In this paper,we propose a novel task of Manipulation Question Answering(MQA),a class of Question Answering (QA) task, where the robot is required to find the answer to the question by actively interacting with the environment via manipulation. Considering the tabletop scenario, a heatmap of the scene is generated to facilitate the robot to have a semantic understanding of the scene and an imitation learning approach with semantic understanding metric is proposed to generate manipulation actions which guide the manipulator to explore the tabletop to find the answer to the question. Besides, a novel dataset which contains a variety of tabletop scenarios and corresponding question-answer pairs is established. Extensive experiments have been conducted to validate the effectiveness of the proposed framework.

Deep learning models on graphs have achieved remarkable performance in various graph analysis tasks, e.g., node classification, link prediction and graph clustering. However, they expose uncertainty and unreliability against the well-designed inputs, i.e., adversarial examples. Accordingly, various studies have emerged for both attack and defense addressed in different graph analysis tasks, leading to the arms race in graph adversarial learning. For instance, the attacker has poisoning and evasion attack, and the defense group correspondingly has preprocessing- and adversarial- based methods. Despite the booming works, there still lacks a unified problem definition and a comprehensive review. To bridge this gap, we investigate and summarize the existing works on graph adversarial learning tasks systemically. Specifically, we survey and unify the existing works w.r.t. attack and defense in graph analysis tasks, and give proper definitions and taxonomies at the same time. Besides, we emphasize the importance of related evaluation metrics, and investigate and summarize them comprehensively. Hopefully, our works can serve as a reference for the relevant researchers, thus providing assistance for their studies. More details of our works are available at https://github.com/gitgiter/Graph-Adversarial-Learning.

Reinforcement learning (RL) is a popular paradigm for addressing sequential decision tasks in which the agent has only limited environmental feedback. Despite many advances over the past three decades, learning in many domains still requires a large amount of interaction with the environment, which can be prohibitively expensive in realistic scenarios. To address this problem, transfer learning has been applied to reinforcement learning such that experience gained in one task can be leveraged when starting to learn the next, harder task. More recently, several lines of research have explored how tasks, or data samples themselves, can be sequenced into a curriculum for the purpose of learning a problem that may otherwise be too difficult to learn from scratch. In this article, we present a framework for curriculum learning (CL) in reinforcement learning, and use it to survey and classify existing CL methods in terms of their assumptions, capabilities, and goals. Finally, we use our framework to find open problems and suggest directions for future RL curriculum learning research.

We prove performance guarantees of two algorithms for approximating $Q^\star$ in batch reinforcement learning. Compared to classical iterative methods such as Fitted Q-Iteration---whose performance loss incurs quadratic dependence on horizon---these methods estimate (some forms of) the Bellman error and enjoy linear-in-horizon error propagation, a property established for the first time for algorithms that rely solely on batch data and output stationary policies. One of the algorithms uses a novel and explicit importance-weighting correction to overcome the infamous "double sampling" difficulty in Bellman error estimation, and does not use any squared losses. Our analyses reveal its distinct characteristics and potential advantages compared to classical algorithms.

Reinforcement learning has been at the center of some of the biggest artificial intelligence(AI) breakthroughs of the last five years. In mastering games like Go, Quake III or StarCraft, reinforcement learning models demonstrated that they can surpass human performance and create unique long-term strategies never explored before. Part of the magic of reinforcement learning relies on regularly rewarding the agents for actions that lead to a better outcome. That models works great in dense reward environments like games in which almost every action correspond to a specific feedback but what happens if that feedback is not available? In reinforcement learning this is known as sparse rewards environments and, unfortunately, it's a representation of most real-world scenarios.

