Reinforcement Learning Domains Research Articles

Depthwise Convolution for Multi-Agent Communication With Enhanced Mean-Field Approximation.

Multi-Agent settings remain a fundamental challenge in the reinforcement learning (RL) domain due to the partial observability and the lack of accurate real-time interactions across agents. In this article, we propose a new method based on local communication learning to tackle the multi-agent RL (MARL) challenge within a large number of agents coexisting. First, we design a new communication protocol that exploits the ability of depthwise convolution to efficiently extract local relations and learn local communication between neighboring agents. To facilitate multi-agent coordination, we explicitly learn the effect of joint actions by taking the policies of neighboring agents as inputs. Second, we introduce the mean-field approximation into our method to reduce the scale of agent interactions. To more effectively coordinate behaviors of neighboring agents, we enhance the mean-field approximation by a supervised policy rectification network (PRN) for rectifying real-time agent interactions and by a learnable compensation term for correcting the approximation bias. The proposed method enables efficient coordination as well as outperforms several baseline approaches on the adaptive traffic signal control (ATSC) task and the StarCraft II multi-agent challenge (SMAC).

Fundamentals of Arthroscopic Surgery Training and beyond: a reinforcement learning exploration and benchmark.

This work presents FASTRL, a benchmark set of instrument manipulation tasks adapted to the domain of reinforcement learning and used in simulated surgical training. This benchmark enables and supports the design and training of human-centric reinforcement learning agents which assist and evaluate human trainees in surgical practice. Simulation tasks from the Fundamentals of Arthroscopic Surgery Training (FAST) program are adapted to the reinforcement learning setting for the purpose of training virtual agents that are capable of providing assistance and scoring to the surgical trainees. A skill performance assessment protocol is presented based on the trained virtual agents. The proposed benchmark suite presents an API for training reinforcement learning agents in the context of arthroscopic skill training. The evaluation scheme based on both heuristic and learned reward functions robustly recovers the ground truth ranking on a diverse test set of human trajectories. The presented benchmark enables the exploration of a novel reinforcement learning-based approach to skill performance assessment and in-procedure assistance for simulated surgical training scenarios. The evaluation protocol based on the learned reward model demonstrates potential for evaluating the performance of surgical trainees in simulation.

Scaling Offline Evaluation of Reinforcement Learning Agents through Abstraction

A critical challenge for the widescale adoption of reinforcement learning (RL) is the need to give domain experts assurance that learned policies will improve decision-making -- and not lead to unacceptable behavior. To meet this challenge, my work aims to develop new methods for offline policy evaluation in real world RL domains. There has been much recent interest in offline evaluation and many advances. However, recent benchmarking efforts have also shown that there remains a substantial gap between current state-of-the-art methods and real world domains such as robotics. Towards scalable offline evaluation, my group is investigating the use of methods for abstraction and representation learning. In this new faculty highlight, I will present our recent results that show the promise of this direction for scaling offline evaluation in RL domains. I will then describe future directions in this line of that work which will further realize the promise of offline policy evaluation for increasing confidence in deployed RL.

A Survey of Deep Reinforcement Learning Based on Multi-Particle Environments

This paper extensively explores the application of deep reinforcement learning within multi-agent environments, a domain marked by intricate challenges and burgeoning opportunities. Multi-agent environments inherently entail the presence of multiple interacting intelligent agents, thereby amplifying the complexity of decision-making and control. Traditional reinforcement learning methods, when confronted with these intricate settings, manifest limitations that have prompted a compelling shift towards the fusion of deep learning and reinforcement learning. The journey commences by elucidating the multifaceted challenges and intricacies pervasive in multi-agent environments. It systematically delineates the constraints of conventional reinforcement learning techniques, underscoring their inadequacies in effectively addressing the multifarious complexities presented by these environments. As the narrative unfolds, it places a spotlight on the transformative potential of deep learning in surmounting the formidable challenges associated with decision-making in multi-agent settings. A critical examination ensues, encompassing a thorough review of pioneering applications of methodologies such as value functions, policy gradients, and actor-critic approaches within the realm of multi-agent systems. Through an exhaustive exploration of existing literature, this article endeavors to unveil the vanguard developments and persisting challenges within the domain of deep reinforcement learning in multi-agent environments. Furthermore, this paper conducts a comparative analysis of the performance of two prominent algorithms, namely Q-learning and the Deep Q Network (DQN) algorithm. This analysis serves as a compass, elucidating the trajectory of development within the sphere of deep reinforcement learning. Ultimately, this paper acts as a compass, guiding future research and innovation in the dynamic and promising field of deep reinforcement learning within multi-agent environments.

