Traditional model-based reinforcement learning approaches learn a model of the environment dynamics without explicitly considering how it will be used by the agent. In the presence of misspecified model classes, this can lead to poor estimates, as some relevant available information is ignored. In this paper, we introduce a novel model-based policy search approach that exploits the knowledge of the current agent policy to learn an approximate transition model, focusing on the portions of the environment that are most relevant for policy improvement. We leverage a weighting scheme, derived from the minimization of the error on the model-based policy gradient estimator, in order to define a suitable objective function that is optimized for learning the approximate transition model. Then, we integrate this procedure into a batch policy improvement algorithm, named Gradient-Aware Model-based Policy Search (GAMPS), which iteratively learns a transition model and uses it, together with the collected trajectories, to compute the new policy parameters. Finally, we empirically validate GAMPS on benchmark domains analyzing and discussing its properties.

We study the problem of identifying the policy space of a learning agent, having access to a set of demonstrations generated by its optimal policy. We introduce an approach based on statistical testing to identify the set of policy parameters the agent can control, within a larger parametric policy space. After presenting two identification rules (combinatorial and simplified), applicable under different assumptions on the policy space, we provide a probabilistic analysis of the simplified one in the case of linear policies belonging to the exponential family. To improve the performance of our identification rules, we frame the problem in the recently introduced framework of the Configurable Markov Decision Processes, exploiting the opportunity of configuring the environment to induce the agent revealing which parameters it can control. Finally, we provide an empirical evaluation, on both discrete and continuous domains, to prove the effectiveness of our identification rules.

We explore fixed-horizon temporal difference (TD) methods, reinforcement learning algorithms for a new kind of value function that predicts the sum of rewards over a $\textit{fixed}$ number of future time steps. To learn the value function for horizon $h$, these algorithms bootstrap from the value function for horizon $h-1$, or some shorter horizon. Because no value function bootstraps from itself, fixed-horizon methods are immune to the stability problems that plague other off-policy TD methods using function approximation (also known as "the deadly triad"). Although fixed-horizon methods require the storage of additional value functions, this gives the agent additional predictive power, while the added complexity can be substantially reduced via parallel updates, shared weights, and $n$-step bootstrapping. We show how to use fixed-horizon value functions to solve reinforcement learning problems competitively with methods such as Q-learning that learn conventional value functions. We also prove convergence of fixed-horizon temporal difference methods with linear and general function approximation. Taken together, our results establish fixed-horizon TD methods as a viable new way of avoiding the stability problems of the deadly triad.

We consider the Markov Decision Process (MDP) of selecting a subset of items at each step, termed the Select-MDP (S-MDP). The large state and action spaces of S-MDPs make them intractable to solve with typical reinforcement learning (RL) algorithms especially when the number of items is huge. In this paper, we present a deep RL algorithm to solve this issue by adopting the following key ideas. First, we convert the original S-MDP into an Iterative Select-MDP (IS-MDP), which is equivalent to the S-MDP in terms of optimal actions. IS-MDP decomposes a joint action of selecting K items simultaneously into K iterative selections resulting in the decrease of actions at the expense of an exponential increase of states. Second, we overcome this state space explo-sion by exploiting a special symmetry in IS-MDPs with novel weight shared Q-networks, which prov-ably maintain sufficient expressive power. Various experiments demonstrate that our approach works well even when the item space is large and that it scales to environments with item spaces different from those used in training.

This episode gives a general introduction into the field of Reinforcement Learning: - High level description of the field - Policy gradients - Biggest challenges (sparse rewards, reward shaping, ...) This video forms the basis for a series on RL where I will dive much deeper into technical details of state-of-the-art methods for RL. Links: - "Pong from Pixels - Karpathy": http://karpathy.github.io/2016/05/31/rl/ - Concept networks for grasp & stack (Paper with heavy reward shaping): https://arxiv.org/abs/1709.06977 If you enjoy my videos, all support is super welcome!

