Human is able to conduct 3D recognition by a limited number of haptic contacts between the target object and his/her fingers without seeing the object. This capability is defined as `haptic glance' in cognitive neuroscience. Most of the existing 3D recognition models were developed based on dense 3D data. Nonetheless, in many real-life use cases, where robots are used to collect 3D data by haptic exploration, only a limited number of 3D points could be collected. In this study, we thus focus on solving the intractable problem of how to obtain cognitively representative 3D key-points of a target object with limited interactions between the robot and the object. A novel reinforcement learning based framework is proposed, where the haptic exploration procedure (the agent iteratively predicts the next position for the robot to explore) is optimized simultaneously with the objective 3D recognition with actively collected 3D points. As the model is rewarded only when the 3D object is accurately recognized, it is driven to find the sparse yet efficient haptic-perceptual 3D representation of the object. Experimental results show that our proposed model outperforms the state of the art models.

Creating incentives for cooperation is a challenge in natural and artificial systems. One potential answer is reputation, whereby agents trade the immediate cost of cooperation for the future benefits of having a good reputation. Game theoretical models have shown that specific social norms can make cooperation stable, but how agents can independently learn to establish effective reputation mechanisms on their own is less understood. We use a simple model of reinforcement learning to show that reputation mechanisms generate two coordination problems: agents need to learn how to coordinate on the meaning of existing reputations and collectively agree on a social norm to assign reputations to others based on their behavior. These coordination problems exhibit multiple equilibria, some of which effectively establish cooperation. When we train agents with a standard Q-learning algorithm in an environment with the presence of reputation mechanisms, convergence to undesirable equilibria is widespread. We propose two mechanisms to alleviate this: (i) seeding a proportion of the system with fixed agents that steer others towards good equilibria; and (ii), intrinsic rewards based on the idea of introspection, i.e., augmenting agents' rewards by an amount proportionate to the performance of their own strategy against themselves. A combination of these simple mechanisms is successful in stabilizing cooperation, even in a fully decentralized version of the problem where agents learn to use and assign reputations simultaneously. We show how our results relate to the literature in Evolutionary Game Theory, and discuss implications for artificial, human and hybrid systems, where reputations can be used as a way to establish trust and cooperation.

People have made remarkable progress in game AIs, especially in domain of perfect information game. However, trick-taking poker game, as a popular form of imperfect information game, has been regarded as a challenge for a long time. Since trick-taking game requires high level of not only reasoning, but also inference to excel, it can be a new milestone for imperfect information game AI. We study Gongzhu, a trick-taking game analogous to, but slightly simpler than contract bridge. Nonetheless, the strategies of Gongzhu are complex enough for both human and computer players. We train a strong Gongzhu AI ScrofaZero from \textit{tabula rasa} by deep reinforcement learning, while few previous efforts on solving trick-taking poker game utilize the representation power of neural networks. Also, we introduce new techniques for imperfect information game including stratified sampling, importance weighting, integral over equivalent class, Bayesian inference, etc. Our AI can achieve human expert level performance. The methodologies in building our program can be easily transferred into a wide range of trick-taking games.

Although model-based and model-free approaches to learning the control of systems have achieved impressive results on standard benchmarks, generalization to task variations is still lacking. Recent results suggest that generalization for standard architectures improves only after obtaining exhaustive amounts of data. We give evidence that generalization capabilities are in many cases bottlenecked by the inability to generalize on the combinatorial aspects of the problem. Furthermore, we show that for a certain subclass of the MDP framework, this can be alleviated by neuro-algorithmic architectures. Many control problems require long-term planning that is hard to solve generically with neural networks alone. We introduce a neuro-algorithmic policy architecture consisting of a neural network and an embedded time-dependent shortest path solver. These policies can be trained end-to-end by blackbox differentiation. We show that this type of architecture generalizes well to unseen variations in the environment already after seeing a few examples.

We study reinforcement learning in an infinite-horizon average-reward setting with linear function approximation, where the transition probability function of the underlying Markov Decision Process (MDP) admits a linear form over a feature mapping of the current state, action, and next state. We propose a new algorithm UCRL2-VTR, which can be seen as an extension of the UCRL2 algorithm with linear function approximation. We show that UCRL2-VTR with Bernstein-type bonus can achieve a regret of $\tilde{O}(d\sqrt{DT})$, where $d$ is the dimension of the feature mapping, $T$ is the horizon, and $\sqrt{D}$ is the diameter of the MDP. We also prove a matching lower bound $\tilde{\Omega}(d\sqrt{DT})$, which suggests that the proposed UCRL2-VTR is minimax optimal up to logarithmic factors. To the best of our knowledge, our algorithm is the first nearly minimax optimal RL algorithm with function approximation in the infinite-horizon average-reward setting.

