Treatment policies learned via reinforcement learning (RL) from observational health data are sensitive to subtle choices in study design. We highlight a simple approach, trajectory inspection, to bring clinicians into an iterative design process for model-based RL studies. We inspect trajectories where the model recommends unexpectedly aggressive treatments or believes its recommendations would lead to much more positive outcomes. Then, we examine clinical trajectories simulated with the learned model and policy alongside the actual hospital course to uncover possible modeling issues. To demonstrate that this approach yields insights, we apply it to recent work on RL for inpatient sepsis management. We find that a design choice around maximum trajectory length leads to a model bias towards discharge, that the RL policy preference for high vasopressor doses may be linked to small sample sizes, and that the model has a clinically implausible expectation of discharge without weaning off vasopressors.

In order to make better use of deep reinforcement learning in the creation of sensing policies for resource-constrained IoT devices, we present and study a novel reward function based on the Fisher information value. This reward function enables IoT sensor devices to learn to spend available energy on measurements at otherwise unpredictable moments, while conserving energy at times when measurements would provide little new information. This is a highly general approach, which allows for a wide range of use cases without significant human design effort or hyper-parameter tuning. We illustrate the approach in a scenario of workplace noise monitoring, where results show that the learned behavior outperforms a uniform sampling strategy and comes close to a near-optimal oracle solution.

This paper studies learning logic rules for reasoning on knowledge graphs. Logic rules provide interpretable explanations when used for prediction as well as being able to generalize to other tasks, and hence are critical to learn. Existing methods either suffer from the problem of searching in a large search space (e.g., neural logic programming) or ineffective optimization due to sparse rewards (e.g., techniques based on reinforcement learning). To address these limitations, this paper proposes a probabilistic model called RNNLogic. RNNLogic treats logic rules as a latent variable, and simultaneously trains a rule generator as well as a reasoning predictor with logic rules. We develop an EM-based algorithm for optimization. In each iteration, the reasoning predictor is first updated to explore some generated logic rules for reasoning. Then in the E-step, we select a set of high-quality rules from all generated rules with both the rule generator and reasoning predictor via posterior inference; and in the M-step, the rule generator is updated with the rules selected in the E-step. Experiments on four datasets prove the effectiveness of RNNLogic.

Simulated environments with procedurally generated content have become popular benchmarks for testing systematic generalization of reinforcement learning agents. Every level in such an environment is algorithmically created, thereby exhibiting a unique configuration of underlying factors of variation, such as layout, positions of entities, asset appearances, or even the rules governing environment transitions. Fixed sets of training levels can be determined to aid comparison and reproducibility, and test levels can be held out to evaluate the generalization and robustness of agents. We introduce Prioritized Level Replay, a general framework for estimating the future learning potential of a level given the current state of the agent's policy. We find that temporal-difference (TD) errors, while previously used to selectively sample past transitions, also prove effective for scoring a level's future learning potential in generating entire episodes that an agent would experience when replaying it. We report significantly improved sample-efficiency and generalization on the majority of Procgen Benchmark environments as well as two challenging MiniGrid environments. Lastly, we present a qualitative analysis showing that Prioritized Level Replay induces an implicit curriculum, taking the agent gradually from easier to harder levels. Environments generated using procedural content generation (PCG) have garnered increasing interest in RL research, leading to a surge of PCG environments such as MiniGrid (Chevalier-Boisvert et al., 2018), the Obstacle Tower Challenge (Juliani et al., 2019), the Procgen Benchmark (Cobbe et al., 2019), and the NetHack Learning Environment (Küttler et al., 2020).

Reliable bipedal walking over complex terrain is a challenging problem, using a curriculum can help learning. Curriculum learning is the idea of starting with an achievable version of a task and increasing the difficulty as a success criteria is met. We propose a 3-stage curriculum to train Deep Reinforcement Learning policies for bipedal walking over various challenging terrains. In the first stage, the agent starts on an easy terrain and the terrain difficulty is gradually increased, while forces derived from a target policy are applied to the robot joints and the base. In the second stage, the guiding forces are gradually reduced to zero. Finally, in the third stage, random perturbations with increasing magnitude are applied to the robot base, so the robustness of the policies are improved. In simulation experiments, we show that our approach is effective in learning walking policies, separate from each other, for five terrain types: flat, hurdles, gaps, stairs, and steps. Moreover, we demonstrate that in the absence of human demonstrations, a simple hand designed walking trajectory is a sufficient prior to learn to traverse complex terrain types. In ablation studies, we show that taking out any one of the three stages of the curriculum degrades the learning performance.

