Who this course is for: Anyone interested in artificial intelligence, data science, machine learning, deep learning and reinforcement learning.

At a 2017 O'Reilly AI conference, Andrew Ng ranked reinforcement learning dead last in terms of its utility for business applications. Compared to other machine learning methods like supervised learning, transfer learning, and even unsupervised learning, deep reinforcement learning (RL) is incredibly data hungry, often unstable, and rarely the best option in terms of performance. RL has historically been successfully applied only in arenas where mountains of simulated data can be generated on demand, such as games and robotics. Despite RL's limitations in solving business use cases, some AI experts believe this approach is the most viable strategy for achieving human or superhuman Artificial General Intelligence (AGI). The recent victory of DeepMind's AlphaStar over top-ranked professional StarCraft players suggests we might be on the cusp of applying deep RL to real world problems with real-time demands, extraordinary complexity, and incomplete information.

Recently, GAIL framework and various variants have shown remarkable possibilities for solving practical MDP problems. However, detailed researches of low-level, and high-dimensional state input in this framework, such as image sequences, has not been conducted. Furthermore, the cost function learned in the traditional GAIL frame-work only lies on a negative range, acting as a non-penalized reward and making the agent difficult to learn the optimal policy. In this paper, we propose a new algorithm based on the GAIL framework that includes a global encoder and the reward penalization mechanism. The global encoder solves two issues that arise when applying GAIL framework to high-dimensional image state. Also, it is shown that the penalization mechanism provides more adequate reward to the agent, resulting in stable performance improvement. Our approach's potential can be backed up by the fact that it is generally applicable to variants of GAIL framework. We conducted in-depth experiments by applying our methods to various variants of the GAIL framework. And, the results proved that our method significantly improves the performances when it comes to low-level and high-dimensional tasks.

The autonomous trading agent is one of the most actively studied areas of artificial intelligence to solve the capital market portfolio management problem. The two primary goals of the portfolio management problem are maximizing profit and restrainting risk. However, most approaches to this problem solely take account of maximizing returns. Therefore, this paper proposes a deep reinforcement learning based trading agent that can manage the portfolio considering not only profit maximization but also risk restraint. We also propose a new target policy to allow the trading agent to learn to prefer low-risk actions. The new target policy can be reflected in the update by adjusting the greediness for the optimal action through the hyper parameter. The proposed trading agent verifies the performance through the data of the cryptocurrency market. The Cryptocurrency market is the best test-ground for testing our trading agents because of the huge amount of data accumulated every minute and the market volatility is extremely large. As a experimental result, during the test period, our agents achieved a return of 1800% and provided the least risky investment strategy among the existing methods. And, another experiment shows that the agent can maintain robust generalized performance even if market volatility is large or training period is short.

Model-free deep reinforcement learning (RL) algorithms have been widely used for a range of complex control tasks. However, slow convergence and sample inefficiency remain challenging problems in RL, especially when handling continuous and high-dimensional state spaces. To tackle this problem, we propose a general acceleration method for model-free, off-policy deep RL algorithms by drawing the idea underlying regularized Anderson acceleration (RAA), which is an effective approach to accelerating the solving of fixed point problems with perturbations. Specifically, we first explain how policy iteration can be applied directly with Anderson acceleration. Then we extend RAA to the case of deep RL by introducing a regularization term to control the impact of perturbation induced by function approximation errors. We further propose two strategies, i.e., progressive update and adaptive restart, to enhance the performance. The effectiveness of our method is evaluated on a variety of benchmark tasks, including Atari 2600 and MuJoCo. Experimental results show that our approach substantially improves both the learning speed and final performance of state-of-the-art deep RL algorithms.

