This paper adds to the growing literature of reinforcement learning (RL) for healthcare by proposing a novel paradigm: augmenting any predictor with Rule-based RL Layer (RRLL) that corrects the model's physiologically impossible predictions. Specifically, RRLL takes as input states predicted labels and outputs corrected labels as actions. The reward of the state-action pair is evaluated by a set of general rules. RRLL is efficient, general and lightweight: it does not require heavy expert knowledge like prior work but only a set of impossible transitions. This set is much smaller than all possible transitions; yet it can effectively reduce physiologically impossible mistakes made by the state-of-the-art predictor models. We verify the utility of RRLL on a variety of important healthcare classification problems and observe significant improvements using the same setup, with only the domain-specific set of impossibility changed. In-depth analysis shows that RRLL indeed improves accuracy by effectively reducing the presence of physiologically impossible predictions.

Sparse continuous policies are distributions that can choose some actions at random yet keep strictly zero probability for the other actions, which are radically different from the Gaussian. They have important real-world implications, e.g. in modeling safety-critical tasks like medicine. The combination of offline reinforcement learning and sparse policies provides a novel paradigm that enables learning completely from logged datasets a safety-aware sparse policy. However, sparse policies can cause difficulty with the existing offline algorithms which require evaluating actions that fall outside of the current support. In this paper, we propose the first offline policy optimization algorithm that tackles this challenge: Fat-to-Thin Policy Optimization (FtTPO). Specifically, we maintain a fat (heavy-tailed) proposal policy that effectively learns from the dataset and injects knowledge to a thin (sparse) policy, which is responsible for interacting with the environment. We instantiate FtTPO with the general $q$-Gaussian family that encompasses both heavy-tailed and sparse policies and verify that it performs favorably in a safety-critical treatment simulation and the standard MuJoCo suite. Our code is available at \url{https://github.com/lingweizhu/fat2thin}.

Many policy optimization approaches in reinforcement learning incorporate a Kullback-Leilbler (KL) divergence to the previous policy, to prevent the policy from changing too quickly. This idea was initially proposed in a seminal paper on Conservative Policy Iteration, with approximations given by algorithms like TRPO and Munchausen Value Iteration (MVI). We continue this line of work by investigating a generalized KL divergence -- called the Tsallis KL divergence -- which use the $q$-logarithm in the definition. The approach is a strict generalization, as $q = 1$ corresponds to the standard KL divergence; $q > 1$ provides a range of new options. We characterize the types of policies learned under the Tsallis KL, and motivate when $q >1$ could be beneficial. To obtain a practical algorithm that incorporates Tsallis KL regularization, we extend MVI, which is one of the simplest approaches to incorporate KL regularization. We show that this generalized MVI($q$) obtains significant improvements over the standard MVI($q = 1$) across 35 Atari games.

Deep reinforcement learning with domain randomization learns a control policy in various simulations with randomized physical and sensor model parameters to become transferable to the real world in a zero-shot setting. However, a huge number of samples are often required to learn an effective policy when the range of randomized parameters is extensive due to the instability of policy updates. To alleviate this problem, we propose a sample-efficient method named cyclic policy distillation (CPD). CPD divides the range of randomized parameters into several small sub-domains and assigns a local policy to each one. Then local policies are learned while cyclically transitioning to sub-domains. CPD accelerates learning through knowledge transfer based on expected performance improvements. Finally, all of the learned local policies are distilled into a global policy for sim-to-real transfers. CPD's effectiveness and sample efficiency are demonstrated through simulations with four tasks (Pendulum from OpenAIGym and Pusher, Swimmer, and HalfCheetah from Mujoco), and a real-robot, ball-dispersal task. We published code and videos from our experiments at https://github.com/yuki-kadokawa/cyclic-policy-distillation.

