Deep Reinforcement Learning (DRL) connects the classic Reinforcement Learning algorithms with Deep Neural Networks. A problem in DRL is that CNNs are black-boxes and it is hard to understand the decision-making process of agents. In order to be able to use RL agents in highly dangerous environments for humans and machines, the developer needs a debugging tool to assure that the agent does what is expected. Currently, rewards are primarily used to interpret how well an agent is learning. However, this can lead to deceptive conclusions if the agent receives more rewards by memorizing a policy and not learning to respond to the environment. In this work, it is shown that this problem can be recognized with the help of gradient visualization techniques. This work brings some of the best-known visualization methods from the field of image classification to the area of Deep Reinforcement Learning. Furthermore, two new visualization techniques have been developed, one of which provides particularly good results. It is being proven to what extent the algorithms can be used in the area of Reinforcement learning. Also, the question arises on how well the DRL algorithms can be visualized across different environments with varying visualization techniques.
Saliency maps are frequently used to support explanations of the behavior of deep reinforcement learning (RL) agents. However, a review of how saliency maps are used in practice indicates that the derived explanations are often unfalsifiable and can be highly subjective. We introduce an empirical approach grounded in counterfactual reasoning to test the hypotheses generated from saliency maps and assess the degree to which they correspond to the semantics of RL environments. We use Atari games, a common benchmark for deep RL, to evaluate three types of saliency maps. Our results show the exten…
IRT Saint-Exupéry（圣埃克苏佩里工业技术研究院）；ISAE-SUPAERO（法国高等航空航天学院）；IMT, INSA Toulouse（图卢兹国立应用科学学院数学与信息学研究所）；ANITI（图卢兹人工智能与自然智能研究所）；IMT, Université Paul Sabatier（保罗·萨巴蒂埃大学数学与信息学研究所）；Université de Toulouse（图卢兹大学）
Deep reinforcement learning policies, despite their outstanding efficiency in simulated visual control tasks, have shown disappointing ability to generalize across disturbances in the input training images. Changes in image statistics or distracting background elements are pitfalls that prevent generalization and real-world applicability of such control policies. We elaborate on the intuition that a good visual policy should be able to identify which pixels are important for its decision, and preserve this identification of important sources of information across images. This implies that trai…
Attention Trajectories as a Diagnostic Axis for Deep Reinforcement Learning
Max Planck Institute for Human Cognitive and Brain Sciences（马克斯·普朗克人类认知与脑科学研究所）；Center for Scalable Data Analytics and Artificial Intelligence (ScaDS.AI)（可扩展数据与人工智能研究中心）；Leipzig University（莱比锡大学）
Samsung AI Center（三星人工智能中心）；Steklov Institute of Mathematics at St. Petersburg（圣彼得堡斯捷克洛夫数学研究所）；Neuromation OU（Neuromation公司）
We propose a new approach to visualize saliency maps for deep neural network models and apply it to deep reinforcement learning agents trained on Atari environments. Our method adds an attention module that we call FLS (Free Lunch Saliency) to the feature extractor from an established baseline (Mnih et al., 2015). This addition results in a trainable model that can produce saliency maps, i.e., visualizations of the importance of different parts of the input for the agent's current decision making. We show experimentally that a network with an FLS module exhibits performance similar to the base…
One of the most prominent methods for explaining the behavior of Deep Reinforcement Learning (DRL) agents is the generation of saliency maps that show how much each pixel attributed to the agents' decision. However, there is no work that computationally evaluates and compares the fidelity of different saliency map approaches specifically for DRL agents. It is particularly challenging to computationally evaluate saliency maps for DRL agents since their decisions are part of an overarching policy. For instance, the output neurons of value-based DRL algorithms encode both the value of the current…
UT Austin（得克萨斯大学奥斯汀分校）；Stanford University（斯坦福大学）；Sony AI（索尼人工智能）
Deep reinforcement learning (RL) algorithms are powerful tools for solving visuomotor decision tasks. However, the trained models are often difficult to interpret, because they are represented as end-to-end deep neural networks. In this paper, we shed light on the inner workings of such trained models by analyzing the pixels that they attend to during task execution, and comparing them with the pixels attended to by humans executing the same tasks. To this end, we investigate the following two questions that, to the best of our knowledge, have not been previously studied. 1) How similar are th…
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) whi…
