Decentralized learning has shown great promise for cooperative multi-agent reinforcement learning (MARL). However, non-stationarity remains a significant challenge in fully decentralized learning. In the paper, we tackle the non-stationarity problem in the simplest and fundamental way and propose multi-agent alternate Q-learning (MA2QL), where agents take turns updating their Q-functions by Q-learning. MA2QL is a minimalist approach to fully decentralized cooperative MARL but is theoretically grounded. We prove that when each agent guarantees $\varepsilon$-convergence at each turn, their joint policy converges to a Nash equilibrium. In practice, MA2QL only requires minimal changes to independent Q-learning (IQL). We empirically evaluate MA2QL on a variety of cooperative multi-agent tasks. Results show MA2QL consistently outperforms IQL, which verifies the effectiveness of MA2QL, despite such minimal changes.

In this work, we attempt to bridge the two fields of finite-agent and infinite-agent games, by studying how the optimal policies of agents evolve with the number of agents (population size) in mean-field games, an agent-centric perspective in contrast to the existing works focusing typically on the convergence of the empirical distribution of the population. To this end, the premise is to obtain the optimal policies of a set of finite-agent games with different population sizes. However, either deriving the closed-form solution for each game is theoretically intractable, training a distinct policy for each game is computationally intensive, or directly applying the policy trained in a game to other games is sub-optimal. We address these challenges through the Population-size-Aware Policy Optimization (PAPO). Our contributions are three-fold. First, to efficiently generate efficient policies for games with different population sizes, we propose PAPO, which unifies two natural options (augmentation and hypernetwork) and achieves significantly better performance. PAPO consists of three components: i) the population-size encoding which transforms the original value of population size to an equivalent encoding to avoid training collapse, ii) a hypernetwork to generate a distinct policy for each game conditioned on the population size, and iii) the population size as an additional input to the generated policy. Next, we construct a multi-task-based training procedure to efficiently train the neural networks of PAPO by sampling data from multiple games with different population sizes. Finally, extensive experiments on multiple environments show the significant superiority of PAPO over baselines, and the analysis of the evolution of the generated policies further deepens our understanding of the two fields of finite-agent and infinite-agent games.

In recent years, Deep Reinforcement Learning emerged as a promising approach for autonomous navigation of ground vehicles and has been utilized in various areas of navigation such as cruise control, lane changing, or obstacle avoidance. However, most research works either focus on providing an end-to-end solution training the whole system using Deep Reinforcement Learning or focus on one specific aspect such as local motion planning. This however, comes along with a number of problems such as catastrophic forgetfulness, inefficient navigation behavior, and non-optimal synchronization between different entities of the navigation stack. In this paper, we propose a holistic Deep Reinforcement Learning training approach in which the training procedure is involving all entities of the navigation stack. This should enhance the synchronization between- and understanding of all entities of the navigation stack and as a result, improve navigational performance. We trained several agents with a number of different observation spaces to study the impact of different input on the navigation behavior of the agent. In profound evaluations against multiple learning-based and classic model-based navigation approaches, our proposed agent could outperform the baselines in terms of efficiency and safety attaining shorter path lengths, less roundabout paths, and less collisions.

Using large datasets in machine learning has led to outstanding results, in some cases outperforming humans in tasks that were believed impossible for machines. However, achieving human-level performance when dealing with physically interactive tasks, e.g., in contact-rich robotic manipulation, is still a big challenge. It is well known that regulating the Cartesian impedance for such operations is of utmost importance for their successful execution. Approaches like reinforcement Learning (RL) can be a promising paradigm for solving such problems. More precisely, approaches that use task-agnostic expert demonstrations to bootstrap learning when solving new tasks have a huge potential since they can exploit large datasets. However, existing data collection systems are expensive, complex, or do not allow for impedance regulation. This work represents a first step towards a data collection framework suitable for collecting large datasets of impedance-based expert demonstrations compatible with the RL problem formulation, where a novel action space is used. The framework is designed according to requirements acquired after an extensive analysis of available data collection frameworks for robotics manipulation. The result is a low-cost and open-access tele-impedance framework which makes human experts capable of demonstrating contact-rich tasks.

