In this paper, we introduce an actor-critic algorithm called Deep Value Model Predictive Control (DMPC), which combines model-based trajectory optimization with value function estimation. The DMPC actor is a Model Predictive Control (MPC) optimizer with an objective function defined in terms of a value function estimated by the critic. We show that our MPC actor is an importance sampler, which minimizes an upper bound of the cross-entropy to the state distribution of the optimal sampling policy. In our experiments with a Ballbot system, we show that our algorithm can work with sparse and binary reward signals to efficiently solve obstacle avoidance and target reaching tasks. Compared to previous work, we show that including the value function in the running cost of the trajectory optimizer speeds up the convergence. We also discuss the necessary strategies to robustify the algorithm in practice.

Healthcare data suffers from both noise and lack of ground truth. The cost of data increases as it is cleaned and annotated in healthcare. Unlike other data sets, medical data annotation, which is critical to accurate ground truth, requires medical domain expertise for a better patient outcome. In this work, we report on the use of reinforcement learning to mimic the decision making process of annotators for medical events, to automate annotation and labelling. The reinforcement agent learns to annotate alarm data based on annotations done by an expert. Our method shows promising results on medical alarm data sets. We trained DQN and A2C agents using the data from monitoring devices annotated by an expert. Initial results from these RL agents learning the expert annotation behavior are promising. The A2C agent performs better in terms of learning the sparse events in a given state, thereby choosing more right actions compared to DQN agent. To the best of our knowledge, this is the first reinforcement learning application for the automation of medical events annotation, which has far-reaching practical use.

-- When learning behavior, training data is often generated by the learner itself; this can result in unstable training dynamics, and this problem has particularly important applications in safety-sensitive real-world control tasks such as robotics. In this work, we propose a principled and model-agnostic approach to mitigate the issue of unstable learning dynamics by maintaining a history of a reinforcement learning agent over the course of training, and reverting to the parameters of a previous agent whenever performance significantly decreases. We develop techniques for evaluating this performance through statistical hypothesis testing of continued improvement, and evaluate them on a standard suite of challenging benchmark tasks involving continuous control of simulated robots. We show improvements over state-of- the-art reinforcement learning algorithms in performance and robustness to hyperparameters, outperforming DDPG in 5 out of 6 evaluation environments and showing no decrease in performance with TD3, which is known to be relatively stable. In this way, our approach takes an important step towards increasing data efficiency and stability in training for real-world robotic applications. Online behavior learning, typically in the form of deep reinforcement learning (RL), has demonstrated significant successes in recent years [1, 2, 3, 4].

Sample-efficient exploration is crucial not only for discovering rewarding experiences but also for adapting to environment changes in a task-agnostic fashion. A principled treatment of the problem of optimal input synthesis for system identification is provided within the framework of sequential Bayesian experimental design. In this paper, we present an effective trajectory-optimization-based approximate solution of this otherwise intractable problem that models optimal exploration in an unknown Markov decision process (MDP). By interleaving episodic exploration with Bayesian nonlinear system identification, our algorithm takes advantage of the inductive bias to explore in a directed manner, without assuming prior knowledge of the MDP. Empirical evaluations indicate a clear advantage of the proposed algorithm in terms of the rate of convergence and the final model fidelity when compared to intrinsic-motivation-based algorithms employing exploration bonuses such as prediction error and information gain. Moreover, our method maintains a computational advantage over a recent model-based active exploration (MAX) algorithm, by focusing on the information gain along trajectories instead of seeking a global exploration policy. A reference implementation of our algorithm and the conducted experiments is publicly available.

TorchBeast is a platform for reinforcement learning (RL) research in PyTorch. It implements a version of the popular IMPALA algorithm for fast, asynchronous, parallel training of RL agents. Additionally, TorchBeast has simplicity as an explicit design goal: We provide both a pure-Python implementation ("MonoBeast") as well as a multi-machine high-performance version ("PolyBeast"). In the latter, parts of the implementation are written in C++, but all parts pertaining to machine learning are kept in simple Python using PyTorch, with the environments provided using the OpenAI Gym interface. This enables researchers to conduct scalable RL research using TorchBeast without any programming knowledge beyond Python and PyTorch. In this paper, we describe the TorchBeast design principles and implementation and demonstrate that it performs on-par with IMPALA on Atari. TorchBeast is released as an open-source package under the Apache 2.0 license and is available at \url{https://github.com/facebookresearch/torchbeast}.