Clinicians prescribe antibiotics by looking at the patient's health record with an experienced eye. However, the therapy might be rendered futile if the patient has drug resistance. Determining drug resistance requires time-consuming laboratory-level testing while applying clinicians' heuristics in an automated way is difficult due to the categorical or binary medical events that constitute health records. In this paper, we propose a novel framework for rapid clinical intervention by viewing health records as graphs whose nodes are mapped from medical events and edges as correspondence between events in given a time window. A novel graph-based model is then proposed to extract informative features and yield automated drug resistance analysis from those high-dimensional and scarce graphs. The proposed method integrates multi-task learning into a common feature extracting graph encoder for simultaneous analyses of multiple drugs as well as stabilizing learning. On a massive dataset comprising over 110,000 patients with urinary tract infections, we verify the proposed method is capable of attaining superior performance on the drug resistance prediction problem. Furthermore, automated drug recommendations resemblant to laboratory-level testing can also be made based on the model resistance analysis.

Cancer subtyping is crucial for understanding the nature of tumors and providing suitable therapy. However, existing labelling methods are medically controversial, and have driven the process of subtyping away from teaching signals. Moreover, cancer genetic expression profiles are high-dimensional, scarce, and have complicated dependence, thereby posing a serious challenge to existing subtyping models for outputting sensible clustering. In this study, we propose a novel clustering method for exploiting genetic expression profiles and distinguishing subtypes in an unsupervised manner. The proposed method adaptively learns categorical correspondence from latent representations of expression profiles to the subtypes output by the model. By maximizing the problem -- agnostic mutual information between input expression profiles and output subtypes, our method can automatically decide a suitable number of subtypes. Through experiments, we demonstrate that our proposed method can refine existing controversial labels, and, by further medical analysis, this refinement is proven to have a high correlation with cancer survival rates.

The oscillating performance of off-policy learning and persisting errors in the actor-critic (AC) setting call for algorithms that can conservatively learn to suit the stability-critical applications better. In this paper, we propose a novel off-policy AC algorithm cautious actor-critic (CAC). The name cautious comes from the doubly conservative nature that we exploit the classic policy interpolation from conservative policy iteration for the actor and the entropy-regularization of conservative value iteration for the critic. Our key observation is the entropy-regularized critic facilitates and simplifies the unwieldy interpolated actor update while still ensuring robust policy improvement. We compare CAC to state-of-the-art AC methods on a set of challenging continuous control problems and demonstrate that CAC achieves comparable performance while significantly stabilizes learning.

In this paper, we propose cautious policy programming (CPP), a novel value-based reinforcement learning (RL) algorithm that can ensure monotonic policy improvement during learning. Based on the nature of entropy-regularized RL, we derive a new entropy regularization-aware lower bound of policy improvement that only requires estimating the expected policy advantage function. CPP leverages this lower bound as a criterion for adjusting the degree of a policy update for alleviating policy oscillation. Different from similar algorithms that are mostly theory-oriented, we also propose a novel interpolation scheme that makes CPP better scale in high dimensional control problems. We demonstrate that the proposed algorithm can trade o? performance and stability in both didactic classic control problems and challenging high-dimensional Atari games.

Reinforcement Learning (RL) (Sutton and Barto 2018) has A significant factor causing the complexity might be its excessive recently achieved impressive successes in fields such as generality (Kakade and Langford 2002; Pirotta et al. robotic manipulation (OpenAI 2019), video game playing 2013); Those bounds do not focus on any particular class (Mnih et al. 2015) and the game of Go (Silver et al. 2016). of value-based RL algorithms. In this paper, in order to develop However, compared with supervised learning that has widerange more tractable bounds, we focus on an RL class known of practical applications, RL applications have primarily as entropy-regularized value-based methods (Azar, Gómez, been limited to casual game playing or laboratory and Kappen 2012; Fox, Pakman, and Tishby 2016; Haarnoja based robotics. A crucial reason for limiting applications et al. 2017, 2018), where the entropies of policies are introduced to these environments is that it is not guaranteed that the