Addressing such diverse ends as safety alignment with human preferences, and the efficiency of learning, a growing line of reinforcement learning research focuses on risk functionals that depend on the entire distribution of returns. Recent work on \emph{off-policy risk assessment} (OPRA) for contextual bandits introduced consistent estimators for the target policy's CDF of returns along with finite sample guarantees that extend to (and hold simultaneously over) all risk. In this paper, we lift OPRA to Markov decision processes (MDPs), where importance sampling (IS) CDF estimators suffer high variance on longer trajectories due to small effective sample size. To mitigate these problems, we incorporate model-based estimation to develop the first doubly robust (DR) estimator for the CDF of returns in MDPs. This estimator enjoys significantly less variance and, when the model is well specified, achieves the Cramer-Rao variance lower bound. Moreover, for many risk functionals, the downstream estimates enjoy both lower bias and lower variance. Additionally, we derive the first minimax lower bounds for off-policy CDF and risk estimation, which match our error bounds up to a constant factor. Finally, we demonstrate the precision of our DR CDF estimates experimentally on several different environments.

Multi-agent reinforcement learning (MARL) has emerged as a useful approach to solving decentralised decision-making problems at scale. Research in the field has been growing steadily with many breakthrough algorithms proposed in recent years. In this work, we take a closer look at this rapid development with a focus on evaluation methodologies employed across a large body of research in cooperative MARL. By conducting a detailed meta-analysis of prior work, spanning 75 papers accepted for publication from 2016 to 2022, we bring to light worrying trends that put into question the true rate of progress. We further consider these trends in a wider context and take inspiration from single-agent RL literature on similar issues with recommendations that remain applicable to MARL. Combining these recommendations, with novel insights from our analysis, we propose a standardised performance evaluation protocol for cooperative MARL. We argue that such a standard protocol, if widely adopted, would greatly improve the validity and credibility of future research, make replication and reproducibility easier, as well as improve the ability of the field to accurately gauge the rate of progress over time by being able to make sound comparisons across different works.

Robotic cloth manipulation is a relevant challenging problem for autonomous robotic systems. Highly deformable objects as textile items can adopt multiple configurations and shapes during their manipulation. Hence, robots should not only understand the current cloth configuration but also be able to predict the future possible behaviors of the cloth. This paper addresses the problem of indirectly controlling the configuration of certain points of a textile object, by applying actions on other parts of the object through the use of a Model Predictive Control (MPC) strategy, which also allows to foresee the behavior of indirectly controlled points. The designed controller finds the optimal control signals to attain the desired future target configuration. The explored scenario in this paper considers tracking a reference trajectory with the lower corners of a square piece of cloth by grasping its upper corners. To do so, we propose and validate a linear cloth model that allows solving the MPC-related optimization problem in real time. Reinforcement Learning (RL) techniques are used to learn the optimal parameters of the proposed cloth model and also to tune the resulting MPC. After obtaining accurate tracking results in simulation, the full control scheme was implemented and executed in a real robot, obtaining accurate tracking even in adverse conditions. While total observed errors reach the 5 cm mark, for a 30x30 cm cloth, an analysis shows the MPC contributes less than 30% to that value.

Combining learned policies in a prioritized, ordered manner is desirable because it allows for modular design and facilitates data reuse through knowledge transfer. In control theory, prioritized composition is realized by null-space control, where low-priority control actions are projected into the null-space of high-priority control actions. Such a method is currently unavailable for Reinforcement Learning. We propose a novel, task-prioritized composition framework for Reinforcement Learning, which involves a novel concept: The indifferent-space of Reinforcement Learning policies. Our framework has the potential to facilitate knowledge transfer and modular design while greatly increasing data efficiency and data reuse for Reinforcement Learning agents. Further, our approach can ensure high-priority constraint satisfaction, which makes it promising for learning in safety-critical domains like robotics. Unlike null-space control, our approach allows learning globally optimal policies for the compound task by online learning in the indifference-space of higher-level policies after initial compound policy construction.

As a relatively new form of sport, esports offers unparalleled data availability. Despite the vast amounts of data that are generated by game engines, it can be challenging to extract them and verify their integrity for the purposes of practical and scientific use. Our work aims to open esports to a broader scientific community by supplying raw and pre-processed files from StarCraft II esports tournaments. These files can be used in statistical and machine learning modeling tasks and related to various laboratory-based measurements (e.g., behavioral tests, brain imaging). We have gathered publicly available game-engine generated "replays" of tournament matches and performed data extraction and cleanup using a low-level application programming interface (API) parser library. Additionally, we open-sourced and published all the custom tools that were developed in the process of creating our dataset. These tools include PyTorch and PyTorch Lightning API abstractions to load and model the data. Our dataset contains replays from major and premiere StarCraft II tournaments since 2016. To prepare the dataset, we processed 55 tournament "replaypacks" that contained 17930 files with game-state information. Based on initial investigation of available StarCraft II datasets, we observed that our dataset is the largest publicly available source of StarCraft II esports data upon its publication. Analysis of the extracted data holds promise for further Artificial Intelligence (AI), Machine Learning (ML), psychological, Human-Computer Interaction (HCI), and sports-related studies in a variety of supervised and self-supervised tasks.