Current end-to-end deep Reinforcement Learning (RL) approaches require jointly learning perception, decision-making and low-level control from very sparse reward signals and high-dimensional inputs, with little capability of incorporating prior knowledge. This results in prohibitively long training times for use on real-world robotic tasks. Existing algorithms capable of extracting task-level representations from high-dimensional inputs, e.g. object detection, often produce outputs of varying lengths, restricting their use in RL methods due to the need for neural networks to have fixed length inputs. In this work, we propose a framework that combines deep sets encoding, which allows for variable-length abstract representations, with modular RL that utilizes these representations, decoupling high-level decision making from low-level control. We successfully demonstrate our approach on the robot manipulation task of object sorting, showing that this method can learn effective policies within mere minutes of highly simplified simulation. The learned policies can be directly deployed on a robot without further training, and generalize to variations of the task unseen during training.

We build a profitable electronic trading agent with Reinforcement Learning that places buy and sell orders in the stock market. An environment model is built only with historical observational data, and the RL agent learns the trading policy by interacting with the environment model instead of with the real-market to minimize the risk and potential monetary loss. Trained in unsupervised and self-supervised fashion, our environment model learned a temporal and causal representation of the market in latent space through deep neural networks. We demonstrate that the trading policy trained entirely within the environment model can be transferred back into the real market and maintain its profitability. We believe that this environment model can serve as a robust simulator that predicts market movement as well as trade impact for further studies.

The aim of the inverse chemical design is to develop new molecules with given optimized molecular properties or objectives. Recently, generative deep learning (DL) networks are considered as the state-of-the-art in inverse chemical design and have achieved early success in generating molecular structures with desired properties in the pharmaceutical and material chemistry fields. However, satisfying a large number (larger than 10 objectives) of molecular objectives is a limitation of current generative models. To improve the model's ability to handle a large number of molecule design objectives, we developed a Reinforcement Learning (RL) based generative framework to optimize chemical molecule generation. Our use of Curriculum Learning (CL) to fine-tune the pre-trained generative network allowed the model to satisfy up to 21 objectives and increase the generative network's robustness. The experiments show that the proposed multiple-objective RL-based generative model can correctly identify unknown molecules with an 83 to 100 percent success rate, compared to the baseline approach of 0 percent. Additionally, this proposed generative model is not limited to just chemistry research challenges; we anticipate that problems that utilize RL with multiple-objectives will benefit from this framework.

The Sampled Policy Gradient (SPG) algorithm is a new offline actor-critic variant that samples in the action space to approximate the policy gradient. It does so by using the critic to evaluate the sampled actions. SPG offers theoretical promise over similar algorithms such as DPG as it searches the action-Q-value space independently of the local gradient, enabling it to avoid local minima. This paper aims to compare SPG to two similar actor-critic algorithms, CACLA and DPG. The comparison is made across two different environments, two different network architectures, as well as training on on-policy transitions in contrast to using an experience buffer. Results seem to show that although SPG does often not perform the worst, it doesn't always match the performance of the best performing algorithm at a particular task. Further experiments are required to get a better estimate of the qualities of SPG.

The human brain often recalls past memories (seemingly) unprompted. As we go throughout our day, we have spontaneous flashes of memory from our lives. While this spontaneous conjuration of memories has long been of interest to neuroscientists, AI research company DeepMind recently published a paper detailing how an AI of theirs replicated this strange pattern of recall. The conjuration of memories in the brain, neural replay, is tightly linked with the hippocampus. The hippocampus is a seahorse-shaped formation in the brain that belongs to the limbic system, and it is associated with the formation of new memories, as well as the emotions that memories spark. Current theories on the role of the hippocampi (there is one in each hemisphere of the brain), state that different regions of the hippocampus are responsible for the handling of different types of memories.