Predictive power allocation is conceived for power-efficient video streaming over mobile networks using deep reinforcement learning. The goal is to minimize the accumulated energy consumption over a complete video streaming session for a mobile user under the quality of service constraint that avoids video playback interruptions. To handle the continuous state and action spaces, we resort to deep deterministic policy gradient (DDPG) algorithm for solving the formulated problem. In contrast to previous predictive resource policies that first predict future information with historical data and then optimize the policy based on the predicted information, the proposed policy operates in an online and end-to-end manner. By judiciously designing the action and state that only depend on slowly-varying average channel gains, the signaling overhead between the edge server and the base stations can be reduced, and the dynamics of the system can be learned effortlessly. To improve the robustness of streaming and accelerate learning, we further exploit the partially known dynamics of the system by integrating the concepts of safer layer, post-decision state, and virtual experience into the basic DDPG algorithm. Our simulation results show that the proposed polices converge to the optimal policy derived based on perfect prediction of the future large-scale channel gains and outperforms the first-predictthen-optimize policy in the presence of prediction errors. By harnessing the partially known model of the system dynamics, the convergence speed can be dramatically improved. I. INTRODUCTION Mobile video traffic is expected to account for more than 75% of the global mobile data by 2021, and video-on-demand (VoD) services represent the main contributor [2]. This paper was presented in part at IEEE Globecom 2019 [1]. To avoid video stalling for a user experiencing hostile channel conditions, a base station (BS) can increase its transmit power for ensuring that the video segment is downloaded before being played.

In this paper, we propose an autonomous UAV path planning framework using deep reinforcement learning approach. The objective is to employ a self-trained UAV as a flying mobile unit to reach spatially distributed moving or static targets in a given three dimensional urban area. In this approach, a Deep Deterministic Policy Gradient (DDPG) with continuous action space is designed to train the UAV to navigate through or over the obstacles to reach its assigned target. A customized reward function is developed to minimize the distance separating the UAV and its destination while penalizing collisions. Numerical simulations investigate the behavior of the UAV in learning the environment and autonomously determining trajectories for different selected scenarios.

The next generation of mobile robots needs to be socially-compliant to be accepted by humans. As simple as this task may seem, defining compliance formally is not trivial. Yet, classical reinforcement learning (RL) relies upon hard-coded reward signals. In this work, we go beyond this approach and provide the agent with intrinsic motivation using empowerment. Empowerment maximizes the influence of an agent on its near future and has been shown to be a good model for biological behaviors. It also has been used for artificial agents to learn complicated and generalized actions. Self-empowerment maximizes the influence of an agent on its future. On the contrary, our robot strives for the empowerment of people in its environment, so they are not disturbed by the robot when pursuing their goals. We show that our robot has a positive influence on humans, as it minimizes the travel time and distance of humans while moving efficiently to its own goal. The method can be used in any multi-agent system that requires a robot to solve a particular task involving humans interactions.

The book AI Crash Course by Hadelin de Ponteves contains a toolkit of four different AI models: Thompson Sampling, Q-Learning, Deep Q-Learning and Deep Convolutional Q-learning. It teaches the theory of these AI models and provides coding examples for solving industry cases based on these models. InfoQ readers can find an excerpt of AI Crash Course on the publisher's website. InfoQ interviewed Hadelin de Ponteves about using different AI models and how to develop AI skills. InfoQ: Why did you write this book?

In this paper, we propose a general theoretical method for analyzing the risk bound in the presence of adversaries. Specifically, we try to fit the adversarial learning problem into the minimax framework. We first show that the original adversarial learning problem can be transformed into a minimax statistical learning problem by introducing a transport map between distributions. Then, we prove a new risk bound for this minimax problem in terms of covering numbers under a weak version of Lipschitz condition. Our method can be applied to multi-class classification and popular loss functions including the hinge loss and ramp loss.

This paper investigates the idea of encoding object-centered representations in the design of the reward function and policy architectures of a language-guided reinforcement learning agent. This is done using a combination of object-wise permutation invariant networks inspired from Deep Sets and gated-attention mechanisms. In a 2D procedurally-generated world where agents targeting goals in natural language navigate and interact with objects, we show that these architectures demonstrate strong generalization capacities to out-of-distribution goals. We study the generalization to varying numbers of objects at test time and further extend the object-centered architectures to goals involving relational reasoning.

This paper focuses on finding reinforcement learning policies for control systems with hard state and action constraints. Despite its success in many domains, reinforcement learning is challenging to apply to problems with hard constraints, especially if both the state variables and actions are constrained. Previous works seeking to ensure constraint satisfaction, or safety, have focused on adding a projection step to a learned policy. Yet, this approach requires solving an optimization problem at every policy execution step, which can lead to significant computational costs. To tackle this problem, this paper proposes a new approach, termed Vertex Networks (VNs), with guarantees on safety during exploration and on learned control policies by incorporating the safety constraints into the policy network architecture. Leveraging the geometric property that all points within a convex set can be represented as the convex combination of its vertices, the proposed algorithm first learns the convex combination weights and then uses these weights along with the pre-calculated vertices to output an action. The output action is guaranteed to be safe by construction. Numerical examples illustrate that the proposed VN algorithm outperforms vanilla reinforcement learning in a variety of benchmark control tasks.

In many real world applications, reinforcement learning agents have to optimize multiple objectives while following certain rules or satisfying a list of constraints. Classical methods based on reward shaping, i.e. a weighted combination of different objectives in the reward signal, or Lagrangian methods, including constraints in the loss function, have no guarantees that the agent satisfies the constraints at all points in time and lack in interpretability. When a discrete policy is extracted from an action-value function, safe actions can be ensured by restricting the action space at maximization, but can lead to sub-optimal solutions among feasible alternatives. In this work, we propose Multi Time-scale Constrained DQN, a novel algorithm restricting the action space directly in the Q-update to learn the optimal Q-function for the constrained MDP and the corresponding safe policy. In addition to single-step constraints referring only to the next action, we introduce a formulation for approximate multi-step constraints under the current target policy based on truncated value-functions to enhance interpretability. We compare our algorithm to reward shaping and Lagrangian methods in the application of high-level decision making in autonomous driving, considering constraints for safety, keeping right and comfort. We train our agent in the open-source simulator SUMO and on the real HighD data set.

In decision-making for autonomous vehicles, we need to predict other vehicle's behaviors or learn their behavior implicitly using machine learning. However, often the predictions and learned models have errors or might be wrong altogether which can lead to dangerous situations. Therefore, decision-making algorithms should consider counterfactual reasoning such as: what would happen if the other agents will behave in a certain way? The approach we present in this paper is two-fold: First, during training, we randomly select behavior models from a behavior model pool and assign them to the other vehicles in the scenario, such as more passive or aggressive behavior models. Second, during the application, we derive several virtual worlds from the actual world that have the same initial state. In each of these worlds, we also assign behavior models from the behavior model pool to others. We then evolve these virtual worlds for a defined time-horizon. This enables us to apply counterfactual reasoning by asking what would happen if the actual world evolves as in the virtual world. In uncertain environments, this makes it possible to generate more probable risk estimates and, thus, to enable safer decision-making. We conduct studies using a lane-change scenario that shows the advantages of counterfactual reasoning using learned policies and virtual worlds to estimate their risk and performance.

At each time step, I need to give a score to a varying number of items n. Depending on the score some different things will happen to the i best and j worst items. I need to optimize the neural network that does the scoring through time.