Clustering-based attack detection for adversarial reinforcement learning

Detecting malicious attacks presents a major challenge in the field of reinforcement learning (RL), as such attacks can force the victim to perform abnormal actions, with potentially severe consequences. To mitigate these risks, current research focuses on the enhancement of RL algorithms with efficient detection mechanisms, especially for real-world applications. Adversarial attacks have the potential to alter the environmental dynamics of a Markov Decision Process (MDP) perceived by an RL agent. Leveraging these changes in dynamics, we propose a novel approach to detect attacks. Our contribution can be summarized in two main aspects. Firstly, we propose a novel formalization of the attack detection problem that entails analyzing modifications made by attacks to the transition and reward dynamics within the environment. This problem can be framed as a context change detection problem, where the goal is to identify the transition from a “free-of-attack” situation to an “under-attack” scenario. To solve this problem, we propose a groundbreaking “model-free” clustering-based countermeasure. This approach consists of two essential steps: first, partitioning the transition space into clusters, and then using this partitioning to identify changes in environmental dynamics caused by adversarial attacks. To assess the efficiency of our detection method, we performed experiments on four established RL domains (grid-world, mountain car, carpole, and acrobot) and subjected them to four advanced attack types. Uniform, Strategically-timed, Q-value, and Multi-objective. Our study proves that our technique has a high potential for perturbation detection, even in scenarios where attackers employ more sophisticated strategies.

The Application of Residual Connection-Based State Normalization Method in GAIL

In the domain of reinforcement learning (RL), deriving efficacious state representations and maintaining algorithmic stability are crucial for optimal agent performance. However, the inherent dynamism of state representations often complicates the normalization procedure. To overcome these challenges, we present an innovative RL framework that integrates state normalization techniques with residual connections and incorporates attention mechanisms into generative adversarial imitation learning (GAIL). This combination not only enhances the expressive capability of state representations, thereby improving the agent’s accuracy in state recognition, but also significantly mitigates the common issues of gradient dissipation and explosion. Compared to traditional RL algorithms, GAIL combined with the residual connection-based state normalization method empowers the agent to markedly reduce the exploration duration such that feedback concerning rewards in the current state can be provided in real time. Empirical evaluations demonstrate the superior efficacy of this methodology across various RL environments.

A novel reinforcement learning framework for disassembly sequence planning using Q-learning technique optimized using an enhanced simulated annealing algorithm

Abstract The increase in Electrical and Electronic Equipment (EEE) usage in various sectors has given rise to repair and maintenance units. Disassembly of parts requires proper planning, which is done by the Disassembly Sequence Planning (DSP) process. Since the manual disassembly process has various time and labor restrictions, it requires proper planning. Effective disassembly planning methods can encourage the reuse and recycling sector, resulting in reduction of raw-materials mining. An efficient DSP can lower the time and cost consumption. To address the challenges in DSP, this research introduces an innovative framework based on Q-Learning (QL) within the domain of Reinforcement Learning (RL). Furthermore, an Enhanced Simulated Annealing (ESA) algorithm is introduced to improve the exploration and exploitation balance in the proposed RL framework. The proposed framework is extensively evaluated against state-of-the-art frameworks and benchmark algorithms using a diverse set of eight products as test cases. The findings reveal that the proposed framework outperforms benchmark algorithms and state-of-the-art frameworks in terms of time consumption, memory consumption, and solution optimality. Specifically, for complex large products, the proposed technique achieves a remarkable minimum reduction of 60% in time consumption and 30% in memory usage compared to other state-of-the-art techniques. Additionally, qualitative analysis demonstrates that the proposed approach generates sequences with high fitness values, indicating more stable and less time-consuming disassembles. The utilization of this framework allows for the realization of various real-world disassembly applications, thereby making a significant contribution to sustainable practices in EEE industries.

Improving Insertion Loss of Sonic Crystal Active Noise Barrier by Reinforcement Learning and Finite Difference Time Domain Simulations

Improving Insertion Loss of Sonic Crystal Active Noise Barrier by Reinforcement Learning and Finite Difference Time Domain Simulations

Multiagent Reinforcement Learning With Graphical Mutual Information Maximization.

