Background Drug‐induced liver injury (DILI) is the most common reason for a drug to be withdrawn from the market. Apart from stopping the offending drug, no regimens are available for treating idiosyncratic DILI in clinical practice. Methods We carried out a randomized, double‐blind, multidoses, active drug controlled, multicentre phase II trial to assess the safety and efficacy of the study drug, magnesium isoglycyrrhizinate (MgIG), as compared to tiopronin, a standard therapy for DILI in China. The primary outcome was the proportion of alanine aminotransferase (ALT) normalization at week 4 after study drug administration. Logistic regression was used to examine the odds of ALT normalization between low dose (Group A) and high dose (Group B) vs active control (Group C). Results One hundred and seventy‐four eligible subjects were randomized and enrolled into three groups: 59 in group A, 56 in group B and 59 in group C. It was shown that group A and group B lowered ALT level even at early stage of study drug administration; when compared with Group C (61.02%), the proportions of ALT normalization at week 4 were significantly greater in Group A (84.75%, P = .0029) and Group B (85.71%, P = .0037) respectively. The results from the univariate logistic model showed that the odds of ALT normalized among subjects in Group A were about 3.6 times greater (OR = 3.55, 95% CI: 1.47‐8.57, P = .0049) than subjects in Group C. Similar effect was observed among subjects in Group B (OR = 3.83, 95% CI: 1.54‐9.55, P = .0039). Conclusions This trial provided preliminary evidence that MgIG is an effective and safe treatment for patients with acute DILI.

The Persistent Monitoring (PM) problem seeks to find a set of trajectories (or controllers) for robots to persistently monitor a changing environment. Each robot has a limited fieldof-view and may need to coordinate with others to ensure no point in the environment is left unmonitored for long periods of time. We model the problem such that there is a penalty that accrues every time step if a point is left unmonitored. However, the dynamics of the penalty are unknown to us. We present a Multi-Agent Reinforcement Learning (MARL) algorithm for the persistent monitoring problem. Specifically, we present a Multi-Agent Graph Attention Proximal Policy Optimization (MA-G-PPO) algorithm that takes as input the local observations of all agents combined with a low resolution global map to learn a policy for each agent. The graph attention allows agents to share their information with others leading to an effective joint policy. Our main focus is to understand how effective MARL is for the PM problem. We investigate five research questions with this broader goal. We find that MA-G-PPO is able to learn a better policy than the non-RL baseline in most cases, the effectiveness depends on agents sharing information with each other, and the policy learnt shows emergent behavior for the agents.

Infection of Mycobacterium tuberculosis (MTB) and nontuberculous mycobacteria (NTM) challenges effective pulmonary infectious disease control. Current phenotypic and molecular assays could not comprehensively and accurately diagnose MTB, NTM, and drug resistance. Next-generation sequencing allows an “all-in-one” approach providing results on expected drug susceptibility testing (DST) and the genotype of NTM strains. In this study, targeted capture sequencing was used to analyze the genetic backgrounds of 4 MTB strains and 32 NTM pathogenic strains in 30 clinical samples, including 14 sputum specimens and 16 bronchoalveolar lavage fluid samples. Through comparing with other TB diagnostic tests, we proved that targeted capture sequencing could be used as a highly sensitive (91.3%) and accurate (83.3%) method to diagnose TB, as well as MGIT 960. Also, we identified 7 NTM strains in 11 patients; among them, seven patients were MTB/NTM co-affected, which indicated that it was a meaningful tool for the diagnosis and treatment of NTM infection diseases in clinic. However, based on a drug-resistant mutation library (1,325 drug resistance loci), only 9 drug resistance strains and 22 drug resistance loci were discovered, having considerable discordance with the drug-resistant results of MGIT 960. Our finding indicated that targeted capture sequencing approach was applicable for the comprehensive and accurate diagnosis of MTB and NTM. However, from data presented here, the DST results identified by next-generation sequencing (NGS) showed a relatively low consistency with MGIT 960, especially in sputum samples. Further work should be done to explore the reasons for low drug-resistance detection rate of NGS.

We study the problem of devising a closed-loop strategy to control the position of a robot that is tracking a possibly moving target. The robot is capable of obtaining noisy measurements of the target's position. The key idea in active target tracking is to choose control laws that drive the robot to measurement locations that will reduce the uncertainty in the target's position. The challenge is that measurement uncertainty often is a function of the (unknown) relative positions of the target and the robot. Consequently, a closed-loop control policy is desired which can map the current estimate of the target's position to an optimal control law for the robot.Our main contribution is to devise a closed-loop control policy for target tracking that plans for a sequence of control actions, instead of acting greedily. We consider scenarios where the noise in measurement is a function of the state of the target. We seek to minimize the maximum uncertainty (trace of the posterior covariance matrix) over all possible measurements. We exploit the structural properties of a Kalman Filter to build a policy tree that is orders of magnitude smaller than naive enumeration while still preserving optimality guarantees. We show how to obtain even more computational savings by relaxing the optimality guarantees. The resulting algorithms are evaluated through simulations.