To meet the growing quest for enhanced network capacity, mobile network operators (MNOs) are deploying dense infrastructures of small cells. This, in turn, increases the power consumption of mobile networks, thus impacting the environment. As a result, we have seen a recent trend of powering mobile networks with harvested ambient energy to achieve both environmental and cost benefits. In this paper, we consider a network of virtualized small cells (vSCs) powered by energy harvesters and equipped with rechargeable batteries, which can opportunistically offload baseband (BB) functions to a grid-connected edge server depending on their energy availability. We formulate the corresponding grid energy and traffic drop rate minimization problem, and propose a distributed deep reinforcement learning (DDRL) solution. Coordination among vSCs is enabled via the exchange of battery state information. The evaluation of the network performance in terms of grid energy consumption and traffic drop rate confirms that enabling coordination among the vSCs via knowledge exchange achieves a performance close to the optimal. Numerical results also confirm that the proposed DDRL solution provides higher network performance, better adaptation to the changing environment, and higher cost savings with respect to a tabular multi-agent reinforcement learning (MRL) solution used as a benchmark.

Consider the following problem, which we term Offline Meta Reinforcement Learning (OMRL): given the complete training histories of $N$ conventional RL agents, trained on $N$ different tasks, design a learning agent that can quickly maximize reward in a new, unseen task from the same task distribution. In particular, while each conventional RL agent explored and exploited its own different task, the OMRL agent must identify regularities in the data that lead to effective exploration/exploitation in the unseen task. To solve OMRL, we take a Bayesian RL (BRL) view, and seek to learn a Bayes-optimal policy from the offline data. We extend the recently proposed VariBAD BRL algorithm to the off-policy setting, and demonstrate learning of Bayes-optimal exploration strategies from offline data using deep neural networks. Furthermore, when applied to the online meta-RL setting (agent simultaneously collects data and improves its meta-RL policy), our method is significantly more sample efficient than the conventional VariBAD.

Modern Reinforcement Learning (RL) algorithms promise to solve difficult motor control problems directly from raw sensory inputs. Their attraction is due in part to the fact that they can represent a general class of methods that allow to learn a solution with a reasonably set reward and minimal prior knowledge, even in situations where it is difficult or expensive for a human expert. For RL to truly make good on this promise, however, we need algorithms and learning setups that can work across a broad range of problems with minimal problem specific adjustments or engineering. In this paper, we study this idea of generality in the locomotion domain. We develop a learning framework that can learn sophisticated locomotion behavior for a wide spectrum of legged robots, such as bipeds, tripeds, quadrupeds and hexapods, including wheeled variants. Our learning framework relies on a data-efficient, off-policy multi-task RL algorithm and a small set of reward functions that are semantically identical across robots. To underline the general applicability of the method, we keep the hyper-parameter settings and reward definitions constant across experiments and rely exclusively on on-board sensing. For nine different types of robots, including a real-world quadruped robot, we demonstrate that the same algorithm can rapidly learn diverse and reusable locomotion skills without any platform specific adjustments or additional instrumentation of the learning setup.

Artificial touch would seem well-suited for Reinforcement Learning (RL), since both paradigms rely on interaction with an environment. Here we propose a new environment and set of tasks to encourage development of tactile reinforcement learning: learning to type on a braille keyboard. Four tasks are proposed, progressing in difficulty from arrow to alphabet keys and from discrete to continuous actions. A simulated counterpart is also constructed by sampling tactile data from the physical environment. Using state-of-the-art deep RL algorithms, we show that all of these tasks can be successfully learnt in simulation, and 3 out of 4 tasks can be learned on the real robot. A lack of sample efficiency currently makes the continuous alphabet task impractical on the robot. To the best of our knowledge, this work presents the first demonstration of successfully training deep RL agents in the real world using observations that exclusively consist of tactile images. To aid future research utilising this environment, the code for this project has been released along with designs of the braille keycaps for 3D printing and a guide for recreating the experiments. A brief video summary is also available at https://youtu.be/eNylCA2uE_E.

We seek to efficiently learn by leveraging shared structure between different tasks and environments. For example, cooking is similar in different kitchens, even though the ingredients may change location. In principle, meta-reinforcement learning approaches can exploit this shared structure, but in practice, they fail to adapt to new environments when adaptation requires targeted exploration (e.g., exploring the cabinets to find ingredients in a new kitchen). We show that existing approaches fail due to a chicken-and-egg problem: learning what to explore requires knowing what information is critical for solving the task, but learning to solve the task requires already gathering this information via exploration. For example, exploring to find the ingredients only helps a robot prepare a meal if it already knows how to cook, but the robot can only learn to cook if it already knows where the ingredients are. To address this, we propose a new exploration objective (DREAM), based on identifying key information in the environment, independent of how this information will exactly be used solve the task. By decoupling exploration from task execution, DREAM explores and consequently adapts to new environments, requiring no reward signal when the task is specified via an instruction. Empirically, DREAM scales to more complex problems, such as sparse-reward 3D visual navigation, while existing approaches fail from insufficient exploration.