Communication learning is an important research direction in the multiagent reinforcement learning (MARL) domain. Graph neural networks (GNNs) can aggregate the information of neighbor nodes for representation learning. In recent years, several MARL methods leverage GNN to model information interactions between agents to coordinate actions and complete cooperative tasks. However, simply aggregating the information of neighboring agents through GNNs may not extract enough useful information, and the topological relationship information is ignored. To tackle this difficulty, we investigate how to efficiently extract and utilize the rich information of neighbor agents as much as possible in the graph structure, so as to obtain high-quality expressive feature representation to complete the cooperation task. To this end, we present a novel GNN-based MARL method with graphical mutual information (MI) maximization to maximize the correlation between input feature information of neighbor agents and output high-level hidden feature representations. The proposed method extends the traditional idea of MI optimization from graph domain to multiagent system, in which the MI is measured from two aspects: agent features information and agent topological relationships. The proposed method is agnostic to specific MARL methods and can be flexibly integrated with various value function decomposition methods. Considerable experiments on various benchmarks demonstrate that the performance of our proposed method is superior to the existing MARL methods.

Instance Weighted Incremental Evolution Strategies for Reinforcement Learning in Dynamic Environments.

Evolution strategies (ESs), as a family of black-box optimization algorithms, recently emerge as a scalable alternative to reinforcement learning (RL) approaches such as Q-learning or policy gradient and are much faster when many central processing units (CPUs) are available due to better parallelization. In this article, we propose a systematic incremental learning method for ES in dynamic environments. The goal is to adjust previously learned policy to a new one incrementally whenever the environment changes. We incorporate an instance weighting mechanism with ES to facilitate its learning adaptation while retaining scalability of ES. During parameter updating, higher weights are assigned to instances that contain more new knowledge, thus encouraging the search distribution to move toward new promising areas of parameter space. We propose two easy-to-implement metrics to calculate the weights: instance novelty and instance quality. Instance novelty measures an instance's difference from the previous optimum in the original environment, while instance quality corresponds to how well an instance performs in the new environment. The resulting algorithm, instance weighted incremental evolution strategies (IW-IESs), is verified to achieve significantly improved performance on challenging RL tasks ranging from robot navigation to locomotion. This article thus introduces a family of scalable ES algorithms for RL domains that enables rapid learning adaptation to dynamic environments.

Dynamic Weights and Prior Reward in Policy Fusion for Compound Agent Learning

In Deep Reinforcement Learning (DRL) domain, a compound learning task is often decomposed into several sub-tasks in a divide-and-conquer manner, each trained separately and then fused concurrently to achieve the original task, referred to as policy fusion. However, the state-of-the-art (SOTA) policy fusion methods treat the importance of sub-tasks equally throughout the task process, eliminating the possibility of the agent relying on different sub-tasks at various stages. To address this limitation, we propose a generic policy fusion approach, referred to as Policy Fusion Learning with Dynamic Weights and Prior Reward (PFLDWPR), to automate the time-varying selection of sub-tasks. Specifically, PFLDWPR produces a time-varying one-hot vector for sub-tasks to dynamically select a suitable sub-task and mask the rest throughout the entire task process, enabling the fused strategy to optimally guide the agent in executing the compound task. The sub-tasks with the dynamic one-hot vector are then aggregated to obtain the action policy for the original task. Moreover, we collect sub-tasks’s rewards at the pre-training stage as a prior reward, which, along with the current reward, is used to train the policy fusion network. Thus, this approach reduces fusion bias by leveraging prior experience. Experimental results under three popular learning tasks demonstrate that the proposed method significantly improves three SOTA policy fusion methods in terms of task duration, episode reward, and score difference.