We introduce an exploration bonus for deep reinforcement learning methods calculated using self-organising feature maps. Our method uses adaptive resonance theory (ART) providing online, unsupervised clustering to quantify the novelty of a state. This heuristic is used to add an intrinsic reward to the extrinsic reward signal for then to optimize the agent to maximize the sum of these two rewards. We find that this method was able to play the game Ordeal at a human level after a comparable number of training epochs to ICM arXiv:1705.05464. Agents augmented with RND arXiv:1810.12894 were unable to achieve the same level of performance in our space of hyperparameters.

We propose DITTO, an offline imitation learning algorithm which uses world models and on-policy reinforcement learning to addresses the problem of covariate shift, without access to an oracle or any additional online interactions. We discuss how world models enable offline, on-policy imitation learning, and propose a simple intrinsic reward defined in the world model latent space that induces imitation learning by reinforcement learning. Theoretically, we show that our formulation induces a divergence bound between expert and learner, in turn bounding the difference in reward. We test our method on difficult Atari environments from pixels alone, and achieve state-of-the-art performance in the offline setting.

The predominant approach in reinforcement learning is to assign credit to actions based on the expected return. However, we show that the return may depend on the policy in a way which could lead to excessive variance in value estimation and slow down learning. Instead, we show that the advantage function can be interpreted as causal effects and shares similar properties with causal representations. Based on this insight, we propose Direct Advantage Estimation (DAE), a novel method that can model the advantage function and estimate it directly from on-policy data while simultaneously minimizing the variance of the return without requiring the (action-)value function. We also relate our method to Temporal Difference methods by showing how value functions can be seamlessly integrated into DAE. The proposed method is easy to implement and can be readily adapted by modern actor-critic methods. We evaluate DAE empirically on three discrete control domains and show that it can outperform generalized advantage estimation (GAE), a strong baseline for advantage estimation, on a majority of the environments when applied to policy optimization.

We aim to automate decades of research and experience in register allocation, leveraging machine learning. We tackle this problem by embedding a multi-agent reinforcement learning algorithm within LLVM, training it with the state of the art techniques. We formalize the constraints that precisely define the problem for a given instruction-set architecture, while ensuring that the generated code preserves semantic correctness. We also develop a gRPC based framework providing a modular and efficient compiler interface for training and inference. Our approach is architecture independent: we show experimental results targeting Intel x86 and ARM AArch64. Our results match or out-perform the heavily tuned, production-grade register allocators of LLVM.

We consider the problem of preference based reinforcement learning (PbRL), where, unlike traditional reinforcement learning, an agent receives feedback only in terms of a 1 bit (0/1) preference over a trajectory pair instead of absolute rewards for them. The success of the traditional RL framework crucially relies on the underlying agent-reward model, which, however, depends on how accurately a system designer can express an appropriate reward function and often a non-trivial task. The main novelty of our framework is the ability to learn from preference-based trajectory feedback that eliminates the need to hand-craft numeric reward models. This paper sets up a formal framework for the PbRL problem with non-markovian rewards, where the trajectory preferences are encoded by a generalized linear model of dimension $d$. Assuming the transition model is known, we then propose an algorithm with almost optimal regret guarantee of $\tilde {\mathcal{O}}\left( SH d \log (T / \delta) \sqrt{T} \right)$. We further, extend the above algorithm to the case of unknown transition dynamics, and provide an algorithm with near optimal regret guarantee $\widetilde{\mathcal{O}}((\sqrt{d} + H^2 + |\mathcal{S}|)\sqrt{dT} +\sqrt{|\mathcal{S}||\mathcal{A}|TH} )$. To the best of our knowledge, our work is one of the first to give tight regret guarantees for preference based RL problems with trajectory preferences.

Abstract--In recent years, mobile robot navigation approaches have become increasingly important due to various application areas ranging from healthcare to warehouse logistics. In particular, Deep Reinforcement Learning approaches have gained popularity for robot navigation but are not easily accessible to non-experts and complex to develop. In recent years, efforts have been made to make these sophisticated approaches accessible to a wider audience. The interface is designed to be intuitive and engaging to appeal to non-experts and make the technology accessible to a wider audience. With Arena-Web and its interface, training and developing Deep Reinforcement Learning agents is simplified and made easy without a single line of code. The web-app is free to use and openly available under the link stated in the supplementary materials. With recent advances in Deep Reinforcement Learning (DRL) for navigation and motion planning, several research works utilized DRL inside their approach [1], [2]. Figure 1: Arena-Web provides a web-based interface to develop, train, and test navigation approaches conveniently on any computer.