Abstractive text summarization is the task of compressing and rewriting a long document into a short summary while maintaining saliency, directed logical entailment, and non-redundancy. In this work, we address these three important aspects of a good summary via a reinforcement learning approach with two novel reward functions: ROUGESal and Entail, on top of a coverage-based baseline. The ROUGESal reward modifies the ROUGE metric by up-weighting the salient phrases/words detected via a keyphrase classifier. The Entail reward gives high (length-normalized) scores to logically-entailed summaries…
National University of Ireland, Galway（爱尔兰国立大学戈尔韦分校）；University of Central Florida（中佛罗里达大学）；Iran University of Science and Technology（伊朗科技大学）
Bottom-Up (BU) saliency models do not perform well in complex interactive environments where humans are actively engaged in tasks (e.g., sandwich making and playing the video games). In this paper, we leverage Reinforcement Learning (RL) to highlight task-relevant locations of input frames. We propose a soft attention mechanism combined with the Deep Q-Network (DQN) model to teach an RL agent how to play a game and where to look by focusing on the most pertinent parts of its visual input. Our evaluations on several Atari 2600 games show that the soft attention based model could predict fixatio…
Fraunhofer Institute for Integrated Circuits IIS（弗劳恩霍夫集成电路研究所IIS）；Friedrich-Alexander-Universität Erlangen-Nürnberg（埃尔朗根-纽伦堡弗里德里希-亚历山大大学）
Autonomous driving has the potential to revolutionize mobility and is hence an active area of research. In practice, the behavior of autonomous vehicles must be acceptable, i.e., efficient, safe, and interpretable. While vanilla reinforcement learning (RL) finds performant behavioral strategies, they are often unsafe and uninterpretable. Safety is introduced through Safe RL approaches, but they still mostly remain uninterpretable as the learned behaviour is jointly optimized for safety and performance without modeling them separately. Interpretable machine learning is rarely applied to RL. Thi…
This paper introduces Dex, a reinforcement learning environment toolkit specialized for training and evaluation of continual learning methods as well as general reinforcement learning problems. We also present the novel continual learning method of incremental learning, where a challenging environment is solved using optimal weight initialization learned from first solving a similar easier environment. We show that incremental learning can produce vastly superior results than standard methods by providing a strong baseline method across ten Dex environments. We finally develop a saliency metho…
Deep reinforcement learning has become popular over recent years, showing superiority on different visual-input tasks such as playing Atari games and robot navigation. Although objects are important image elements, few work considers enhancing deep reinforcement learning with object characteristics. In this paper, we propose a novel method that can incorporate object recognition processing to deep reinforcement learning models. This approach can be adapted to any existing deep reinforcement learning frameworks. State-of-the-art results are shown in experiments on Atari games. We also propose a…
Technische Universität Berlin（柏林工业大学）；Zuse Institute Berlin（柏林祖斯研究所）
One of the goals of Explainable AI (XAI) is to determine which input components were relevant for a classifier decision. This is commonly know as saliency attribution. Characteristic functions (from cooperative game theory) are able to evaluate partial inputs and form the basis for theoretically "fair" attribution methods like Shapley values. Given only a standard classifier function, it is unclear how partial input should be realised. Instead, most XAI-methods for black-box classifiers like neural networks consider counterfactual inputs that generally lie off-manifold. This makes them hard to…
Universität Augsburg（奥格斯堡大学）；Technion - Israel Institute of Technology（以色列理工学院）
With advances in reinforcement learning (RL), agents are now being developed in high-stakes application domains such as healthcare and transportation. Explaining the behavior of these agents is challenging, as the environments in which they act have large state spaces, and their decision-making can be affected by delayed rewards, making it difficult to analyze their behavior. To address this problem, several approaches have been developed. Some approaches attempt to convey the $\textit{global}$ behavior of the agent, describing the actions it takes in different states. Other approaches devised…
Higher Resolution, Better Generalization: Unlocking Visual Scaling in Deep Reinforcement Learning
TUM School of Engineering and Design（技术大学慕尼黑工程与设计学院）；Technical University of Munich（技术大学慕尼黑）；KUIS AI Center（KUIS人工智能中心）；Koç University（科克大学）；Department of Computer Engineering（计算机工程系）
School of Computer Science, McGill University（麦吉尔大学计算机科学学院）；Department of Electrical, Computer, and Software Engineering, McGill University（麦吉尔大学电气、计算机与软件工程系）
Reinforcement learning has shown promise in learning policies that can solve complex problems. However, manually specifying a good reward function can be difficult, especially for intricate tasks. Inverse reinforcement learning offers a useful paradigm to learn the underlying reward function directly from expert demonstrations. Yet in reality, the corpus of demonstrations may contain trajectories arising from a diverse set of underlying reward functions rather than a single one. Thus, in inverse reinforcement learning, it is useful to consider such a decomposition. The options framework in rei…
