Peter Stone

Saman Khamesian, Sri Harini Balaji, Di Yang Shi et al.

Type 1 Diabetes (T1D) management requires continuous adjustment of insulin and lifestyle behaviors to maintain blood glucose within a safe target range. Although automated insulin delivery (AID) systems have improved glycemic outcomes, many patients still fail to achieve recommended clinical targets, warranting new approaches to improve glucose control in patients with T1D. While reinforcement learning (RL) has been utilized as a promising approach, current RL-based methods focus primarily on insulin-only treatment and do not provide behavioral recommendations for glucose control. To address this gap, we propose GUIDE, an RL-based decision-support framework designed to complement AID technologies by providing behavioral recommendations to prevent abnormal glucose events. GUIDE generates structured actions defined by intervention type, magnitude, and timing, including bolus insulin administration and carbohydrate intake events. GUIDE integrates a patient-specific glucose level predictor trained on real-world continuous glucose monitoring data and supports both offline and online RL algorithms within a unified environment. We evaluate both off-policy and on-policy methods across 25 individuals with T1D using standardized glycemic metrics. Among the evaluated approaches, the CQL-BC algorithm demonstrates the highest average time-in-range, reaching 85.49% while maintaining low hypoglycemia exposures. Behavioral similarity analysis further indicates that the learned CQL-BC policy preserves key structural characteristics of patient action patterns, achieving a mean cosine similarity of 0.87 $\pm$ 0.09 across subjects. These findings suggest that conservative offline RL with a structured behavioral action space can provide clinically meaningful and behaviorally plausible decision support for personalized diabetes management.

Linji Wang, Zifan Xu, Peter Stone et al.

The high cost of real-world data for robotics Reinforcement Learning (RL) leads to the wide usage of simulators. Despite extensive work on building better dynamics models for simulators to match with the real world, there is another, often-overlooked mismatch between simulations and the real world, namely the distribution of available training tasks. Such a mismatch is further exacerbated by existing curriculum learning techniques, which automatically vary the simulation task distribution without considering its relevance to the real world. Considering these challenges, we posit that curriculum learning for robotics RL needs to be grounded in real-world task distributions. To this end, we propose Grounded Curriculum Learning (GCL), which aligns the simulated task distribution in the curriculum with the real world, as well as explicitly considers what tasks have been given to the robot and how the robot has performed in the past. We validate GCL using the BARN dataset on complex navigation tasks, achieving a 6.8% and 6.5% higher success rate compared to a state-of-the-art CL method and a curriculum designed by human experts, respectively. These results show that GCL can enhance learning efficiency and navigation performance by grounding the simulation task distribution in the real world within an adaptive curriculum.

Bram Grooten, Patrick MacAlpine, Kaushik Subramanian et al.

Generalization to unseen environments is a significant challenge in the field of robotics and control. In this work, we focus on contextual reinforcement learning, where agents act within environments with varying contexts, such as self-driving cars or quadrupedal robots that need to operate in different terrains or weather conditions than they were trained for. We tackle the critical task of generalizing to out-of-distribution (OOD) settings, without access to explicit context information at test time. Recent work has addressed this problem by training a context encoder and a history adaptation module in separate stages. While promising, this two-phase approach is cumbersome to implement and train. We simplify the methodology and introduce SPARC: single-phase adaptation for robust control. We test SPARC on varying contexts within the high-fidelity racing simulator Gran Turismo 7 and wind-perturbed MuJoCo environments, and find that it achieves reliable and robust OOD generalization.

Cameron Angliss, Jiaxun Cui, Jiaheng Hu et al.

Developing AI agents that can robustly adapt to dramatically different strategic landscapes without retraining is a central challenge for multi-agent learning. Pokémon Video Game Championships (VGC) is a domain with an extraordinarily large space of possible team configurations of approximately $10^{139}$ - far larger than those of Dota or Starcraft. The highly discrete, combinatorial nature of team building in Pokémon VGC causes optimal strategies to shift dramatically depending on both the team being piloted and the opponent's team, making generalization uniquely challenging. To advance research on this problem, we introduce VGC-Bench: a benchmark that provides critical infrastructure, standardizes evaluation protocols, and supplies human-play datasets and a range of baselines - from large-language-model agents and behavior cloning to reinforcement learning and empirical game-theoretic methods such as self-play, fictitious play, and double oracle. In the restricted setting where an agent is trained and evaluated on a single-team configuration, our methods are able to win against a professional VGC competitor. We extensively evaluated all baseline methods over progressively larger team sets and find that even the best-performing algorithm in the single-team setting struggles at scaling up as team size grows. Thus, policy generalization across diverse team strategies remains an open challenge for the community. Our code is open sourced at https://github.com/cameronangliss/VGC-Bench.

