When applying offline reinforcement learning (RL) in healthcare scenarios, the out-of-distribution (OOD) issues pose significant risks, as inappropriate generalization beyond clinical expertise can result in potentially harmful recommendations. While existing methods like conservative Q-learning (CQL) attempt to address the OOD issue, their effectiveness is limited by only constraining action selection by suppressing uncertain actions. This action-only regularization imitates clinician actions that prioritize short-term rewards, but it fails to regulate downstream state trajectories, thereby limiting the discovery of improved long-term treatment strategies. To safely improve policy beyond clinician recommendations while ensuring that state-action trajectories remain in-distribution, we propose \textit{Offline Guarded Safe Reinforcement Learning} ($\mathsf{OGSRL}$), a theoretically grounded model-based offline RL framework. $\mathsf{OGSRL}$ introduces a novel dual constraint mechanism for improving policy with reliability and safety. First, the OOD guardian is established to specify clinically validated regions for safe policy exploration. By constraining optimization within these regions, it enables the reliable exploration of treatment strategies that outperform clinician behavior by leveraging the full patient state history, without drifting into unsupported state-action trajectories. Second, we introduce a safety cost constraint that encodes medical knowledge about physiological safety boundaries, providing domain-specific safeguards even in areas where training data might contain potentially unsafe interventions. Notably, we provide theoretical guarantees on safety and near-optimality: policies that satisfy these constraints remain in safe and reliable regions and achieve performance close to the best possible policy supported by the data. When evaluated on the MIMIC-III sepsis treatment dataset, $\mathsf{OGSRL}$ demonstrated significantly better OOD handling than baselines. $\mathsf{OGSRL}$ achieved a 78\% reduction in mortality estimates and a 51\% increase in reward compared to clinician decisions.

The event-triggered control strategy is capable of significantly reducing the number of control task executions while achieving desired control objectives, such as stability. In this article, we introduce a novel learning-based method for event-triggered model predictive control with initially unknown dynamics. The formulation of optimal control problems (OCPs) is based on predictive states derived from Gaussian process (GP) regression under terminal constraints. The event-triggered condition proposed in this article is derived from the recursive feasibility, so that the OCPs are solved only when an error between the predictive and the actual states exceeds a certain threshold. This article analyzes the convergence of the closed-loop system under the event-triggered condition, demonstrating that the system's state will enter the terminal set within a finite time, assuming small-enough uncertainty in the GP model. We validate this approach through a tracking control problem, illustrating its practical effectiveness.

In safety-critical engineering applications, such as robust prediction against adversarial noise, it is necessary to quantify neural networks' uncertainty. Interval neural networks (INNs) are effective models for uncertainty quantification, giving an interval of predictions instead of a single value for a corresponding input. This article formulates the problem of training an INN as a chance-constrained optimization problem. The optimal solution of the formulated chance-constrained optimization naturally forms an INN that gives the tightest interval of predictions with a required confidence level. Since the chance-constrained optimization problem is intractable, a sample-based continuous approximate method is used to obtain approximate solutions to the chance-constrained optimization problem. We prove the uniform convergence of the approximation, showing that it gives the optimal INN consistently with the original ones. Additionally, we investigate the reliability of the approximation with finite samples, giving the probability bound for violation with finite samples. Through a numerical example and an application case study of anomaly detection in wind power data, we evaluate the effectiveness of the proposed INN against existing approaches, including Bayesian neural networks, highlighting its capability to significantly improve the performance of applying INNs for regression and unsupervised anomaly detection.

This paper addresses the wind power curve's abnormal detection and regression problems in a unified way. We formulate a chance-constrained optimization problem to obtain an interval neural network (or a set-valued regression model) of the wind power curve. The obtained interval neural network has two properties: (1) the interval locks the area of the normal data, which can be used for abnormal detection; (2) the center of the interval is the fitted wind power curve. We propose a sample-based sigmoidal approximation-based method to solve the formulated chance-constrained optimization problem.

Safe reinforcement learning (RL) is a promising approach for many real-world decision-making problems where ensuring safety is a critical necessity. In safe RL research, while expected cumulative safety constraints (ECSCs) are typically the first choices, chance constraints are often more pragmatic for incorporating safety under uncertainties. This paper proposes a \textit{flipping-based policy} for Chance-Constrained Markov Decision Processes (CCMDPs). The flipping-based policy selects the next action by tossing a potentially distorted coin between two action candidates. The probability of the flip and the two action candidates vary depending on the state. We establish a Bellman equation for CCMDPs and further prove the existence of a flipping-based policy within the optimal solution sets. Since solving the problem with joint chance constraints is challenging in practice, we then prove that joint chance constraints can be approximated into Expected Cumulative Safety Constraints (ECSCs) and that there exists a flipping-based policy in the optimal solution sets for constrained MDPs with ECSCs. As a specific instance of practical implementations, we present a framework for adapting constrained policy optimization to train a flipping-based policy. This framework can be applied to other safe RL algorithms. We demonstrate that the flipping-based policy can improve the performance of the existing safe RL algorithms under the same limits of safety constraints on Safety Gym benchmarks.