Reinforcement Learning faces an important challenge in partial observable environments that has long-term dependencies. In order to learn in an ambiguous environment, an agent has to keep previous perceptions in a memory. Earlier memory based approaches use a fixed method to determine what to keep in the memory, which limits them to certain problems. In this study, we follow the idea of giving the control of the memory to the agent by allowing it to have memory-changing actions. This learning mechanism is supported by an intrinsic motivation to memorize rare observations that can help the agen…
Carnegie Mellon University（卡内基梅隆大学）；Google Inc.（谷歌公司）；University of Pittsburgh（匹兹堡大学）
Autonomous AI systems will be entering human society in the near future to provide services and work alongside humans. For those systems to be accepted and trusted, the users should be able to understand the reasoning process of the system, i.e. the system should be transparent. System transparency enables humans to form coherent explanations of the system's decisions and actions. Transparency is important not only for user trust, but also for software debugging and certification. In recent years, Deep Neural Networks have made great advances in multiple application areas. However, deep neural…
SafeRL-Lite: A Lightweight, Explainable, and Constrained Reinforcement Learning Library
Vision Mentors Ltd.（愿景导师有限公司）；School of Computer Science & AI, SR University（计算机科学与人工智能学院，SR大学）；Department of Computer Science, MANIT（计算机科学系，MANIT）；University of West London（西伦敦大学）
All Models are Wrong, Knowing Where is Useful: On Model Uncertainty in Reinforcement Learning
German Federal Ministry of Research, Technology and Space (BMFTR)（德国联邦研究、技术和空间部）；Robotics Institute Germany (RIG)（德国机器人研究所）；Institute for Data Science in Mechanical Engineering, RWTH Aachen University（机械工程数据科学研究所，亚琛工业大学）；NHR Center NHR4CES at RWTH Aachen University（亚琛工业大学NHR4CES中心）
This paper presents a novel state representation for reward-free Markov decision processes. The idea is to learn, in a self-supervised manner, an embedding space where distances between pairs of embedded states correspond to the minimum number of actions needed to transition between them. Compared to previous methods, our approach does not require any domain knowledge, learning from offline and unlabeled data. We show how this representation can be leveraged to learn goal-conditioned policies, providing a notion of similarity between states and goals and a useful heuristic distance to guide pl…
University of Alberta（阿尔伯塔大学）；Noah’s Ark Lab, Huawei Technologies Canada（华为加拿大诺亚方舟实验室）
Auxiliary tasks have been argued to be useful for representation learning in reinforcement learning. Although many auxiliary tasks have been empirically shown to be effective for accelerating learning on the main task, it is not yet clear what makes useful auxiliary tasks. Some of the most promising results are on the pixel control, reward prediction, and the next state prediction auxiliary tasks; however, the empirical results are mixed, showing substantial improvements in some cases and marginal improvements in others. Careful investigations of how auxiliary tasks help the learning of the ma…
Shanghai Key Laboratory of Intelligent Information Processing（上海智能信息处理重点实验室）；School of Computer Science, Fudan University（复旦大学计算机科学学院）
We explore the black-box adversarial attack on video recognition models. Attacks are only performed on selected key regions and key frames to reduce the high computation cost of searching adversarial perturbations on a video due to its high dimensionality. To select key frames, one way is to use heuristic algorithms to evaluate the importance of each frame and choose the essential ones. However, it is time inefficient on sorting and searching. In order to speed up the attack process, we propose a reinforcement learning based frame selection strategy. Specifically, the agent explores the differ…
Several machine learning and deep learning frameworks have been proposed to solve remaining useful life estimation and failure prediction problems in recent years. Having access to the remaining useful life estimation or likelihood of failure in near future helps operators to assess the operating conditions and, therefore, provides better opportunities for sound repair and maintenance decisions. However, many operators believe remaining useful life estimation and failure prediction solutions are incomplete answers to the maintenance challenge. They argue that knowing the likelihood of failure …
Tel Aviv University（特拉维夫大学）；Google Research（谷歌研究院）；Duke University（杜克大学）
Interpretability is an essential building block for trustworthiness in reinforcement learning systems. However, interpretability might come at the cost of deteriorated performance, leading many researchers to build complex models. Our goal is to analyze the cost of interpretability. We show that in certain cases, one can achieve policy interpretability while maintaining its optimality. We focus on a classical problem from reinforcement learning: mazes with $k$ obstacles in $\mathbb{R}^d$. We prove the existence of a small decision tree with a linear function at each inner node and depth $O(\lo…