Shivin Dass, Wensi Ai, Yuqian Jiang et al.

A critical bottleneck limiting imitation learning in robotics is the lack of data. This problem is more severe in mobile manipulation, where collecting demonstrations is harder than in stationary manipulation due to the lack of available and easy-to-use teleoperation interfaces. In this work, we demonstrate TeleMoMa, a general and modular interface for whole-body teleoperation of mobile manipulators. TeleMoMa unifies multiple human interfaces including RGB and depth cameras, virtual reality controllers, keyboard, joysticks, etc., and any combination thereof. In its more accessible version, TeleMoMa works using simply vision (e.g., an RGB-D camera), lowering the entry bar for humans to provide mobile manipulation demonstrations. We demonstrate the versatility of TeleMoMa by teleoperating several existing mobile manipulators - PAL Tiago++, Toyota HSR, and Fetch - in simulation and the real world. We demonstrate the quality of the demonstrations collected with TeleMoMa by training imitation learning policies for mobile manipulation tasks involving synchronized whole-body motion. Finally, we also show that TeleMoMa's teleoperation channel enables teleoperation on site, looking at the robot, or remote, sending commands and observations through a computer network, and perform user studies to evaluate how easy it is for novice users to learn to collect demonstrations with different combinations of human interfaces enabled by our system. We hope TeleMoMa becomes a helpful tool for the community enabling researchers to collect whole-body mobile manipulation demonstrations. For more information and video results, https://robin-lab.cs.utexas.edu/telemoma-web.

Harshit Sikchi, Siddhant Agarwal, Pranaya Jajoo et al.

The reward hypothesis states that all goals and purposes can be understood as the maximization of a received scalar reward signal. However, in practice, defining such a reward signal is notoriously difficult, as humans are often unable to predict the optimal behavior corresponding to a reward function. Natural language offers an intuitive alternative for instructing reinforcement learning (RL) agents, yet previous language-conditioned approaches either require costly supervision or test-time training given a language instruction. In this work, we present a new approach that uses a pretrained RL agent trained using only unlabeled, offline interactions--without task-specific supervision or labeled trajectories--to get zero-shot test-time policy inference from arbitrary natural language instructions. We introduce a framework comprising three steps: imagine, project, and imitate. First, the agent imagines a sequence of observations corresponding to the provided language description using video generative models. Next, these imagined observations are projected into the target environment domain. Finally, an agent pretrained in the target environment with unsupervised RL instantly imitates the projected observation sequence through a closed-form solution. To the best of our knowledge, our method, RLZero, is the first approach to show direct language-to-behavior generation abilities on a variety of tasks and environments without any in-domain supervision. We further show that components of RLZero can be used to generate policies zero-shot from cross-embodied videos, such as those available on YouTube, even for complex embodiments like humanoids.

Siddhant Agarwal, Harshit Sikchi, Peter Stone et al.

Having explored an environment, intelligent agents should be able to transfer their knowledge to most downstream tasks within that environment without additional interactions. Referred to as "zero-shot learning", this ability remains elusive for general-purpose reinforcement learning algorithms. While recent works have attempted to produce zero-shot RL agents, they make assumptions about the nature of the tasks or the structure of the MDP. We present Proto Successor Measure: the basis set for all possible behaviors of a Reinforcement Learning Agent in a dynamical system. We prove that any possible behavior (represented using visitation distributions) can be represented using an affine combination of these policy-independent basis functions. Given a reward function at test time, we simply need to find the right set of linear weights to combine these bases corresponding to the optimal policy. We derive a practical algorithm to learn these basis functions using reward-free interaction data from the environment and show that our approach can produce the optimal policy at test time for any given reward function without additional environmental interactions. Project page: https://agarwalsiddhant10.github.io/projects/psm.html.