Progress in deep learning highlights the tremendous potential of utilizing diverse robotic datasets for attaining effective generalization and makes it enticing to consider leveraging broad datasets for attaining robust generalization in robotic learning as well. However, in practice, we often want to learn a new skill in a new environment that is unlikely to be contained in the prior data. Therefore we ask: how can we leverage existing diverse offline datasets in combination with small amounts of task-specific data to solve new tasks, while still enjoying the generalization benefits of training on large amounts of data? In this paper, we demonstrate that end-to-end offline RL can be an effective approach for doing this, without the need for any representation learning or vision-based pre-training. We present pre-training for robots (PTR), a framework based on offline RL that attempts to effectively learn new tasks by combining pre-training on existing robotic datasets with rapid fine-tuning on a new task, with as few as 10 demonstrations. PTR utilizes an existing offline RL method, conservative Q-learning (CQL), but extends it to include several crucial design decisions that enable PTR to actually work and outperform a variety of prior methods. To our knowledge, PTR is the first RL method that succeeds at learning new tasks in a new domain on a real WidowX robot with as few as 10 task demonstrations, by effectively leveraging an existing dataset of diverse multi-task robot data collected in a variety of toy kitchens. We also demonstrate that PTR can enable effective autonomous fine-tuning and improvement in a handful of trials, without needing any demonstrations. An accompanying overview video can be found in the supplementary material and at thi URL: https://sites.google.com/view/ptr-final/

Deep reinforcement learning (DRL) for resource allocation has been investigated extensively owing to its ability of handling model-free and end-to-end problems. Yet the high training complexity of DRL hinders its practical use in dynamic wireless systems. To reduce the training cost, we resort to graph reinforcement learning for exploiting two kinds of relational priors inherent in many problems in wireless communications: topology information and permutation properties. To design graph reinforcement learning framework systematically for harnessing the two priors, we first conceive a method to transform state matrix into state graph, and then propose a general method for graph neural networks to satisfy desirable permutation properties. To demonstrate how to apply the proposed methods, we take deep deterministic policy gradient (DDPG) as an example for optimizing two representative resource allocation problems. One is predictive power allocation that minimizes the energy consumed for ensuring the quality-ofservice of each user that requests video streaming. The other is link scheduling that maximizes the sum-rate for device-to-device communications. Simulation results show that the graph DDPG algorithm converges much faster and needs much lower space complexity than existing DDPG algorithms to achieve the same learning performance. Deep reinforcement learning (DRL) has been introduced to optimize a variety of resource allocation problems, thanks to its ability of learning wireless policies from the optimization problems without closed-form objectives and constraints, making decision in an end-to-end manner, and online training [1-8]. When learning a resource allocation policy to be operated in non-stationary wireless channels, a DRL algorithm needs to be online trained consistently for adapting to the dynamic environments. In particular, the agent of DRL interacts with the environment to gather a sample (i.e., an experience in reinforcement learning parlance) in each time step and updates deep neural networks (DNNs) with a batch of experiences every several time steps.

Multi-Agent Reinforcement Learning (MARL) is a promising candidate for realizing efficient control of microscopic particles, of which micro-robots are a subset. However, the microscopic particles' environment presents unique challenges, such as Brownian motion at sufficiently small length-scales. In this work, we explore the role of temperature in the emergence and efficacy of strategies in MARL systems using particle-based Langevin molecular dynamics simulations as a realistic representation of micro-scale environments. To this end, we perform experiments on two different multi-agent tasks in microscopic environments at different temperatures, detecting the source of a concentration gradient and rotation of a rod. We find that at higher temperatures, the RL agents identify new strategies for achieving these tasks, highlighting the importance of understanding this regime and providing insight into optimal training strategies for bridging the generalization gap between simulation and reality. We also introduce a novel Python package for studying microscopic agents using reinforcement learning (RL) to accompany our results.

Pre-training on Internet data has proven to be a key ingredient for broad generalization in many modern ML systems. What would it take to enable such capabilities in robotic reinforcement learning (RL)? Offline RL methods, which learn from datasets of robot experience, offer one way to leverage prior data into the robotic learning pipeline. However, these methods have a "type mismatch" with video data (such as Ego4D), the largest prior datasets available for robotics, since video offers observation-only experience without the action or reward annotations needed for RL methods. In this paper, we develop a system for leveraging large-scale human video datasets in robotic offline RL, based entirely on learning value functions via temporal-difference learning. We show that value learning on video datasets learns representations that are more conducive to downstream robotic offline RL than other approaches for learning from video data. Our system, called V-PTR, combines the benefits of pre-training on video data with robotic offline RL approaches that train on diverse robot data, resulting in value functions and policies for manipulation tasks that perform better, act robustly, and generalize broadly. On several manipulation tasks on a real WidowX robot, our framework produces policies that greatly improve over prior methods. Our video and additional details can be found at https://dibyaghosh.com/vptr/

Humans learn by interacting with their environments and perceiving the outcomes of their actions. A landmark in artificial intelligence has been the development of deep reinforcement learning (dRL) algorithms capable of doing the same in video games, on par with or better than humans. However, it remains unclear whether the successes of dRL models reflect advances in visual representation learning, the effectiveness of reinforcement learning algorithms at discovering better policies, or both. To address this question, we introduce the Learning Challenge Diagnosticator (LCD), a tool that separately measures the perceptual and reinforcement learning demands of a task. We use LCD to discover a novel taxonomy of challenges in the Procgen benchmark, and demonstrate that these predictions are both highly reliable and can instruct algorithmic development. More broadly, the LCD reveals multiple failure cases that can occur when optimizing dRL algorithms over entire video game benchmarks like Procgen, and provides a pathway towards more efficient progress.

