Reducing unexpected urban traffic congestion caused by en-route events (e.g., road closures, car crashes, etc.) often requires fast and accurate reactions to choose the best-fit traffic signals. Traditional traffic light control systems, such as SCATS and SCOOT, are not efficient as their traffic data provided by induction loops has a low update frequency (i.e., longer than 1 minute). Moreover, the traffic light signal plans used by these systems are selected from a limited set of candidate plans pre-programmed prior to unexpected events' occurrence. Recent research demonstrates that camera-based traffic light systems controlled by deep reinforcement learning (DRL) algorithms are more effective in reducing traffic congestion, in which the cameras can provide high-frequency high-resolution traffic data. However, these systems are costly to deploy in big cities due to the excessive potential upgrades required to road infrastructure. In this paper, we argue that Unmanned Aerial Vehicles (UAVs) can play a crucial role in dealing with unexpected traffic congestion because UAVs with onboard cameras can be economically deployed when and where unexpected congestion occurs. Then, we propose a system called "AVARS" that explores the potential of using UAVs to reduce unexpected urban traffic congestion using DRL-based traffic light signal control. This approach is validated on a widely used open-source traffic simulator with practical UAV settings, including its traffic monitoring ranges and battery lifetime. Our simulation results show that AVARS can effectively recover the unexpected traffic congestion in Dublin, Ireland, back to its original un-congested level within the typical battery life duration of a UAV.

In this paper, we discuss the Dutch power market, which is comprised of a day-ahead market and an intraday balancing market that operates like an auction. Due to fluctuations in power supply and demand, there is often an imbalance that leads to different prices in the two markets, providing an opportunity for arbitrage. To address this issue, we restructure the problem and propose a collaborative dual-agent reinforcement learning approach for this bi-level simulation and optimization of European power arbitrage trading. We also introduce two new implementations designed to incorporate domain-specific knowledge by imitating the trading behaviours of power traders. By utilizing reward engineering to imitate domain expertise, we are able to reform the reward system for the RL agent, which improves convergence during training and enhances overall performance. Additionally, the tranching of orders increases bidding success rates and significantly boosts profit and loss (P&L). Our study demonstrates that by leveraging domain expertise in a general learning problem, the performance can be improved substantially, and the final integrated approach leads to a three-fold improvement in cumulative P&L compared to the original agent. Furthermore, our methodology outperforms the highest benchmark policy by around 50% while maintaining efficient computational performance.

Reinforcement learning (RL) is a powerful tool for solving complex decision-making problems, but its lack of transparency and interpretability has been a major challenge in domains where decisions have significant real-world consequences. In this paper, we propose a novel Advantage Actor-Critic with Reasoner (A2CR), which can be easily applied to Actor-Critic-based RL models and make them interpretable. A2CR consists of three interconnected networks: the Policy Network, the Value Network, and the Reasoner Network. By predefining and classifying the underlying purpose of the actor's actions, A2CR automatically generates a more comprehensive and interpretable paradigm for understanding the agent's decision-making process. It offers a range of functionalities such as purpose-based saliency, early failure detection, and model supervision, thereby promoting responsible and trustworthy RL. Evaluations conducted in action-rich Super Mario Bros environments yield intriguing findings: Reasoner-predicted label proportions decrease for ``Breakout" and increase for ``Hovering" as the exploration level of the RL algorithm intensifies. Additionally, purpose-based saliencies are more focused and comprehensible.

Maximizing influences in complex networks is a practically important but computationally challenging task for social network analysis, due to its NP- hard nature. Most current approximation or heuristic methods either require tremendous human design efforts or achieve unsatisfying balances between effectiveness and efficiency. Recent machine learning attempts only focus on speed but lack performance enhancement. In this paper, different from previous attempts, we propose an effective deep reinforcement learning model that achieves superior performances over traditional best influence maximization algorithms. Specifically, we design an end-to-end learning framework that combines graph neural network as the encoder and reinforcement learning as the decoder, named DREIM. Trough extensive training on small synthetic graphs, DREIM outperforms the state-of-the-art baseline methods on very large synthetic and real-world networks on solution quality, and we also empirically show its linear scalability with regard to the network size, which demonstrates its superiority in solving this problem.

We propose a framework for verifiable and compositional reinforcement learning (RL) in which a collection of RL subsystems, each of which learns to accomplish a separate subtask, are composed to achieve an overall task. The framework consists of a high-level model, represented as a parametric Markov decision process, which is used to plan and analyze compositions of subsystems, and of the collection of low-level subsystems themselves. The subsystems are implemented as deep RL agents operating under partial observability. By defining interfaces between the subsystems, the framework enables automatic decompositions of task specifications, e.g., reach a target set of states with a probability of at least 0.95, into individual subtask specifications, i.e. achieve the subsystem's exit conditions with at least some minimum probability, given that its entry conditions are met. This in turn allows for the independent training and testing of the subsystems. We present theoretical results guaranteeing that if each subsystem learns a policy satisfying its subtask specification, then their composition is guaranteed to satisfy the overall task specification. Conversely, if the subtask specifications cannot all be satisfied by the learned policies, we present a method, formulated as the problem of finding an optimal set of parameters in the high-level model, to automatically update the subtask specifications to account for the observed shortcomings. The result is an iterative procedure for defining subtask specifications, and for training the subsystems to meet them. Experimental results demonstrate the presented framework's novel capabilities in environments with both full and partial observability, discrete and continuous state and action spaces, as well as deterministic and stochastic dynamics.