Purpose: AI in radiology is hindered chiefly by: 1) Requiring large annotated data sets. 2) Non-generalizability that limits deployment to new scanners / institutions. And 3) Inadequate explainability and interpretability. We believe that reinforcement learning can address all three shortcomings, with robust and intuitive algorithms trainable on small datasets. To the best of our knowledge, reinforcement learning has not been directly applied to computer vision tasks for radiological images. In this proof-of-principle work, we train a deep reinforcement learning network to predict brain tumor location. Materials and Methods: Using the BraTS brain tumor imaging database, we trained a deep Q network on 70 post-contrast T1-weighted 2D image slices. We did so in concert with image exploration, with rewards and punishments designed to localize lesions. To compare with supervised deep learning, we trained a keypoint detection convolutional neural network on the same 70 images. We applied both approaches to a separate 30 image testing set. Results: Reinforcement learning predictions consistently improved during training, whereas those of supervised deep learning quickly diverged. Reinforcement learning predicted testing set lesion locations with 85% accuracy, compared to roughly 7% accuracy for the supervised deep network. Conclusion: Reinforcement learning predicted lesions with high accuracy, which is unprecedented for such a small training set. We believe that reinforcement learning can propel radiology AI well past the inherent limitations of supervised deep learning, with more clinician-driven research and finally toward true clinical applicability.

This note aims to provide a basic intuition on the concept of filtrations as used in the context of reinforcement learning (RL). Filtrations are often used to formally define RL problems, yet their implications might not be eminent for those without a background in measure theory. Essentially, a filtration is a construct that captures partial knowledge up to time $t$, without revealing any future information that has already been simulated, yet not revealed to the decision-maker. We illustrate this with simple examples from the finance domain on both discrete and continuous outcome spaces. Furthermore, we show that the notion of filtration is not needed, as basing decisions solely on the current problem state (which is possible due to the Markovian property) suffices to eliminate future knowledge from the decision-making process.

Deep Q-Network (DQN) based multi-agent systems (MAS) for reinforcement learning (RL) use various schemes where in the agents have to learn and communicate. The learning is however specific to each agent and communication may be satisfactorily designed for the agents. As more complex Deep Q-Networks come to the fore, the overall complexity of the multi-agent system increases leading to issues like difficulty in training, need for higher resources and more training time, difficulty in fine-tuning, etc. To address these issues we propose a simple but efficient DQN based MAS for RL which uses shared state and rewards, but agent-specific actions, for updation of the experience replay pool of the DQNs, where each agent is a DQN. The benefits of the approach are overall simplicity, faster convergence and better performance as compared to conventional DQN based approaches. It should be noted that the method can be extended to any DQN. As such we use simple DQN and DDQN (Double Q-learning) respectively on three separate tasks i.e. Cartpole-v1 (OpenAI Gym environment), LunarLander-v2 (OpenAI Gym environment) and Maze Traversal (customized environment). The proposed approach outperforms the baseline on these tasks by decent margins respectively.

We aim to help users estimate the state of the world in tasks like robotic teleoperation and navigation with visual impairments, where users may have systematic biases that lead to suboptimal behavior: they might struggle to process observations from multiple sensors simultaneously, receive delayed observations, or overestimate distances to obstacles. While we cannot directly change the user's internal beliefs or their internal state estimation process, our insight is that we can still assist them by modifying the user's observations. Instead of showing the user their true observations, we synthesize new observations that lead to more accurate internal state estimates when processed by the user. We refer to this method as assistive state estimation (ASE): an automated assistant uses the true observations to infer the state of the world, then generates a modified observation for the user to consume (e.g., through an augmented reality interface), and optimizes the modification to induce the user's new beliefs to match the assistant's current beliefs. We evaluate ASE in a user study with 12 participants who each perform four tasks: two tasks with known user biases -- bandwidth-limited image classification and a driving video game with observation delay -- and two with unknown biases that our method has to learn -- guided 2D navigation and a lunar lander teleoperation video game. A different assistance strategy emerges in each domain, such as quickly revealing informative pixels to speed up image classification, using a dynamics model to undo observation delay in driving, identifying nearby landmarks for navigation, and exaggerating a visual indicator of tilt in the lander game. The results show that ASE substantially improves the task performance of users with bandwidth constraints, observation delay, and other unknown biases.

In the paper, we propose a class of efficient momentum-based policy gradient methods for the model-free reinforcement learning, which use adaptive learning rates and do not require any large batches. Specifically, we propose a fast important-sampling momentum-based policy gradient (IS-MBPG) method based on a new momentum-based variance reduced technique and the importance sampling technique. We also propose a fast Hessian-aided momentum-based policy gradient (HA-MBPG) method based on the momentum-based variance reduced technique and the Hessian-aided technique. Moreover, we prove that both the IS-MBPG and HA-MBPG methods reach the best known sample complexity of $O(\epsilon^{-3})$ for finding an $\epsilon$-stationary point of the non-concave performance function, which only require one trajectory at each iteration. In particular, we present a non-adaptive version of IS-MBPG method, i.e., IS-MBPG*, which also reaches the best known sample complexity of $O(\epsilon^{-3})$ without any large batches. In the experiments, we apply four benchmark tasks to demonstrate the effectiveness of our algorithms.