In this work, we use optimal control to change the behavior of a deep reinforcement learning policy by optimizing directly in the policy's latent space. We hypothesize that distinct behavioral patterns, termed behavioral modes, can be identified within certain regions of a deep reinforcement learning policy's latent space, meaning that specific actions or strategies are preferred within these regions. We identify these behavioral modes using latent space dimension-reduction with \ac*{pacmap}. Using the actions generated by the optimal control procedure, we move the system from one behavioral mode to another. We subsequently utilize these actions as a filter for interpreting the neural network policy. The results show that this approach can impose desired behavioral modes in the policy, demonstrated by showing how a failed episode can be made successful and vice versa using the lunar lander reinforcement learning environment.

Simulating a single trajectory of a dynamical system under some state-dependent policy is a core bottleneck in policy optimization algorithms. The many inherently serial policy evaluations that must be performed in a single simulation constitute the bulk of this bottleneck. To wit, in applying policy optimization to supply chain optimization (SCO) problems, simulating a single month of a supply chain can take several hours. We present an iterative algorithm for policy simulation, which we dub Picard Iteration. This scheme carefully assigns policy evaluation tasks to independent processes. Within an iteration, a single process evaluates the policy only on its assigned tasks while assuming a certain 'cached' evaluation for other tasks; the cache is updated at the end of the iteration. Implemented on GPUs, this scheme admits batched evaluation of the policy on a single trajectory. We prove that the structure afforded by many SCO problems allows convergence in a small number of iterations, independent of the horizon. We demonstrate practical speedups of 400x on large-scale SCO problems even with a single GPU, and also demonstrate practical efficacy in other RL environments.

In recent years, reinforcement learning (RL) has emerged as a valuable tool in drug design, offering the potential to propose and optimize molecules with desired properties. However, striking a balance between capabilities, flexibility, reliability, and efficiency remains challenging due to the complexity of advanced RL algorithms and the significant reliance on specialized code. In this work, we introduce ACEGEN, a comprehensive and streamlined toolkit tailored for generative drug design, built using TorchRL, a modern RL library that offers thoroughly tested reusable components. We validate ACEGEN by benchmarking against other published generative modeling algorithms and show comparable or improved performance. We also show examples of ACEGEN applied in multiple drug discovery case studies. ACEGEN is accessible at \url{https://github.com/acellera/acegen-open} and available for use under the MIT license.

Generative Flow Networks (GFlowNets) are amortized sampling methods for learning a stochastic policy to sequentially generate compositional objects with probabilities proportional to their rewards. GFlowNets exhibit a remarkable ability to generate diverse sets of high-reward objects, in contrast to standard return maximization reinforcement learning approaches, which often converge to a single optimal solution. Recent works have arisen for learning goal-conditioned GFlowNets to acquire various useful properties, aiming to train a single GFlowNet capable of achieving different goals as the task specifies. However, training a goal-conditioned GFlowNet poses critical challenges due to extremely sparse rewards, which is further exacerbated in large state spaces. In this work, we propose a novel method named Retrospective Backward Synthesis (RBS) to address these challenges. Specifically, RBS synthesizes a new backward trajectory based on the backward policy in GFlowNets to enrich training trajectories with enhanced quality and diversity, thereby efficiently solving the sparse reward problem. Extensive empirical results show that our method improves sample efficiency by a large margin and outperforms strong baselines on various standard evaluation benchmarks.

A key aspect of intelligence is the ability to demonstrate a broad spectrum of behaviors for adapting to unexpected situations. Over the past decade, advancements in deep reinforcement learning have led to groundbreaking achievements to solve complex continuous control tasks. However, most approaches return only one solution specialized for a specific problem. We introduce Quality-Diversity Actor-Critic (QDAC), an off-policy actor-critic deep reinforcement learning algorithm that leverages a value function critic and a successor features critic to learn high-performing and diverse behaviors. In this framework, the actor optimizes an objective that seamlessly unifies both critics using constrained optimization to (1) maximize return, while (2) executing diverse skills. Compared with other Quality-Diversity methods, QDAC achieves significantly higher performance and more diverse behaviors on six challenging continuous control locomotion tasks. We also demonstrate that we can harness the learned skills to adapt better than other baselines to five perturbed environments. Finally, qualitative analyses showcase a range of remarkable behaviors: adaptive-intelligent-robotics.github.io/QDAC.