We study multi-marginal optimal transport, a generalization of optimal transport that allows us to define discrepancies between multiple measures. It provides a framework to solve multi-task learning problems and to perform barycentric averaging. However, multi-marginal distances between multiple measures are typically challenging to compute because they require estimating a transport plan with $N^P$ variables. In this paper, we address this issue in the following way: 1) we efficiently solve the one-dimensional multi-marginal Monge-Wasserstein problem for a classical cost function in closed form, and 2) we propose a higher-dimensional multi-marginal discrepancy via slicing and study its generalized metric properties. We show that computing the sliced multi-marginal discrepancy is massively scalable for a large number of probability measures with support as large as $10^7$ samples. Our approach can be applied to solving problems such as barycentric averaging, multi-task density estimation and multi-task reinforcement learning.

Training deep reinforcement learning agents on environments with multiple levels / scenes from the same task, has become essential for many applications aiming to achieve generalization and domain transfer from simulation to the real world. While such a strategy is helpful with generalization, the use of multiple scenes significantly increases the variance of samples collected for policy gradient computations. Current methods, effectively continue to view this collection of scenes as a single Markov decision process (MDP), and thus learn a scene-generic value function V(s). However, we argue that the sample variance for a multi-scene environment is best minimized by treating each scene as a distinct MDP, and then learning a joint value function V(s,M) dependent on both state s and MDP M. We further demonstrate that the true joint value function for a multi-scene environment, follows a multi-modal distribution which is not captured by traditional CNN / LSTM based critic networks. To this end, we propose a dynamic value estimation (DVE) technique, which approximates the true joint value function through a sparse attention mechanism over multiple value function hypothesis / modes. The resulting agent not only shows significant improvements in the final reward score across a range of OpenAI ProcGen environments, but also exhibits enhanced navigation efficiency and provides an implicit mechanism for unsupervised state-space skill decomposition.

The number of connected smart devices has been increasing exponentially for different Internet-of-Things (IoT) applications. Security has been a long run challenge in the IoT systems which has many attack vectors, security flaws and vulnerabilities. Securing billions of B connected devices in IoT is a must task to realize the full potential of IoT applications. Recently, researchers have proposed many security solutions for IoT. Machine learning has been proposed as one of the emerging solutions for IoT security and Reinforcement learning is gaining more popularity for securing IoT systems. Reinforcement learning, unlike other machine learning techniques, can learn the environment by having minimum information about the parameters to be learned. It solves the optimization problem by interacting with the environment adapting the parameters on the fly. In this paper, we present an comprehensive survey of different types of cyber-attacks against different IoT systems and then we present reinforcement learning and deep reinforcement learning based security solutions to combat those different types of attacks in different IoT systems. Furthermore, we present the Reinforcement learning for securing CPS systems (i.e., IoT with feedback and control) such as smart grid and smart transportation system. The recent important attacks and countermeasures using reinforcement learning B in IoT are also summarized in the form of tables. With this paper, readers can have a more thorough understanding of IoT security attacks and countermeasures using Reinforcement Learning, as well as research trends in this area.

Cyber Physical Systems (CPS) are characterized by their ability to integrate the physical and information or cyber worlds. Their deployment in critical infrastructure have demonstrated a potential to transform the world. However, harnessing this potential is limited by their critical nature and the far reaching effects of cyber attacks on human, infrastructure and the environment. An attraction for cyber concerns in CPS rises from the process of sending information from sensors to actuators over the wireless communication medium, thereby widening the attack surface. Traditionally, CPS security has been investigated from the perspective of preventing intruders from gaining access to the system using cryptography and other access control techniques. Most research work have therefore focused on the detection of attacks in CPS. However, in a world of increasing adversaries, it is becoming more difficult to totally prevent CPS from adversarial attacks, hence the need to focus on making CPS resilient. Resilient CPS are designed to withstand disruptions and remain functional despite the operation of adversaries. One of the dominant methodologies explored for building resilient CPS is dependent on machine learning (ML) algorithms. However, rising from recent research in adversarial ML, we posit that ML algorithms for securing CPS must themselves be resilient. This paper is therefore aimed at comprehensively surveying the interactions between resilient CPS using ML and resilient ML when applied in CPS. The paper concludes with a number of research trends and promising future research directions. Furthermore, with this paper, readers can have a thorough understanding of recent advances on ML-based security and securing ML for CPS and countermeasures, as well as research trends in this active research area.

We consider the problem of generalization in reinforcement learning where visual aspects of the observations might differ, e.g. when there are different backgrounds or change in contrast, brightness, etc. We assume that our agent has access to only a few of the MDPs from the MDP distribution during training. The performance of the agent is then reported on new unknown test domains drawn from the distribution (e.g. unseen backgrounds). For this "zero-shot RL" task, we enforce invariance of the learned representations to visual domains via a domain adversarial optimization process. We empirically show that this approach allows achieving a significant generalization improvement to new unseen domains.