We introduce the receding-horizon policy gradient (RHPG) algorithm, the first PG algorithm with provable global convergence in learning the optimal linear estimator designs, i.e., the Kalman filter (KF). Notably, the RHPG algorithm does not require any prior knowledge of the system for initialization and does not require the target system to be open-loop stable. The key of RHPG is that we integrate vanilla PG (or any other policy search directions) into a dynamic programming outer loop, which iteratively decomposes the infinite-horizon KF problem that is constrained and non-convex in the policy parameter into a sequence of static estimation problems that are unconstrained and strongly-convex, thus enabling global convergence. We further provide fine-grained analyses of the optimization landscape under RHPG and detail the convergence and sample complexity guarantees of the algorithm. This work serves as an initial attempt to develop reinforcement learning algorithms specifically for control applications with performance guarantees by utilizing classic control theory in both algorithmic design and theoretical analyses. Lastly, we validate our theories by deploying the RHPG algorithm to learn the Kalman filter design of a large-scale convection-diffusion model. We open-source the code repository at \url{https://github.com/xiangyuan-zhang/LearningKF}.

Robot manipulation is an important part of human-robot interaction technology. However, traditional pre-programmed methods can only accomplish simple and repetitive tasks. To enable effective communication between robots and humans, and to predict and adapt to uncertain environments, this paper reviews recent autonomous and adaptive learning in robotic manipulation algorithms. It includes typical applications and challenges of human-robot interaction, fundamental tasks of robot manipulation and one of the most widely used formulations of robot manipulation, Markov Decision Process. Recent research focusing on robot manipulation is mainly based on Reinforcement Learning and Imitation Learning. This review paper shows the importance of Deep Reinforcement Learning, which plays an important role in manipulating robots to complete complex tasks in disturbed and unfamiliar environments. With the introduction of Imitation Learning, it is possible for robot manipulation to get rid of reward function design and achieve a simple, stable and supervised learning process. This paper reviews and compares the main features and popular algorithms for both Reinforcement Learning and Imitation Learning.

Artificial intelligence has made remarkable progress in handling complex tasks, thanks to advances in hardware acceleration and machine learning algorithms. However, to acquire more accurate outcomes and solve more complex issues, algorithms should be trained with more data. Processing this huge amount of data could be time-consuming and require a great deal of computation. To address these issues, distributed machine learning has been proposed, which involves distributing the data and algorithm across several machines. There has been considerable effort put into developing distributed machine learning algorithms, and different methods have been proposed so far. We divide these algorithms in classification and clustering (traditional machine learning), deep learning and deep reinforcement learning groups. Distributed deep learning has gained more attention in recent years and most of the studies have focused on this approach. Therefore, we mostly concentrate on this category. Based on the investigation of the mentioned algorithms, we highlighted the limitations that should be addressed in future research. Keywords: Artificial intelligence, Machine learning, Distributed machine learning, Distributed deep learning, Ditributed reinforcement learning, Data-parallelism, Model-parallelism. Introduction Artificial intelligence (AI) is a rapidly developing field that uses knowledge to simulate human behaviors (1) and train computers to learn, make judgments, and make decisions similarly to humans (2, 3). In other words, AI involves developing techniques and algorithms that are capable of thinking, acting, and implementing tasks using protocols that are otherwise beyond human comprehension (4). Machine learning (ML) is a subset of AI that learns from historical data, without being explicitly programmed (5). ML algorithms can be used to analyze data and build data-driven systems, including classification, clustering, regression, association rule mining, and reinforcement learning (6, 7). Deep learning is a branch of machine learning that uses artificial neural networks to intelligently analyze large amounts of data (8, 9).

In this paper, we propose a novel model-based multi-agent reinforcement learning approach named Value Decomposition Framework with Disentangled World Model to address the challenge of achieving a common goal of multiple agents interacting in the same environment with reduced sample complexity. Due to scalability and non-stationarity problems posed by multi-agent systems, model-free methods rely on a considerable number of samples for training. In contrast, we use a modularized world model, composed of action-conditioned, action-free, and static branches, to unravel the environment dynamics and produce imagined outcomes based on past experience, without sampling directly from the real environment. We employ variational auto-encoders and variational graph auto-encoders to learn the latent representations for the world model, which is merged with a value-based framework to predict the joint action-value function and optimize the overall training objective. We present experimental results in Easy, Hard, and Super-Hard StarCraft II micro-management challenges to demonstrate that our method achieves high sample efficiency and exhibits superior performance in defeating the enemy armies compared to other baselines.

Deep reinforcement learning agents need to be trained over millions of episodes to decently solve navigation tasks grounded to instructions. Furthermore, their ability to generalize to novel combinations of instructions is unclear. Interestingly however, children can decompose language-based instructions and navigate to the referred object, even if they have not seen the combination of queries prior. Hence, we created three 3D environments to investigate how deep RL agents learn and compose color-shape based combinatorial instructions to solve novel combinations in a spatial navigation task. First, we explore if agents can perform compositional learning, and whether they can leverage on frozen text encoders (e.g. CLIP, BERT) to learn word combinations in fewer episodes. Next, we demonstrate that when agents are pretrained on the shape or color concepts separately, they show a 20 times decrease in training episodes needed to solve unseen combinations of instructions. Lastly, we show that agents pretrained on concept and compositional learning achieve significantly higher reward when evaluated zero-shot on novel color-shape1-shape2 visual object combinations. Overall, our results highlight the foundations needed to increase an agent's proficiency in composing word groups through reinforcement learning and its ability for zero-shot generalization to new combinations.