In this paper, we propose a novel algorithm named STOchastic Recursive Momentum for Policy Gradient (STORM-PG), which operates a SARAH-type stochastic recursive variance-reduced policy gradient in an exponential moving average fashion. STORM-PG enjoys a provably sharp $O(1/\epsilon^3)$ sample complexity bound for STORM-PG, matching the best-known convergence rate for policy gradient algorithm. In the mean time, STORM-PG avoids the alternations between large batches and small batches which persists in comparable variance-reduced policy gradient methods, allowing considerably simpler parameter tuning. Numerical experiments depicts the superiority of our algorithm over comparative policy gradient algorithms.

Trust Region Policy Optimization (TRPO) and Proximal Policy Optimization (PPO) are among the most successful policy gradient approaches in deep reinforcement learning (RL). While these methods achieve state-of-the-art performance across a wide range of challenging tasks, there is room for improvement in the stabilization of the policy learning and how the off-policy data are used. In this paper we revisit the theoretical foundations of these algorithms and propose a new algorithm which stabilizes the policy improvement through a proximity term that constrains the discounted state-action visitation distribution induced by consecutive policies to be close to one another. This proximity term, expressed in terms of the divergence between the visitation distributions, is learned in an off-policy and adversarial manner. We empirically show that our proposed method can have a beneficial effect on stability and improve final performance in benchmark high-dimensional control tasks.

Despite the wealth of research into provably efficient reinforcement learning algorithms, most works focus on tabular representation and thus struggle to handle exponentially or infinitely large state-action spaces. In this paper, we consider episodic reinforcement learning with a continuous state-action space which is assumed to be equipped with a natural metric that characterizes the proximity between different states and actions. We propose ZoomRL, an online algorithm that leverages ideas from continuous bandits to learn an adaptive discretization of the joint space by zooming in more promising and frequently visited regions while carefully balancing the exploitation-exploration trade-off. We show that ZoomRL achieves a worst-case regret $\tilde{O}(H^{\frac{5}{2}} K^{\frac{d+1}{d+2}})$ where $H$ is the planning horizon, $K$ is the number of episodes and $d$ is the covering dimension of the space with respect to the metric. Moreover, our algorithm enjoys improved metric-dependent guarantees that reflect the geometry of the underlying space. Finally, we show that our algorithm is robust to small misspecification errors.

Traffic signal controllers play an essential role in the traffic system, while the current majority of them are not sufficiently flexible or adaptive to make optimal traffic schedules. In this paper we present an approach to learn policies for the signal controllers using deep reinforcement learning. Our method uses a novel formulation of the reward function that simultaneously considers efficiency and equity. We furthermore present a general approach to find the bound for the proposed equity factor. Moreover, we introduce the adaptive discounting approach that greatly stabilizes learning, which helps to keep high flexibility of green light duration. The experimental evaluations on both simulated and real-world data demonstrate that our proposed algorithm achieves state-of-the-art performance (previously held by traditional non-learning methods) on a wide range of traffic situations. A video of our experimental results can be found at: https://youtu.be/3rc5-ac3XX0

Exploration is a key problem in reinforcement learning. Recently bonus-based methods have achieved considerable successes in environments where exploration is difficult such as Montezuma's Revenge, which assign additional bonus (e.g., intrinsic reward) to guide the agent to rarely visited states. Since the bonus is calculated according to the novelty of the next state after performing an action, we call such methods the next-state bonus methods. However, the next-state bonus methods bring extra issues. It may lead agent to be trapped in states that fewer being visited and ignore to explore unknown states. Moreover, the behavior policy of the agent is also influenced by the bonus added to the state (or state-action) values indirectly. In contrast to the bonus-based methods which explore in known states, in this paper, we focus on the other part of exploration: exploration for finding unknown states. We propose the action balance exploration method to overcome the defects of the next-state bonus methods, which balances the chosen time of each action in each state and can be treated as an extension of upper confidence bound (UCB) to deep reinforcement learning. To take both the advantages of the next-state bonus method and our action balance exploration method, we propose the action balance RND method, which takes both parts of exploration into consideration. The experiments on grid world and Atari games demonstrate action balance exploration has a better capability in finding unknown states and can improve the real performance of RND in some hard exploration environments respectively.