In Reinforcement Learning (RL), agents are trained on a reward and punishment mechanism. The agent is rewarded for correct moves and punished for the wrong ones. In doing so, the agent tries to minimize wrong moves and maximize the right ones. Various papers have proposed Deep Reinforcement Learning for autonomous driving. In self-driving cars, there are various aspects to consider, such as speed limits at various places, drivable zones, avoiding collisions -- just to mention a few.

Deep reinforcement learning has a variety of different algorithms that solves many types of complex problems in various situations, one class of these algorithms is policy gradient (PG), which applies to a wide range of problems in both discrete and continuous action spaces, but applying it naively is inefficient, because of its poor sample complexity and high variance, which result in slower learning, to mitigate this we can use a baseline. The cause of the high variance problem is the reward scale, we think of policy gradient as it increases the probability of taking good actions and decreases it for bad actions, but mostly this is not the case, imagine a situation where the "good" episode return was 10 and the "bad" one was 5, then both probabilities of the actions in those episodes will be increased, which is not what we want, this problem is what baselines can solve. Mathematically, a baseline is a function when added to an expectation, does not change the expected value (or does not introduce bias), but at the same time, it can significantly affect the variance. Following this definition, we want a baseline for the policy gradient that can reduce its high variance and does not change its direction, a natural thing to do is to take the actions that are better than average, increase their probability, and decrease the probability of the actions that are worse than average, this is implemented by calculating the average reward over the trajectory and subtract it from the reward at the current timestep, this kind of baselines is called the average reward baseline. Now, we will show how baselines do not change the expected value, and we can choose any baselines we want.

For strategic problems, intelligent systems based on Deep Reinforcement Learning (DRL) have demonstrated an impressive ability to learn advanced solutions that can go far beyond human capabilities, especially when dealing with complex scenarios. While this creates new opportunities for the development of intelligent assistance systems with groundbreaking functionalities, applying this technology to real-world problems carries significant risks and therefore requires trust in their transparency and reliability. With superhuman strategies being non-intuitive and complex by definition and real-world scenarios prohibiting a reliable performance evaluation, the key components for trust in these systems are difficult to achieve. Explainable AI (XAI) has successfully increased transparency for modern AI systems through a variety of measures, however, XAI research has not yet provided approaches enabling domain level insights for expert users in strategic situations. In this paper, we discuss the existence of superhuman DRL-based strategies, their properties, the requirements and challenges for transforming them into real-world environments, and the implications for trust through explainability as a key technology.

Reinforcement learning is a powerful framework for robots to acquire skills from experience, but often requires a substantial amount of online data collection. As a result, it is difficult to collect sufficiently diverse experiences that are needed for robots to generalize broadly. Videos of humans, on the other hand, are a readily available source of broad and interesting experiences. In this paper, we consider the question: can we perform reinforcement learning directly on experience collected by humans? This problem is particularly difficult, as such videos are not annotated with actions and exhibit substantial visual domain shift relative to the robot's embodiment. To address these challenges, we propose a framework for reinforcement learning with videos (RLV). RLV learns a policy and value function using experience collected by humans in combination with data collected by robots. In our experiments, we find that RLV is able to leverage such videos to learn challenging vision-based skills with less than half as many samples as RL methods that learn from scratch.

In searching for a generalizable representation of temporally extended tasks, we spot two necessary constituents: the utility needs to be non-Markovian to transfer temporal relations invariant to a probability shift, the utility also needs to be lifted to abstract out specific grounding objects. In this work, we study learning such utility from human demonstrations. While inverse reinforcement learning (IRL) has been accepted as a general framework of utility learning, its fundamental formulation is one concrete Markov Decision Process. Thus the learned reward function does not specify the task independently of the environment. Going beyond that, we define a domain of generalization that spans a set of planning problems following a schema. We hence propose a new quest, Generalized Inverse Planning, for utility learning in this domain. We further outline a computational framework, Maximum Entropy Inverse Planning (MEIP), that learns non-Markovian utility and associated concepts in a generative manner. The learned utility and concepts form a task representation that generalizes regardless of probability shift or structural change. Seeing that the proposed generalization problem has not been widely studied yet, we carefully define an evaluation protocol, with which we illustrate the effectiveness of MEIP on two proof-of-concept domains and one challenging task: learning to fold from demonstrations.