Text-based games have emerged as an important test-bed for Reinforcement Learning (RL) research, requiring RL agents to combine grounded language understanding with sequential decision making. In this paper, we examine the problem of infusing RL agents with commonsense knowledge. Such knowledge would allow agents to efficiently act in the world by pruning out implausible actions, and to perform look-ahead planning to determine how current actions might affect future world states. We design a new text-based gaming environment called TextWorld Commonsense (TWC) for training and evaluating RL agents with a specific kind of commonsense knowledge about objects, their attributes, and affordances. We also introduce several baseline RL agents which track the sequential context and dynamically retrieve the relevant commonsense knowledge from ConceptNet. We show that agents which incorporate commonsense knowledge in TWC perform better, while acting more efficiently. We conduct user-studies to estimate human performance on TWC and show that there is ample room for future improvement.

Given a simple request (e.g., Put a washed apple in the kitchen fridge), humans can reason in purely abstract terms by imagining action sequences and scoring their likelihood of success, prototypicality, and efficiency, all without moving a muscle. Once we see the kitchen in question, we can update our abstract plans to fit the scene. Embodied agents require the same abilities, but existing work does not yet provide the infrastructure necessary for both reasoning abstractly and executing concretely. We address this limitation by introducing ALFWorld, a simulator that enables agents to learn abstract, text-based policies in TextWorld (C\^ot\'e et al., 2018) and then execute goals from the ALFRED benchmark (Shridhar et al., 2020) in a rich visual environment. ALFWorld enables the creation of a new BUTLER agent whose abstract knowledge, learned in TextWorld, corresponds directly to concrete, visually grounded actions. In turn, as we demonstrate empirically, this fosters better agent generalization than training only in the visually grounded environment. BUTLER's simple, modular design factors the problem to allow researchers to focus on models for improving every piece of the pipeline (language understanding, planning, navigation, visual scene understanding, and so forth).

A course that will help you implement reinforcement learning in your projects!! In the last few years, we heard about Google's AlphaGo defeating the GO champion; we heard that the latest AIs are now playing Super Mario or Dota2, or even AI-powered self-driving cars (Tesla) have started carrying passengers without human assistance. If all this sounds crazy, then brace yourself for the future because development in AI is increasing at a pace like never before. Reinforcement learning is one such development in AI that has opened a whole new world. To help you learn this concept, we are set to launch an entire curation dedicated to Reinforcement Learning.

In this paper, we revisit variational intrinsic control (VIC), an unsupervised reinforcement learning method for finding the largest set of intrinsic options available to an agent. In the original work by Gregor et al. (2016), two VIC algorithms were proposed: one that represents the options explicitly, and the other that does it implicitly. We show that the intrinsic reward used in the latter is subject to bias in stochastic environments, causing convergence to suboptimal solutions. To correct this behavior, we propose two methods respectively based on the transitional probability model and Gaussian Mixture Model. We substantiate our claims through rigorous mathematical derivations and experimental analyses.

In this paper, we propose new problem-independent lower bounds on the sample complexity and regret in episodic MDPs, with a particular focus on the non-stationary case in which the transition kernel is allowed to change in each stage of the episode. Our main contribution is a novel lower bound of $\Omega((H^3SA/\epsilon^2)\log(1/\delta))$ on the sample complexity of an $(\varepsilon,\delta)$-PAC algorithm for best policy identification in a non-stationary MDP. This lower bound relies on a construction of "hard MDPs" which is different from the ones previously used in the literature. Using this same class of MDPs, we also provide a rigorous proof of the $\Omega(\sqrt{H^3SAT})$ regret bound for non-stationary MDPs. Finally, we discuss connections to PAC-MDP lower bounds.