With the recent evolution of mobile health technologies, health scientists are increasingly interested in delivering interventions via notifications on mobile device at the moments when they can most readily help the user prevent negative health outcomes and promote the adoption and maintenance of healthy behaviors. The type and timing of the mobile health interventions should ideally adapt to the real-time collected user's context, e.g., the time of the day, the location, current activity and stress level. This gives rise to the concept of a justin-time adaptive intervention (JITAI) [28]. Operationally, JITAI includes a sequence of decision rules (e.g., treatment policy) that takes the user's current context as input and specifies whether and what type of an intervention should be provided at the moment. In practice, behavioral theory along with expert opinion and analyses of existing data is often used to design the decision rules. However, these theories are often insufficiently mature to precisely specify which particular intervention and when it should be delivered in order to ensure the interventions have the intended effects and optimize the long-term efficacy of the interventions. As a result, there is much interest in how best to use data to inform the design of JITAIs [12, 39, 3, 35, 26, 41, 33, 10, 34, 42] This paper develops a Reinforcement Learning (RL) algorithm to continuously learn, e.g., online, and optimize the treatment policy in the JITAI as the user experiences the intervention.

Task-specific scores are often used to optimize for and evaluate the performance of conditional text generation systems. However, such scores are non-differentiable and cannot be used in the standard supervised learning paradigm. Hence, policy gradient methods are used since the gradient can be computed without requiring a differentiable objective. However, we argue that current n-gram overlap based measures that are used as rewards can be improved by using model-based rewards transferred from tasks that directly compare the similarity of sentence pairs. These reward models either output a score of sentence-level syntactic and semantic similarity between entire predicted and target sentences as the expected return, or for intermediate phrases as segmented accumulative rewards. We demonstrate that using a \textit{Transferable Reward Learner} leads to improved results on semantical evaluation measures in policy-gradient models for image captioning tasks. Our InferSent actor-critic model improves over a BLEU trained actor-critic model on MSCOCO when evaluated on a Word Mover's Distance similarity measure by 6.97 points, also improving on a Sliding Window Cosine Similarity measure by 10.48 points. Similar performance improvements are also obtained on the smaller Flickr-30k dataset, demonstrating the general applicability of the proposed transfer learning method.

The neural architecture search (NAS) algorithm with reinforcement learning can be a powerful and novel framework for the automatic discovering process of neural architectures. However, its application is restricted by noncontinuous and high-dimensional search spaces, which result in difficulty in optimization. To resolve these problems, we proposed NAS in embedding space (NASES), which is a novel framework. Unlike other NAS with reinforcement learning approaches that search over a discrete and high-dimensional architecture space, this approach enables reinforcement learning to search in an embedding space by using architecture encoders and decoders. The current experiment demonstrated that the performance of the final architecture network using the NASES procedure is comparable with that of other popular NAS approaches for the image classification task on CIFAR-10. The beneficial-performance and effectiveness of NASES was impressive even when only the architecture-embedding searching and pre-training controller were applied without other NAS tricks such as parameter sharing. Specifically, considerable reduction in searches was achieved by reducing the average number of searching to 100 architectures to achieve a final architecture for the NASES procedure. Introduction Deep neural networks have enabled advances in image recognition, sequential pattern recognition, recommendation systems, and various tasks in the past decades.

In this paper, we consider the problem of online learning of Markov decision processes (MDPs) with very large state spaces. Under the assumptions of realizable function approximation and low Bellman ranks, we develop an online learning algorithm that learns the optimal value function while at the same time achieving very low cumulative regret during the learning process. Our learning algorithm, Adaptive Value-function Elimination (AVE), is inspired by the policy elimination algorithm proposed in (Jiang et al., 2017), known as OLIVE. One of our key technical contributions in AVE is to formulate the elimination steps in OLIVE as contextual bandit problems. This technique enables us to apply the active elimination and expert weighting methods from (Dudik et al., 2011), instead of the random action exploration scheme used in the original OLIVE algorithm, for more efficient exploration and better control of the regret incurred in each policy elimination step. To the best of our knowledge, this is the first $\sqrt{n}$-regret result for reinforcement learning in stochastic MDPs with general value function approximation.

This paper considers the problem of self-driving algorithm based on deep learning. This is a hot topic because self-driving is the most important application field of artificial intelligence. Existing work focused on deep learning which has the ability to learn end-to-end self-driving control directly from raw sensory data, but this method is just a mapping between images and driving. We prefer deep reinforcement learning to train a self-driving car in a virtual simulation environment created by Unity and then migrate to reality. Deep reinforcement learning makes the machine own the driving descision-making ability like human. The virtual to realistic training method can efficiently handle the problem that reinforcement learning requires reward from the environment which probably cause cars damge. We have derived a theoretical model and analysis on how to use Deep Q-learning to control a car to drive. We have carried out simulations in the Unity virtual environment for evaluating the performance. Finally, we successfully migrate te model to the real world and realize self-driving.