Psychological research shows that enjoyment of many goods is subject to satiation, with enjoyment declining after repeated exposures to the same item. Nevertheless, proposed algorithms for powering recommender systems seldom model these dynamics, instead proceeding as though user preferences were fixed in time. In this work, we adopt a multi-armed bandit setup, modeling satiation dynamics as a time-invariant linear dynamical system. In our model, the expected rewards for each arm decline monotonically with consecutive exposures and rebound towards the initial reward whenever that arm is not pulled. We analyze this model, showing that, when the arms exhibit deterministic identical dynamics, our problem is equivalent to a specific instance of Max K-Cut. In this case, a greedy policy, which plays the arms in a cyclic order, is optimal. In the general setting, where each arm's satiation dynamics are stochastic and governed by different (unknown) parameters, we propose an algorithm that first uses offline data to estimate each arm's reward model and then plans using a generalization of the greedy policy.

This article aims to combine factor investing and reinforcement learning (RL). The agent learns through sequential random allocations which rely on firms' characteristics. Using Dirichlet distributions as the driving policy, we derive closed forms for the policy gradients and analytical properties of the performance measure. This enables the implementation of REINFORCE methods, which we perform on a large dataset of US equities. Across a large range of implementation choices, our result indicates that RL-based portfolios are very close to the equally-weighted (1/N) allocation. This implies that the agent learns to be agnostic with regard to factors. This is partly consistent with cross-sectional regressions showing a strong time variation in the relationship between returns and firm characteristics.

This paper proves that the episodic learning environment of every finite-horizon decision task has a unique steady state under any behavior policy, and that the marginal distribution of the agent's input indeed approaches to the steady-state distribution in essentially all episodic learning processes. This observation supports an interestingly reversed mindset against conventional wisdom: While steady states are usually presumed to exist in continual learning and are considered less relevant in episodic learning, it turns out they are guaranteed to exist for the latter. Based on this insight, the paper further develops connections between episodic and continual RL for several important concepts that have been separately treated in the two RL formalisms. Practically, the existence of unique and approachable steady state enables a general, reliable, and efficient way to collect data in episodic RL tasks, which the paper applies to policy gradient algorithms as a demonstration, based on a new steady-state policy gradient theorem. The paper also proposes and empirically evaluates a perturbation method that facilitates rapid mixing in real-world tasks.

Active reinforcement learning (ARL) is a variant on reinforcement learning where the agent does not observe the reward unless it chooses to pay a query cost c > 0. The central question of ARL is how to quantify the long-term value of reward information. Even in multi-armed bandits, computing the value of this information is intractable and we have to rely on heuristics. We propose and evaluate several heuristic approaches for ARL in multi-armed bandits and (tabular) Markov decision processes, and discuss and illustrate some challenging aspects of the ARL problem.

Live fire creates a dynamic, rapidly changing environment that presents a worthy challenge for deep learning and artificial intelligence methodologies to assist firefighters with scene comprehension in maintaining their situational awareness, tracking and relay of important features necessary for key decisions as they tackle these catastrophic events. We propose a deep Q-learning based agent who is immune to stress induced disorientation and anxiety and thus able to make clear decisions for navigation based on the observed and stored facts in live fire environments. As a proof of concept, we imitate structural fire in a gaming engine called Unreal Engine which enables the interaction of the agent with the environment. The agent is trained with a deep Q-learning algorithm based on a set of rewards and penalties as per its actions on the environment. We exploit experience replay to accelerate the learning process and augment the learning of the agent with human-derived experiences. The agent trained under this deep Q-learning approach outperforms agents trained through alternative path planning systems and demonstrates this methodology as a promising foundation on which to build a path planning navigation assistant capable of safely guiding fire fighters through live fire environments.