This paper investigates trajectory tracking problem for a class of underactuated autonomous underwater vehicles (AUVs) with unknown dynamics and constrained inputs. Different from existing policy gradient methods which employ single actor-critic but cannot realize satisfactory tracking control accuracy and stable learning, our proposed algorithm can achieve high-level tracking control accuracy of AUVs and stable learning by applying a hybrid actors-critics architecture, where multiple actors and critics are trained to learn a deterministic policy and action-value function, respectively. Specifically, for the critics, the expected absolute Bellman error based updating rule is used to choose the worst critic to be updated in each time step. Subsequently, to calculate the loss function with more accurate target value for the chosen critic, Pseudo Q-learning, which uses sub-greedy policy to replace the greedy policy in Q-learning, is developed for continuous action spaces, and Multi Pseudo Q-learning (MPQ) is proposed to reduce the overestimation of action-value function and to stabilize the learning. As for the actors, deterministic policy gradient is applied to update the weights, and the final learned policy is defined as the average of all actors to avoid large but bad updates. Moreover, the stability analysis of the learning is given qualitatively. The effectiveness and generality of the proposed MPQ-based Deterministic Policy Gradient (MPQ-DPG) algorithm are verified by the application on AUV with two different reference trajectories. And the results demonstrate high-level tracking control accuracy and stable learning of MPQ-DPG. Besides, the results also validate that increasing the number of the actors and critics will further improve the performance.

Maximum entropy deep reinforcement learning (RL) methods have been demonstrated on a range of challenging continuous tasks. However, existing methods either suffer from severe instability when training on large off-policy data or cannot scale to tasks with very high state and action dimensionality such as 3D humanoid locomotion. Besides, the optimality of desired Boltzmann policy set for non-optimal soft value function is not persuasive enough. In this paper, we first derive soft policy gradient based on entropy regularized expected reward objective for RL with continuous actions. Then, we present an off-policy actor-critic, model-free maximum entropy deep RL algorithm called deep soft policy gradient (DSPG) by combining soft policy gradient with soft Bellman equation. To ensure stable learning while eliminating the need of two separate critics for soft value functions, we leverage double sampling approach to making the soft Bellman equation tractable. The experimental results demonstrate that our method outperforms in performance over off-policy prior methods.

Recent research on reinforcement learning has shown that trained agents are vulnerable to maliciously crafted adversarial samples. In this work, we show how adversarial samples against RL agents can be generalised from White-box and Grey-box attacks to a strong Black-box case, namely where the attacker has no knowledge of the agents and their training methods. We use sequence-to-sequence models to predict a single action or a sequence of future actions that a trained agent will make. Our approximation model, based on time-series information from the agent, successfully predicts agents' future actions with consistently above 80% accuracy on a wide range of games and training methods. Second, we find that although such adversarial samples are transferable, they do not outperform random Gaussian noise as a means of reducing the game scores of trained RL agents. This highlights a serious methodological deficiency in previous work on such agents; random jamming should have been taken as the baseline for evaluation. Third, we do find a novel use for adversarial samples in this context: they can be used to trigger a trained agent to misbehave after a specific delay. This appears to be a genuinely new type of attack; it potentially enables an attacker to use devices controlled by RL agents as time bombs.

Full-sampling (e.g., Q-learning) and pure-expectation (e.g., Expected Sarsa) algorithms are efficient and frequently used techniques in reinforcement learning. Q$(\sigma,\lambda)$ is the first approach unifies them with eligibility trace through the sampling degree $\sigma$. However, it is limited to the tabular case, for large-scale learning, the Q$(\sigma,\lambda)$ is too expensive to require a huge volume of tables to accurately storage value functions. To address above problem, we propose a GQ$(\sigma,\lambda)$ that extends tabular Q$(\sigma,\lambda)$ with linear function approximation. We prove the convergence of GQ$(\sigma,\lambda)$. Empirical results on some standard domains show that GQ$(\sigma,\lambda)$ with a combination of full-sampling with pure-expectation reach a better performance than full-sampling and pure-expectation methods.

We present DRLViz, a visual analytics interface to interpret the internal memory of an agent (e.g. a robot) trained using deep reinforcement learning. This memory is composed of large temporal vectors updated when the agent moves in an environment and is not trivial to understand. It is often referred to as a black box as only inputs (images) and outputs (actions) are intelligible for humans. Using DRLViz, experts are assisted to interpret using memory reduction interactions, to investigate parts of the memory role when errors have been made, and ultimately to improve the agent training process. We report on several examples of use of DRLViz, in the context of video games simulators (ViZDoom) for a navigation scenario with item gathering tasks. We also report on experts evaluation using DRLViz, and applicability of DRLViz to other scenarios and navigation problems beyond simulation games, as well as its contribution to black box models interpret-ability and explain-ability in the field of visual analytics.