The control problem of the flexible wing aircraft is challenging due to the prevailing and high nonlinear deformations in the flexible wing system. This urged for new control mechanisms that are robust to the real-time variations in the wing's aerodynamics. An online control mechanism based on a value iteration reinforcement learning process is developed for flexible wing aerial structures. It employs a model-free control policy framework and a guaranteed convergent adaptive learning architecture to solve the system's Bellman optimality equation. A Riccati equation is derived and shown to be equivalent to solving the underlying Bellman equation. The online reinforcement learning solution is implemented using means of an adaptive-critic mechanism. The controller is proven to be asymptotically stable in the Lyapunov sense. It is assessed through computer simulations and its superior performance is demonstrated on two scenarios under different operating conditions.

Policy gradient methods are appealing in deep reinforcement learning but suffer from high variance of gradient estimate. To reduce the variance, the state value function is applied commonly. However, the effect of the state value function becomes limited in stochastic dynamic environments, where the unexpected state dynamics and rewards will increase the variance. In this paper, we propose to replace the state value function with a novel hindsight value function, which leverages the information from the future to reduce the variance of the gradient estimate for stochastic dynamic environments. Particularly, to obtain an ideally unbiased gradient estimate, we propose an information-theoretic approach, which optimizes the embeddings of the future to be independent of previous actions. In our experiments, we apply the proposed hindsight value function in stochastic dynamic environments, including discrete-action environments and continuous-action environments. Compared with the standard state value function, the proposed hindsight value function consistently reduces the variance, stabilizes the training, and improves the eventual policy.

All the sessions from Transform 2021 are available on-demand now. This article is part of our reviews of AI research papers, a series of posts that explore the latest findings in artificial intelligence. One of the key challenges of deep reinforcement learning models -- the kind of AI systems that have mastered Go, StarCraft 2, and other games -- is their inability to generalize their capabilities beyond their training domain. This limit makes it very hard to apply these systems to real-world settings, where situations are much more complicated and unpredictable than the environments where AI models are trained. But scientists at AI research lab DeepMind claim to have taken the "first steps to train an agent capable of playing many different games without needing human interaction data," according to a blog post about their new "open-ended learning" initiative. Their new project includes a 3D environment with realistic dynamics and deep reinforcement learning agents that can learn to solve a wide range of challenges.

Policy learning for delicate industrial insertion tasks (e.g., PC board assembly) is challenging. This paper considers two major problems: how to learn a diversified policy (instead of just one average policy) that can efficiently handle different workpieces with minimum amount of training data, and how to handle defects of workpieces during insertion. To address the problems, we propose tolerance-guided policy learning. To encourage transferability of the learned policy to different workpieces, we add a task embedding to the policy's input space using the insertion tolerance. Then we train the policy using generative adversarial imitation learning with reward shaping (RS-GAIL) on a variety of representative situations. To encourage adaptability of the learned policy to handle defects, we build a probabilistic inference model that can output the best inserting pose based on failed insertions using the tolerance model. The best inserting pose is then used as a reference to the learned policy. This proposed method is validated on a sequence of IC socket insertion tasks in simulation. The results show that 1) RS-GAIL can efficiently learn optimal policies under sparse rewards; 2) the tolerance embedding can enhance the transferability of the learned policy; 3) the probabilistic inference makes the policy robust to defects on the workpieces.

This work proposes a new class of proactive attacks called the Informational Denial-of-Service (IDoS) attacks that exploit the attentional human vulnerability. By generating a large volume of feints, IDoS attacks deplete the cognition resources of human operators to prevent humans from identifying the real attacks hidden among feints. This work aims to formally define IDoS attacks, quantify their consequences, and develop human-assistive security technologies to mitigate the severity level and risks of IDoS attacks. To this end, we model the feint and real attacks' sequential arrivals with category labels as a semi-Markov process. The assistive technology strategically manages human attention by highlighting selective alerts periodically to prevent the distraction of other alerts. A data-driven approach is applied to evaluate human performance under different Attention Management (AM) strategies. Under a representative special case, we establish the computational equivalency between two dynamic programming representations to simplify the theoretical computation and the online learning. A case study corroborates the effectiveness of the learning framework. The numerical results illustrate how AM strategies can alleviate the severity level and the risk of IDoS attacks. Furthermore, we characterize the fundamental limits of the minimum severity level under all AM strategies and the maximum length of the inspection period to reduce the IDoS risks.