Mean-Field Game (MFG) serves as a crucial mathematical framework in modeling the collective behavior of individual agents interacting stochastically with a large population. In this work, we aim at solving a challenging class of MFGs in which the differentiability of these interacting preferences may not be available to the solver, and the population is urged to converge exactly to some desired distribution. These setups are, despite being well-motivated for practical purposes, complicated enough to paralyze most (deep) numerical solvers. Nevertheless, we show that Schr\"odinger Bridge - as an entropy-regularized optimal transport model - can be generalized to accepting mean-field structures, hence solving these MFGs. This is achieved via the application of Forward-Backward Stochastic Differential Equations theory, which, intriguingly, leads to a computational framework with a similar structure to Temporal Difference learning. As such, it opens up novel algorithmic connections to Deep Reinforcement Learning that we leverage to facilitate practical training. We show that our proposed objective function provides necessary and sufficient conditions to the mean-field problem. Our method, named Deep Generalized Schr\"odinger Bridge (DeepGSB), not only outperforms prior methods in solving classical population navigation MFGs, but is also capable of solving 1000-dimensional opinion depolarization, setting a new state-of-the-art numerical solver for high-dimensional MFGs. Our code will be made available at https://github.com/ghliu/DeepGSB.

We present MIDGARD, an open-source simulation platform for autonomous robot navigation in outdoor unstructured environments. MIDGARD is designed to enable the training of autonomous agents (e.g., unmanned ground vehicles) in photorealistic 3D environments, and to support the generalization skills of learning-based agents through the variability in training scenarios. MIDGARD's main features include a configurable, extensible, and difficulty-driven procedural landscape generation pipeline, with fast and photorealistic scene rendering based on Unreal Engine. Additionally, MIDGARD has built-in support for OpenAI Gym, a programming interface for feature extension (e.g., integrating new types of sensors, customizing exposing internal simulation variables), and a variety of simulated agent sensors (e.g., RGB, depth and instance/semantic segmentation). We evaluate MIDGARD's capabilities as a benchmarking tool for robot navigation utilizing a set of state-of-the-art reinforcement learning algorithms. The results demonstrate MIDGARD's suitability as a simulation and training environment, as well as the effectiveness of our procedural generation approach in controlling scene difficulty, which directly reflects on accuracy metrics. MIDGARD build, source code and documentation are available at https://midgardsim.org/.

We present a deep reinforcement learning-based framework for automatically discovering patterns available in any given initial configuration of fat robot swarms. In particular, we model the problem of collision-less gathering and mutual visibility in fat robot swarms and discover patterns for solving them using our framework. We show that by shaping reward signals based on certain constraints like mutual visibility and safe proximity, the robots can discover collision-less trajectories leading to well-formed gathering and visibility patterns.

This paper addresses the problem of optimizing charging/discharging schedules of electric vehicles (EVs) when participate in demand response (DR). As there exist uncertainties in EVs' remaining energy, arrival and departure time, and future electricity prices, it is quite difficult to make charging decisions to minimize charging cost while guarantee that the EV's battery state-of-the-charge (SOC) is within certain range. To handle with this dilemma, this paper formulates the EV charging scheduling problem as a constrained Markov decision process (CMDP). By synergistically combining the augmented Lagrangian method and soft actor critic algorithm, a novel safe off-policy reinforcement learning (RL) approach is proposed in this paper to solve the CMDP. The actor network is updated in a policy gradient manner with the Lagrangian value function. A double-critics network is adopted to synchronously estimate the action-value function to avoid overestimation bias. The proposed algorithm does not require strong convexity guarantee of examined problems and is sample efficient. Comprehensive numerical experiments with real-world electricity price demonstrate that our proposed algorithm can achieve high solution optimality and constraints compliance.

The Euclidean Steiner tree problem seeks the min-cost network to connect a collection of target locations, and it underlies many applications of wireless networks. In this paper, we present a study on solving the Euclidean Steiner tree problem using reinforcement learning enhanced by graph representation learning. Different from the commonly studied connectivity problems like travelling salesman problem or vehicle routing problem where the search space is finite, the Euclidean Steiner tree problem requires to search over the entire Euclidean space, thereby making the existing methods not applicable. In this paper, we design discretization methods by leveraging the unique characteristics of the Steiner tree, and propose new training schemes for handling the dynamic Steiner points emerging during the incremental construction. Our design is examined through a sanity check using experiments on a collection of datasets, with encouraging results demonstrating the utility of our method as an alternative to classic combinatorial methods.