Hierarchical Episodic Control

Deep reinforcement learning is one of the research hotspots in artificial intelligence and has been successfully applied in many research areas; however, the low training efficiency and high demand for samples are problems that limit the application. Inspired by the rapid learning mechanisms of the hippocampus, to address these problems, a hierarchical episodic control model extending episodic memory to the domain of hierarchical reinforcement learning is proposed in this paper. The model is theoretically justified and employs a hierarchical implicit memory planning approach for counterfactual trajectory value estimation. Starting from the final step and recursively moving back along the trajectory, a hidden plan is formed within the episodic memory. Experience is aggregated both along trajectories and across trajectories, and the model is updated using a multi-headed backpropagation similar to bootstrapped neural networks. This model extends the parameterized episodic memory framework to the realm of hierarchical reinforcement learning and is theoretically analyzed to demonstrate its convergence and effectiveness. Experiments conducted in four-room games, Mujoco, and UE4-based active tracking highlight that the hierarchical episodic control model effectively enhances training efficiency. It demonstrates notable improvements in both low-dimensional and high-dimensional environments, even in cases of sparse rewards. This model can enhance the training efficiency of reinforcement learning and is suitable for application scenarios that do not rely heavily on exploration, such as unmanned aerial vehicles, robot control, computer vision applications, and so on.

A hierarchical multi-agent allocation-action learning framework for multi-subtask games

Great progress has been made in the domain of multi-agent reinforcement learning in recent years. Most work concentrates on solving a single task by learning the cooperative behaviors of agents. However, many real-world problems are normally composed of a set of subtasks in which the execution order follows a certain procedure. Cooperative behaviors should be learned on the premise that agents are first allocated to those subtasks. In this paper, we propose a hierarchical framework for learning the dynamic allocation of agents among subtasks, as well as cooperative behaviors. We design networks corresponding to agents and subnetworks, respectively, which constitute the whole hierarchical network. For the upper layer, a novel allocation learning mechanism is devised to map an agent network to a subtask network. Each agent network could only be assigned to only one subtask network at each time step. For the lower layer, an action learning module is designed to compute appropriate actions for each agent with the allocation result. The agent networks together with the subtask networks are updated by a common reward obtained from the environment. To evaluate the effectiveness of our framework, we conduct experiments in two challenging environments, i.e., Google Research Football and SAVETHECITY. Empirical results show that our framework achieves much better performance than other recent methods.

Achieving efficient interpretability of reinforcement learning via policy distillation and selective input gradient regularization

Although deep Reinforcement Learning (RL) has proven successful in a wide range of tasks, one challenge it faces is interpretability when applied to real-world problems. Saliency maps are frequently used to provide interpretability for deep neural networks. However, in the RL domain, existing saliency map approaches are either computationally expensive and thus cannot satisfy the real-time requirement of real-world scenarios or cannot produce interpretable saliency maps for RL policies. In this work, we propose an approach of Distillation with selective Input Gradient Regularization (DIGR) which uses policy distillation and input gradient regularization to produce new policies that achieve both high interpretability and computation efficiency in generating saliency maps. Our approach is also found to improve the robustness of RL policies to multiple adversarial attacks. We conduct experiments on three tasks, MiniGrid (Fetch Object), Atari (Breakout) and CARLA Autonomous Driving, to demonstrate the importance and effectiveness of our approach.

Goal Agnostic Learning and Planning without Reward Functions

In this paper we present an algorithm, the Goal Agnostic Planner (GAP), which combines elements of Reinforcement Learning (RL) and Markov Decision Processes (MDPs) into an elegant, effective system for learning to solve sequential problems. The GAP algorithm does not require the design of either an explicit world model or a reward function to drive policy determination, and is capable of operating on both MDP and RL domain problems. The construction of the GAP lends itself to several analytic guarantees such as policy optimality, exponential goal achievement rates, reciprocal learning rates, measurable robustness to error, and explicit convergence conditions for abstracted states. Empirical results confirm these predictions, demonstrate effectiveness over a wide range of domains, and show that the GAP algorithm performance is an order of magnitude faster than standard reinforcement learning and produces plans of equal quality to MDPs, without requiring design of reward functions.

Automated cloud resources provisioning with the use of the proximal policy optimization

Many modern applications, both scientific and commercial, are deployed to cloud environments and often employ multiple types of resources. That allows them to efficiently allocate only the resources which are actually needed to achieve their goals. However, in many workloads the actual usage of the infrastructure varies over time, which results in over-provisioning and unnecessarily high costs. In such cases, automatic resource scaling can provide significant cost savings by provisioning only the amount of resources which are necessary to support the current workload. Unfortunately, due to the complex nature of distributed systems, automatic scaling remains a challenge. Reinforcement learning domain has been recently a very active field of research. Thanks to combining it with Deep Learning, many newly designed algorithms improve the state of the art in many complex domains. In this paper we present the results of our attempt to use the recent advancements in Reinforcement Learning to optimize the cost of running a compute-intensive evolutionary process by automating the scaling of heterogeneous resources in a compute cloud environment. We describe the architecture of our system and present evaluation results. The experiments include autonomous management of a sample workload and a comparison of its performance to the traditional automatic threshold-based management approach. We also provide the details of training of the management policy using the proximal policy optimization algorithm. Finally, we discuss the feasibility to extend the presented approach to further scenarios.