Safe exploration is essential for the practical use of reinforcement learning (RL) in many real-world scenarios. In this paper, we present a generalized safe exploration (GSE) problem as a unified formulation of common safe exploration problems. We then propose a solution of the GSE problem in the form of a meta-algorithm for safe exploration, MASE, which combines an unconstrained RL algorithm with an uncertainty quantifier to guarantee safety in the current episode while properly penalizing unsafe explorations before actual safety violation to discourage them in future episodes. The advantage of MASE is that we can optimize a policy while guaranteeing with a high probability that no safety constraint will be violated under proper assumptions. Specifically, we present two variants of MASE with different constructions of the uncertainty quantifier: one based on generalized linear models with theoretical guarantees of safety and near-optimality, and another that combines a Gaussian process to ensure safety with a deep RL algorithm to maximize the reward. Finally, we demonstrate that our proposed algorithm achieves better performance than state-of-the-art algorithms on grid-world and Safety Gym benchmarks without violating any safety constraints, even during training.

This article introduces a neural approximation-based method for solving continuous optimization problems with probabilistic constraints. After reformulating the probabilistic constraints as the quantile function, a sample-based neural network model is used to approximate the quantile function. The statistical guarantees of the neural approximation are discussed by showing the convergence and feasibility analysis. Then, by introducing the neural approximation, a simulated annealing-based algorithm is revised to solve the probabilistic constrained programs. An interval predictor model (IPM) of wind power is investigated to validate the proposed method.

In this paper, a multispectral pansharpening based on high-pass modulation regression (HPMR) is proposed. Firstly, full-scale estimation is applied to improving the quality of the injection coefficient estimation. Then the injection coefficient is performed in the HPM which accomplishes the fusion by building a detailed relationship between the panchromatic (PAN) image and the multispectral (MS) image. Experiments on two data sets assessed both at reduced resolution and at full resolution show that the proposed method can acquire better performance than the state-of-the-art pansharpening methods.

With the help of the auxiliary information provided by the appropriate prior fine spectral image (PFSI) in the same region, spatiotemporal super-resolution mapping (SSM) shows greater potential and better performance than the traditional super-resolution mapping (SM) models based on only monotemporal image. However, the temporal dependence of the existing SSM models usually describes the relationship between the coarse fractional images from original coarse spectral image (OCSI) and the fine fractional images from the PFSI, and the scale of temporal dependence information is not accurate and rich due to the different scales and properties of two fractional images. In addition, the existing SSM models usually do not consider point spread function (PSF) effect, resulting in affecting the accuracy of mapping result. To resolve the abovementioned issues, this letter proposes a general SSM model based on fine and coarse scales temporal dependence (FCSTD) by considering PSF effect. The experimental results demonstrate that the proposed model produces better mapping results than the traditional SM models, as well as the SSM models.

Under the background of the rapid change of energy technology and the deep integration of artificial intelligence into the power system, it is of great significance to study the intelligent decision-making method of security-constrained unit commitment (SCUC) with high adaptability and high accuracy. Thus, in this article, an expanded sequence-to-sequence (E-Seq2Seq)-based data-driven SCUC expert system for dynamic multiple-sequence mapping samples is proposed. First, dynamic multiple-sequence mapping samples of SCUC are reconstructed by analyzing the input–output sequence characteristics. Then, an E-Seq2Seq approach with a multiple-encoder–decoder architecture and a fully connected extension layer is proposed. On this basis, the simple recurrent unit is introduced as a neuron of the E-Seq2Seq approach to construct deep learning models, and an intelligent data-driven expert system for SCUC is further developed. The proposed approach has been simulated on a typical IEEE 118-bus system and a practical system in Hunan province in China. The results indicate that the proposed approach could possess strong generality, high solution accuracy, and efficiency over traditional methods.

Elderly drivers often tend to disobey stop signs, and the number of vehicle accidents associated with this is increasing. The problem we address in this work is that of failure (or inability) of elderly drivers to identify potential conflicts with other road users at stop-sign intersections. The purpose of the study is to investigate the influence of deceleration control with brake hold on the driving habits of elderly drivers in potentially hazardous situations. This study proposes a driver assistance system with three functionalities: 1) information provision to warn drivers that they are approaching a stop-sign intersection; 2) deceleration control to stop the vehicle; and 3) brake hold to ensure that the vehicle has stopped completely. The timeline for a stop-sign intersection scenario is divided into pre- and post-vehicle-stop phases. The effectiveness of the proposed approach in the context of braking-assistance intervention was evaluated by conducting a public-road driving experiment involving 34 elderly drivers. It was observed that the participants could be guided to exhibit rule-following driver behavior voluntarily. The proposed system increased safety for elderly drivers, not only by avoiding stop-sign violations, but also by increasing the available time for a driver to search for hidden hazards in blind spots. We conclude that the braking-assistance intervention system is effective in helping drivers avoid failure or inability to search for potential conflict owing to stop-sign violations.

This paper proposes a statistical framework to assess the risk of passing a non-signalized intersection for vehicles. First, an intensity model of the near-accident event is established by regarding the near-accident event as a non-homogeneous Poisson process. The non-homogeneous Poisson process is defined on the sigma-algebra of the 2-dimension plane of vehicle velocity and distance to the intersection instead of in the time axis. On the other hand, the pedestrian intention is defined as a binary variable with 1 as passing through the crosswalk and 0 as stopping. Logistic function is applied to model the probability of pedestrian intention. The proposed statistical models are evaluated by the residual analysis-based model checking method. Besides, based on the two models, the pedestrian-aware risk model is established to give a predictive risk metric quantitatively when pedestrian appears.