the vector-inverse-Jacobian product (vijp), which inverts gradient flow outside the cokernel. Our mixed-mode algorithm first computes input gradients with a memory-efficient reverse pass, then reconstructs parameter gradients in a forward sweep using the vijp, eliminating the need to store activations. We implement this method in Moonwalk and show that it matches backpropagation's runtime while training networks more than twice as deep under the same memory budget.

  PublisherOA PDFDOI

• Realizable Continuous-Space Shields for Safe Reinforcement Learning

  arXiv (Cornell University) · 2024-10-02

  preprintOpen accessSenior author

  While Deep Reinforcement Learning (DRL) has achieved remarkable success across various domains, it remains vulnerable to occasional catastrophic failures without additional safeguards. An effective solution to prevent these failures is to use a shield that validates and adjusts the agent's actions to ensure compliance with a provided set of safety specifications. For real-world robotic domains, it is essential to define safety specifications over continuous state and action spaces to accurately account for system dynamics and compute new actions that minimally deviate from the agent's original decision. In this paper, we present the first shielding approach specifically designed to ensure the satisfaction of safety requirements in continuous state and action spaces, making it suitable for practical robotic applications. Our method builds upon realizability, an essential property that confirms the shield will always be able to generate a safe action for any state in the environment. We formally prove that realizability can be verified for stateful shields, enabling the incorporation of non-Markovian safety requirements, such as loop avoidance. Finally, we demonstrate the effectiveness of our approach in ensuring safety without compromising the policy's success rate by applying it to a navigation problem and a multi-agent particle environment.

  PublisherOA PDFDOI

• Skill Set Optimization: Reinforcing Language Model Behavior via Transferable Skills

  arXiv (Cornell University) · 2024-02-05 · 1 citations

  preprintOpen accessSenior author

  Large language models (LLMs) have recently been used for sequential decision making in interactive environments. However, leveraging environment reward signals for continual LLM actor improvement is not straightforward. We propose Skill Set Optimization (SSO) for improving LLM actor performance through constructing and refining sets of transferable skills. SSO constructs skills by extracting common subtrajectories with high rewards and generating subgoals and instructions to represent each skill. These skills are provided to the LLM actor in-context to reinforce behaviors with high rewards. Then, SSO further refines the skill set by pruning skills that do not continue to result in high rewards. We evaluate our method in the classic videogame NetHack and the text environment ScienceWorld to demonstrate SSO's ability to optimize a set of skills and perform in-context policy improvement. SSO outperforms baselines by 40% in our custom NetHack task and outperforms the previous state-of-the-art in ScienceWorld by 35%.

  PublisherOA PDFDOI

• Learning to Design Analog Circuits to Meet Threshold Specifications

  arXiv (Cornell University) · 2023-07-25

  preprintOpen accessSenior author

  Automated design of analog and radio-frequency circuits using supervised or reinforcement learning from simulation data has recently been studied as an alternative to manual expert design. It is straightforward for a design agent to learn an inverse function from desired performance metrics to circuit parameters. However, it is more common for a user to have threshold performance criteria rather than an exact target vector of feasible performance measures. In this work, we propose a method for generating from simulation data a dataset on which a system can be trained via supervised learning to design circuits to meet threshold specifications. We moreover perform the to-date most extensive evaluation of automated analog circuit design, including experimenting in a significantly more diverse set of circuits than in prior work, covering linear, nonlinear, and autonomous circuit configurations, and show that our method consistently reaches success rate better than 90% at 5% error margin, while also improving data efficiency by upward of an order of magnitude. A demo of this system is available at circuits.streamlit.app

  PublisherOA PDFDOI

• Count-Based Temperature Scheduling for Maximum Entropy Reinforcement Learning

  arXiv (Cornell University) · 2021-11-28

  preprintOpen accessSenior author

  Maximum Entropy Reinforcement Learning (MaxEnt RL) algorithms such as Soft Q-Learning (SQL) and Soft Actor-Critic trade off reward and policy entropy, which has the potential to improve training stability and robustness. Most MaxEnt RL methods, however, use a constant tradeoff coefficient (temperature), contrary to the intuition that the temperature should be high early in training to avoid overfitting to noisy value estimates and decrease later in training as we increasingly trust high value estimates to truly lead to good rewards. Moreover, our confidence in value estimates is state-dependent, increasing every time we use more evidence to update an estimate. In this paper, we present a simple state-based temperature scheduling approach, and instantiate it for SQL as Count-Based Soft Q-Learning (CBSQL). We evaluate our approach on a toy domain as well as in several Atari 2600 domains and show promising results.

  PublisherOA PDFDOI

• Modular Framework for Visuomotor Language Grounding

  arXiv (Cornell University) · 2021 · 5 citations

  • Computer Science
  • Computer Science
  • Artificial Intelligence

  Natural language instruction following tasks serve as a valuable test-bed for grounded language and robotics research. However, data collection for these tasks is expensive and end-to-end approaches suffer from data inefficiency. We propose the structuring of language, acting, and visual tasks into separate modules that can be trained independently. Using a Language, Action, and Vision (LAV) framework removes the dependence of action and vision modules on instruction following datasets, making them more efficient to train. We also present a preliminary evaluation of LAV on the ALFRED task for visual and interactive instruction following.

  PublisherOA PDFDOI

Frequent coauthors

• Ken Goldberg

  University of California, Berkeley

  27 shared

• Ion Stoica

  17 shared

• Naftali Tishby

  Hebrew University of Jerusalem

  12 shared

• Michael Laskey

  11 shared

• Sanjay Krishnan

  University of Chicago

  11 shared

• Stephen McAleer

  9 shared

• Pierre Baldi

  8 shared

• Jonathan Lee

  University of Manitoba

  8 shared

Awards & honors

• Rina Dechter, with Others, Receives a $5M NSF Grant to Impro…