Even though Google Research Football (GRF) was initially benchmarked and studied as a single-agent environment in its original paper, recent years have witnessed an increasing focus on its multi-agent nature by researchers utilizing it as a testbed for Multi-Agent Reinforcement Learning (MARL). However, the absence of standardized environment settings and unified evaluation metrics for multi-agent scenarios hampers the consistent understanding of various studies. Furthermore, the challenging 5-vs-5 and 11-vs-11 full-game scenarios have received limited thorough examination due to their substantial training complexities. To address these gaps, this paper extends the original environment by not only standardizing the environment settings and benchmarking cooperative learning algorithms across different scenarios, including the most challenging full-game scenarios, but also by discussing approaches to enhance football AI from diverse perspectives and introducing related research tools. Specifically, we provide a distributed and asynchronous population-based self-play framework with diverse pre-trained policies for faster training, two football-specific analytical tools for deeper investigation, and an online leaderboard for broader evaluation. The overall expectation of this work is to advance the study of Multi-Agent Reinforcement Learning on Google Research Football environment, with the ultimate goal of benefiting real-world sports beyond virtual games.

Crowd simulation is important for video-games design, since it enables to populate virtual worlds with autonomous avatars that navigate in a human-like manner. Reinforcement learning has shown great potential in simulating virtual crowds, but the design of the reward function is critical to achieving effective and efficient results. In this work, we explore the design of reward functions for reinforcement learning-based crowd simulation. We provide theoretical insights on the validity of certain reward functions according to their analytical properties, and evaluate them empirically using a range of scenarios, using the energy efficiency as the metric. Our experiments show that directly minimizing the energy usage is a viable strategy as long as it is paired with an appropriately scaled guiding potential, and enable us to study the impact of the different reward components on the behavior of the simulated crowd. Our findings can inform the development of new crowd simulation techniques, and contribute to the wider study of human-like navigation.

In this work, we introduce OmniDrones, an efficient and flexible platform tailored for reinforcement learning in drone control, built on Nvidia's Omniverse Isaac Sim. It employs a bottom-up design approach that allows users to easily design and experiment with various application scenarios on top of GPU-parallelized simulations. It also offers a range of benchmark tasks, presenting challenges ranging from single-drone hovering to over-actuated system tracking. In summary, we propose an open-sourced drone simulation platform, equipped with an extensive suite of tools for drone learning. It includes 4 drone models, 5 sensor modalities, 4 control modes, over 10 benchmark tasks, and a selection of widely used RL baselines. To showcase the capabilities of OmniDrones and to support future research, we also provide preliminary results on these benchmark tasks. We hope this platform will encourage further studies on applying RL to practical drone systems.

We address the challenge of enhancing navigation autonomy for planetary space rovers using reinforcement learning (RL). The ambition of future space missions necessitates advanced autonomous navigation capabilities for rovers to meet mission objectives. RL's potential in robotic autonomy is evident, but its reliance on simulations poses a challenge. Transferring policies to real-world scenarios often encounters the "reality gap", disrupting the transition from virtual to physical environments. The reality gap is exacerbated in the context of mapless navigation on Mars and Moon-like terrains, where unpredictable terrains and environmental factors play a significant role. Effective navigation requires a method attuned to these complexities and real-world data noise. We introduce a novel two-stage RL approach using offline noisy data. Our approach employs a teacher-student policy learning paradigm, inspired by the "learning by cheating" method. The teacher policy is trained in simulation. Subsequently, the student policy is trained on noisy data, aiming to mimic the teacher's behaviors while being more robust to real-world uncertainties. Our policies are transferred to a custom-designed rover for real-world testing. Comparative analyses between the teacher and student policies reveal that our approach offers improved behavioral performance, heightened noise resilience, and more effective sim-to-real transfer.

Robot multimodal locomotion encompasses the ability to transition between walking and flying, representing a significant challenge in robotics. This work presents an approach that enables automatic smooth transitions between legged and aerial locomotion. Leveraging the concept of Adversarial Motion Priors, our method allows the robot to imitate motion datasets and accomplish the desired task without the need for complex reward functions. The robot learns walking patterns from human-like gaits and aerial locomotion patterns from motions obtained using trajectory optimization. Through this process, the robot adapts the locomotion scheme based on environmental feedback using reinforcement learning, with the spontaneous emergence of mode-switching behavior. The results highlight the potential for achieving multimodal locomotion in aerial humanoid robotics through automatic control of walking and flying modes, paving the way for applications in diverse domains such as search and rescue, surveillance, and exploration missions. This research contributes to advancing the capabilities of aerial humanoid robots in terms of versatile locomotion in various environments.