Strategic maneuver and disruption with reinforcement learning approaches for multi-agent coordination

Reinforcement learning (RL) approaches can illuminate emergent behaviors that facilitate coordination across teams of agents as part of a multi-agent system (MAS), which can provide windows of opportunity in various military tasks. Technologically advancing adversaries pose substantial risks to a friendly nation’s interests and resources. Superior resources alone are not enough to defeat adversaries in modern complex environments because adversaries create standoff in multiple domains against predictable military doctrine-based maneuvers. Therefore, as part of a defense strategy, friendly forces must use strategic maneuvers and disruption to gain superiority in complex multi-faceted domains, such as multi-domain operations (MDOs). One promising avenue for implementing strategic maneuver and disruption to gain superiority over adversaries is through coordination of MAS in future military operations. In this paper, we present overviews of prominent works in the RL domain with their strengths and weaknesses for overcoming the challenges associated with performing autonomous strategic maneuver and disruption in military contexts.

Reinforcement Learning for Classical Planning: Viewing Heuristics as Dense Reward Generators

Recent advances in reinforcement learning (RL) have led to a growing interest in applying RL to classical planning domains or applying classical planning methods to some complex RL domains. However, the long-horizon goal-based problems found in classical planning lead to sparse rewards for RL, making direct application inefficient. In this paper, we propose to leverage domain-independent heuristic functions commonly used in the classical planning literature to improve the sample efficiency of RL. These classical heuristics act as dense reward generators to alleviate the sparse-rewards issue and enable our RL agent to learn domain-specific value functions as residuals on these heuristics, making learning easier. Correct application of this technique requires consolidating the discounted metric used in RL and the non-discounted metric used in heuristics. We implement the value functions using Neural Logic Machines, a neural network architecture designed for grounded first-order logic inputs. We demonstrate on several classical planning domains that using classical heuristics for RL allows for good sample efficiency compared to sparse-reward RL. We further show that our learned value functions generalize to novel problem instances in the same domain. The source code and the appendix are available at github.com/ibm/pddlrl and arxiv.org/abs/2109.14830.

Scanning the Issue

<b>Deep Reinforcement Learning for Autonomous Driving: A Survey</b><i>R. Kiran, I. Sobh, V. Talpaert, P. Mannion, A. A. Al Sallab, S. Yogamani, and P. P&#x00E9;rez</i>With the development of deep representation learning, the domain of reinforcement learning (RL) has become a powerful learning framework now capable of learning complex policies in high-dimensional environments. This review summarizes deep reinforcement learning (DRL) algorithms and provides a taxonomy of automated driving tasks where (D)RL methods have been employed, while addressing key computational challenges in the real-world deployment of autonomous driving agents. It also delineates adjacent domains such as behavior cloning, imitation learning, and inverse reinforcement learning that are related but are not classical RL algorithms. The role of simulators in training agents, methods to validate, test, and robustify existing solutions in RL are discussed.

Test-Driven Reward Function for Reinforcement Learning: A Contribution towards Applicable Machine Learning Algorithms for Production Systems

Reinforcement Learning algorithms find more and more application in fields where complex tasks need to be solved. The automation of production systems is one of those fields. Normally, programming a control system defines the automation strategy. Previous contributions by the authors have shown that a so-called agent can learn automation strategies for production systems using a Reinforcement Learning setup. However, the development of the reward function for the agent can be challenging and needs Reinforcement Learning domain knowledge. This paper introduces a novel approach in combining Test-Driven Development with Reinforcement Learning in order to solve the problem of a suitable reward function. In the presented approach predefined test cases are used to derive rewards for the agent. The use of an automated test framework allows for continuous learning sequences until all test cases are passed. An application example of a robot cell is used to demonstrate the novel approach and verify its suitability and usability. The first application shows promising results for further examination towards more application scenarios.