Traditional reinforcement learning (RL) methods typically employ a fixed control loop, where each cycle corresponds to an action. This rigidity poses challenges in practical applications, as the optimal control frequency is task-dependent. A suboptimal choice can lead to high computational demands and reduced exploration efficiency. Variable Time Step Reinforcement Learning (VTS-RL) addresses these issues by using adaptive frequencies for the control loop, executing actions only when necessary. This approach, rooted in reactive programming principles, reduces computational load and extends the action space by including action durations. However, VTS-RL's implementation is often complicated by the need to tune multiple hyperparameters that govern exploration in the multi-objective action-duration space (i.e., balancing task performance and number of time steps to achieve a goal). To overcome these challenges, we introduce the Multi-Objective Soft Elastic Actor-Critic (MOSEAC) method. This method features an adaptive reward scheme that adjusts hyperparameters based on observed trends in task rewards during training. This scheme reduces the complexity of hyperparameter tuning, requiring a single hyperparameter to guide exploration, thereby simplifying the learning process and lowering deployment costs. We validate the MOSEAC method through simulations in a Newtonian kinematics environment, demonstrating high task and training performance with fewer time steps, ultimately lowering energy consumption. This validation shows that MOSEAC streamlines RL algorithm deployment by automatically tuning the agent control loop frequency using a single parameter. Its principles can be applied to enhance any RL algorithm, making it a versatile solution for various applications.

Efficient resource utilization and perfect user experience usually conflict with each other in cloud computing platforms. Great efforts have been invested in increasing resource utilization but trying not to affect users' experience for cloud computing platforms. In order to better utilize the remaining pieces of computing resources spread over the whole platform, deferrable jobs are provided with a discounted price to users. For this type of deferrable jobs, users are allowed to submit jobs that will run for a specific uninterrupted duration in a flexible range of time in the future with a great discount. With these deferrable jobs to be scheduled under the remaining capacity after deploying those on-demand jobs, it remains a challenge to achieve high resource utilization and meanwhile shorten the waiting time for users as much as possible in an online manner. In this paper, we propose an online deferrable job scheduling method called \textit{Online Scheduling for DEferrable jobs in Cloud} (\OSDEC{}), where a deep reinforcement learning model is adopted to learn the scheduling policy, and several auxiliary tasks are utilized to provide better state representations and improve the performance of the model. With the integrated reinforcement learning framework, the proposed method can well plan the deployment schedule and achieve a short waiting time for users while maintaining a high resource utilization for the platform. The proposed method is validated on a public dataset and shows superior performance.

This paper introduces a novel transfer learning framework for deep multi-agent reinforcement learning. The approach automatically combines goal-conditioned policies with temporal contrastive learning to discover meaningful sub-goals. The approach involves pre-training a goal-conditioned agent, finetuning it on the target domain, and using contrastive learning to construct a planning graph that guides the agent via sub-goals. Experiments on multi-agent coordination Overcooked tasks demonstrate improved sample efficiency, the ability to solve sparse-reward and long-horizon problems, and enhanced interpretability compared to baselines. The results highlight the effectiveness of integrating goal-conditioned policies with unsupervised temporal abstraction learning for complex multi-agent transfer learning. Compared to state-of-the-art baselines, our method achieves the same or better performances while requiring only 21.7% of the training samples.

The on-orbit intelligent planning of satellites swarm has attracted increasing attention from scholars. Especially in tasks such as the pursuit and attachment of non-cooperative satellites, satellites swarm must achieve coordinated cooperation with limited resources. The study proposes a reinforcement learning framework that integrates the transformer and expert networks. Firstly, under the constraints of incomplete information about non-cooperative satellites, an implicit multi-satellites cooperation strategy was designed using a communication sharing mechanism. Subsequently, for the characteristics of the pursuit-attachment tasks, the multi-agent reinforcement learning framework is improved by introducing transformers and expert networks inspired by transfer learning ideas. To address the issue of satellites swarm scalability, sequence modelling based on transformers is utilized to craft memory-augmented policy networks, meanwhile increasing the scalability of the swarm. By comparing the convergence curves with other algorithms, it is shown that the proposed method is qualified for pursuit-attachment tasks of satellites swarm. Additionally, simulations under different maneuvering strategies of non-cooperative satellites respectively demonstrate the robustness of the algorithm and the task efficiency of the swarm system. The success rate of pursuit-attachment tasks is analyzed through Monte Carlo simulations.

Benchmarks play a crucial role in the development and analysis of reinforcement learning (RL) algorithms. We identify that existing benchmarks used for research into open-ended learning fall into one of two categories. Either they are too slow for meaningful research to be performed without enormous computational resources, like Crafter, NetHack and Minecraft, or they are not complex enough to pose a significant challenge, like Minigrid and Procgen. To remedy this, we first present Craftax-Classic: a ground-up rewrite of Crafter in JAX that runs up to 250x faster than the Python-native original. A run of PPO using 1 billion environment interactions finishes in under an hour using only a single GPU and averages 90% of the optimal reward. To provide a more compelling challenge we present the main Craftax benchmark, a significant extension of the Crafter mechanics with elements inspired from NetHack. Solving Craftax requires deep exploration, long term planning and memory, as well as continual adaptation to novel situations as more of the world is discovered. We show that existing methods including global and episodic exploration, as well as unsupervised environment design fail to make material progress on the benchmark. We believe that Craftax can for the first time allow researchers to experiment in a complex, open-ended environment with limited computational resources.