Quantum Annealing Enhanced Reinforcement Learning for Accurate Remaining Useful Lifetime Prediction
Central University of Karnataka（卡纳塔克中央大学）；University College of Engineering, Anna University（安娜大学工程学院）；AIONOS India Pvt Ltd（AIONOS印度私人有限公司）；National Institute of Technology Tiruchirappalli（蒂鲁吉拉帕利国立理工学院）
University College London（伦敦大学学院）；Meta AI；University of Oxford（牛津大学）
Practising and honing skills forms a fundamental component of how humans learn, yet artificial agents are rarely specifically trained to perform them. Instead, they are usually trained end-to-end, with the hope being that useful skills will be implicitly learned in order to maximise discounted return of some extrinsic reward function. In this paper, we investigate how skills can be incorporated into the training of reinforcement learning (RL) agents in complex environments with large state-action spaces and sparse rewards. To this end, we created SkillHack, a benchmark of tasks and associated …
In this work, we investigate the properties of data that cause popular representation learning approaches to fail. In particular, we find that in environments where states do not significantly overlap, variational autoencoders (VAEs) fail to learn useful features. We demonstrate this failure in a simple gridworld domain, and then provide a solution in the form of metric learning. However, metric learning requires supervision in the form of a distance function, which is absent in reinforcement learning. To overcome this, we leverage the sequential nature of states in a replay buffer to approxim…
STRIDE: Strategic Trajectory Reasoning via Discriminative Estimation for Verifiable Reinforcement Learning
Kean University（基恩大学）；Case Western Reserve University（凯斯西储大学）；University of Texas at Austin（德克萨斯大学奥斯汀分校）；The Ohio State University（俄亥俄州立大学）；Tongji University（同济大学）；Duke Kunshan University（昆山杜克大学）；Royal Melbourne Institute of Technology（皇家墨尔本理工大学）
While recent progress in deep reinforcement learning has enabled robots to learn complex behaviors, tasks with long horizons and sparse rewards remain an ongoing challenge. In this work, we propose an effective reward shaping method through predictive coding to tackle sparse reward problems. By learning predictive representations offline and using these representations for reward shaping, we gain access to reward signals that understand the structure and dynamics of the environment. In particular, our method achieves better learning by providing reward signals that 1) understand environment dy…
University of Alberta（阿尔伯塔大学）；Tsinghua University（清华大学）
Episodic memory is a psychology term which refers to the ability to recall specific events from the past. We suggest one advantage of this particular type of memory is the ability to easily assign credit to a specific state when remembered information is found to be useful. Inspired by this idea, and the increasing popularity of external memory mechanisms to handle long-term dependencies in deep learning systems, we propose a novel algorithm which uses a reservoir sampling procedure to maintain an external memory consisting of a fixed number of past states. The algorithm allows a deep reinforc…
In deep reinforcement learning (RL), data augmentation is widely considered as a tool to induce a set of useful priors about semantic consistency and improve sample efficiency and generalization performance. However, even when the prior is useful for generalization, distilling it to RL agent often interferes with RL training and degenerates sample efficiency. Meanwhile, the agent is forgetful of the prior due to the non-stationary nature of RL. These observations suggest two extreme schedules of distillation: (i) over the entire training; or (ii) only at the end. Hence, we devise a stand-alone…
Tokyo University of Science（东京理科大学）；The University of Tokyo（东京大学）
This paper tackles a new problem setting: reinforcement learning with pixel-wise rewards (pixelRL) for image processing. After the introduction of the deep Q-network, deep RL has been achieving great success. However, the applications of deep reinforcement learning (RL) for image processing are still limited. Therefore, we extend deep RL to pixelRL for various image processing applications. In pixelRL, each pixel has an agent, and the agent changes the pixel value by taking an action. We also propose an effective learning method for pixelRL that significantly improves the performance by consid…
In this paper, we utilize results from convex analysis and monotone operator theory to derive additional properties of the softmax function that have not yet been covered in the existing literature. In particular, we show that the softmax function is the monotone gradient map of the log-sum-exp function. By exploiting this connection, we show that the inverse temperature parameter determines the Lipschitz and co-coercivity properties of the softmax function. We then demonstrate the usefulness of these properties through an application in game-theoretic reinforcement learning.