The introduction of the generative adversarial imitation learning (GAIL) algorithm has spurred the development of scalable imitation learning approaches using deep neural networks. The GAIL objective can be thought of as 1) matching the expert policy's state distribution; 2) penalising the learned policy's state distribution; and 3) maximising entropy. While theoretically motivated, in practice GAIL can be difficult to apply, not least due to the instabilities of adversarial training. In this paper, we take a pragmatic look at GAIL and related imitation learning algorithms. We implement and automatically tune a range of algorithms in a unified experimental setup, presenting a fair evaluation between the competing methods. From our results, our primary recommendation is to consider non-adversarial methods. Furthermore, we discuss the common components of imitation learning objectives, and present promising avenues for future research.

To increase autonomy in reinforcement learning, agents need to learn useful behaviours without reliance on manually designed reward functions. To that end, skill discovery methods have been used to learn the intrinsic options available to an agent using task-agnostic objectives. However, without the guidance of task-specific rewards, emergent behaviours are generally useless due to the under-constrained problem of skill discovery in complex and high-dimensional spaces. This paper proposes a framework for guiding the skill discovery towards the subset of expert-visited states using a learned state projection. We apply our method in various reinforcement learning (RL) tasks and show that such a projection results in more useful behaviours.

Risk-bounded motion planning is an important yet difficult problem for safety-critical tasks. While existing mathematical programming methods offer theoretical guarantees in the context of constrained Markov decision processes, they either lack scalability in solving larger problems or produce conservative plans. Recent advances in deep reinforcement learning improve scalability by learning policy networks as function approximators. In this paper, we propose an extension of soft actor critic model to estimate the execution risk of a plan through a risk critic and produce risk-bounded policies efficiently by adding an extra risk term in the loss function of the policy network. We define the execution risk in an accurate form, as opposed to approximating it through a summation of immediate risks at each time step that leads to conservative plans. Our proposed model is conditioned on a continuous spectrum of risk bounds, allowing the user to adjust the risk-averse level of the agent on the fly. Through a set of experiments, we show the advantage of our model in terms of both computational time and plan quality, compared to a state-of-the-art mathematical programming baseline, and validate its performance in more complicated scenarios, including nonlinear dynamics and larger state space.

Training-time safety violations have been a major concern when we deploy reinforcement learning algorithms in the real world. This paper explores the possibility of safe RL algorithms with zero training-time safety violations in the challenging setting where we are only given a safe but trivial-reward initial policy without any prior knowledge of the dynamics model and additional offline data. We propose an algorithm, Co-trained Barrier Certificate for Safe RL (CRABS), which iteratively learns barrier certificates, dynamics models, and policies. The barrier certificates, learned via adversarial training, ensure the policy's safety assuming calibrated learned dynamics model. We also add a regularization term to encourage larger certified regions to enable better exploration. Empirical simulations show that zero safety violations are already challenging for a suite of simple environments with only 2-4 dimensional state space, especially if high-reward policies have to visit regions near the safety boundary. Prior methods require hundreds of violations to achieve decent rewards on these tasks, whereas our proposed algorithms incur zero violations.

When one agent interacts with a multi-agent environment, it is challenging to deal with various opponents unseen before. Modeling the behaviors, goals, or beliefs of opponents could help the agent adjust its policy to adapt to different opponents. In addition, it is also important to consider opponents who are learning simultaneously or capable of reasoning. However, existing work usually tackles only one of the aforementioned types of opponent. In this paper, we propose model-based opponent modeling (MBOM), which employs the environment model to adapt to all kinds of opponent. MBOM simulates the recursive reasoning process in the environment model and imagines a set of improving opponent policies. To effectively and accurately represent the opponent policy, MBOM further mixes the imagined opponent policies according to the similarity with the real behaviors of opponents. Empirically, we show that MBOM achieves more effective adaptation than existing methods in competitive and cooperative environments, respectively with different types of opponent, i.e., fixed policy, na\"ive learner, and reasoning learner.