We hypothesize that empirically studying the sample complexity of offline reinforcement learning (RL) is crucial for the practical applications of RL in the real world. Several recent works have demonstrated the ability to learn policies directly from offline data. In this work, we ask the question of the dependency on the number of samples for learning from offline data. Our objective is to emphasize that studying sample complexity for offline RL is important, and is an indicator of the usefulness of existing offline algorithms. We propose an evaluation approach for sample complexity analysis…
In deep reinforcement learning (RL), data augmentation is widely considered as a tool to induce a set of useful priors about semantic consistency and improve sample efficiency and generalization performance. However, even when the prior is useful for generalization, distilling it to RL agent often interferes with RL training and degenerates sample efficiency. Meanwhile, the agent is forgetful of the prior due to the non-stationary nature of RL. These observations suggest two extreme schedules of distillation: (i) over the entire training; or (ii) only at the end. Hence, we devise a stand-alone…
Deep reinforcement learning has the potential to train robots to perform complex tasks in the real world without requiring accurate models of the robot or its environment. A practical approach is to train agents in simulation, and then transfer them to the real world. One popular method for achieving transferability is to use domain randomisation, which involves randomly perturbing various aspects of a simulated environment in order to make trained agents robust to the reality gap. However, less work has gone into understanding such agents - which are deployed in the real world - beyond task p…
Google Brain（谷歌大脑）；DeepMind（深度思考）；University of Alberta（阿尔伯塔大学）；University of Oxford（牛津大学）
We propose a new perspective on representation learning in reinforcement learning based on geometric properties of the space of value functions. We leverage this perspective to provide formal evidence regarding the usefulness of value functions as auxiliary tasks. Our formulation considers adapting the representation to minimize the (linear) approximation of the value function of all stationary policies for a given environment. We show that this optimization reduces to making accurate predictions regarding a special class of value functions which we call adversarial value functions (AVFs). We …
Deakin University（迪肯大学）；Institute for Intelligent Systems Research and Innovation (IISRI)（智能系统研究与创新研究所（IISRI））
Reinforcement learning has been applied to human movement through physiologically-based biomechanical models to add insights into the neural control of these movements; it is also useful in the design of prosthetics and robotics. In this paper, we extend the use of reinforcement learning into controlling an ocular biomechanical system to perform saccades, which is one of the fastest eye movement systems. We describe an ocular environment and an agent trained using Deep Deterministic Policy Gradients method to perform saccades. The agent was able to match the desired eye position with a mean de…
Reinforcement learning has received significant interest in recent years, due primarily to the successes of deep reinforcement learning at solving many challenging tasks such as playing Chess, Go and online computer games. However, with the increasing focus on reinforcement learning, applications outside of gaming and simulated environments require understanding the robustness, stability, and resilience of reinforcement learning methods. To this end, we conducted a comprehensive literature review to characterize the available literature on these three behaviors as they pertain to reinforcement…
Inverse Reinforcement Learning (IRL) is the problem of finding a reward function which describes observed/known expert behavior. The IRL setting is remarkably useful for automated control, in situations where the reward function is difficult to specify manually or as a means to extract agent preference. In this work, we provide a new IRL algorithm for the continuous state space setting with unknown transition dynamics by modeling the system using a basis of orthonormal functions. Moreover, we provide a proof of correctness and formal guarantees on the sample and time complexity of our algorith…
The problem of sparse rewards is one of the hardest challenges in contemporary reinforcement learning. Hierarchical reinforcement learning (HRL) tackles this problem by using a set of temporally-extended actions, or options, each of which has its own subgoal. These subgoals are normally handcrafted for specific tasks. Here, though, we introduce a generic class of subgoals with broad applicability in the visual domain. Underlying our approach (in common with work using "auxiliary tasks") is the hypothesis that the ability to control aspects of the environment is an inherently useful skill to ha…
Some phenomena related to statistical noise which have been investigated by various authors under the framework of deep reinforcement learning (RL) algorithms are discussed. The following algorithms are examined: the deep Q-network (DQN), double DQN, deep deterministic policy gradient (DDPG), twin-delayed DDPG (TD3), and hill climbing algorithm. First, we consider overestimation, which is a harmful property resulting from noise. Then we deal with noise used for exploration, this is the useful noise. We discuss setting the noise parameter in the TD3 for typical PyBullet environments associated …
University of Oxford（牛津大学）；Beihang University（北京航空航天大学）；Hebei University of Technology（河北工业大学）
Hierarchical reinforcement learning (HRL) has recently shown promising advances on speeding up learning, improving the exploration, and discovering intertask transferable skills. Most recent works focus on HRL with two levels, i.e., a master policy manipulates subpolicies, which in turn manipulate primitive actions. However, HRL with multiple levels is usually needed in many real-world scenarios, whose ultimate goals are highly abstract, while their actions are very primitive. Therefore, in this paper, we propose a diversity-driven extensible HRL (DEHRL), where an extensible and scalable frame…
