Responsible integration of deep neural networks into the design of trustworthy systems requires the ability to explain decisions made by these models. Explainability and transparency are critical for system analysis, certification, and human-machine teaming. We have recently demonstrated that neural stochastic differential equations (SDEs) present an explanation-friendly DNN architecture. In this paper, we present ExplainIt, an online tool for explaining AI decisions that uses neural SDEs to create visually sharper and more robust attributions than traditional residual neural networks. Our tool shows that the injection of noise in every layer of a residual network often leads to less noisy and less fragile integrated gradient attributions. The discrete neural stochastic differential equation model is trained on the ImageNet data set with a million images, and the demonstration produces robust attributions on images in the ImageNet validation library and on a variety of images in the wild. Our online tool is hosted publicly for educational purposes.

The adoption of health information technology (HIT) has facilitated efforts to increase the quality and efficiency of health care services and decrease health care overhead while simultaneously generating massive amounts of digital information stored in electronic health records (EHRs). However, due to patient safety issues resulting from the use of HIT systems, there is an emerging need to develop and implement hazard detection tools to identify and mitigate risks to patients. This paper presents a new methodological framework to develop hazard detection models and to demonstrate its capability by using the US Department of Veterans Affairs' (VA) Corporate Data Warehouse, the data repository for the VA's EHR. The overall purpose of the framework is to provide structure for research and communication about research results. One objective is to decrease the communication barriers between interdisciplinary research stakeholders and to provide structure for detecting hazards and risks to patient safety introduced by HIT systems through errors in the collection, transmission, use, and processing of data in the EHR, as well as potential programming or configuration errors in these HIT systems. A nine-stage framework was created, which comprises programs about feature extraction, detector development, and detector optimization, as well as a support environment for evaluating detector models. The framework forms the foundation for developing hazard detection tools and the foundation for adapting methods to particular HIT systems.

Over the past decade, health information technology (IT) has enabled the amount of digital information stored in electronic health records (EHRs) to expand greatly. However, according to some studies, hazards in health IT can lead to changes in clinical decisions, care processes, and care outcomes, as well as other issues. Thus, the effects of health IT hazards on patient safety have been at the forefront of recent patient safety research. Nonetheless, hazard detection in health IT remains a challenge. In this paper, the authors assume that safety-related issues in health IT would exhibit anomalous characteristics in EHR data. Although all hazards will exhibit some anomalous characteristics, not all anomalies can be regarded as hazards. The authors hypothesize that errors in health IT could lead to interruptions in the sequence of clinical actions. To this end, the problem of detecting anomalous sequences in big EHR data is considered. This paper focuses on dynamic event sequences, which are a series of clinical actions in motion. The authors propose an adaptive anomaly detection approach that uses higher-order network representation to detect anomalous sequences. Furthermore, the authors propose a contiguous subsequence anomaly detection approach that identifies abnormal subsequences in the detected anomalous sequences. The proposed approaches are tested by using synthetic and real-world EHR data. The proposed methods outperform existing state of the art anomaly detection techniques. To reduce the computational complexity associated with the operational implementation of the proposed approaches, the Apache Spark environment was leveraged, and a much shorter run time together with improved performance were achieved, especially for data with more than 60,000 sequences.

Given a contact network and coarse-grained diagnostic information like electronic Healthcare Reimbursement Claims (eHRC) data, can we develop efficient intervention policies to control an epidemic? Immunization is an important problem in multiple areas especially epidemiology and public health. However, most existing studies focus on developing pre-emptive strategies assuming prior epidemiological models. In practice, disease spread is usually complicated, hence assuming an underlying model may deviate from true spreading patterns, leading to possibly inaccurate interventions. Additionally, the abundance of health care surveillance data (like eHRC) makes it possible to study data-driven strategies without too many restrictive assumptions. Hence, such an approach can help public-health experts take more practical decisions. In this paper, we take into account propagation log and contact networks for controlling propagation. We formulate the novel and challenging Data-Driven Immunization problem without assuming classical epidemiological models. To solve it, we first propose an efficient sampling approach to align surveillance data with contact networks, then develop an efficient algorithm with the provably approximate guarantee for immunization. Finally, we show the effectiveness and scalability of our methods via extensive experiments on multiple datasets, and conduct case studies on nation-wide real medical surveillance data.

We examine the use of electronic healthcare reimbursement claims (EHRC) for analyzing healthcare delivery and practice patterns across the United States (US). We show that EHRCs are correlated with disease incidence estimates published by the Centers for Disease Control. Further, by analyzing over 1 billion EHRCs, we track patterns of clinical procedures administered to patients with autism spectrum disorder (ASD), heart disease (HD) and breast cancer (BC) using sequential pattern mining algorithms. Our analyses reveal that in contrast to treating HD and BC, clinical procedures for ASD diagnoses are highly varied leading up to and after the ASD diagnoses. The discovered clinical procedure sequences also reveal significant differences in the overall costs incurred across different parts of the US, indicating a lack of consensus amongst practitioners in treating ASD patients. We show that a data-driven approach to understand clinical trajectories using EHRC can provide quantitative insights into how to better manage and treat patients. Based on our experience, we also discuss emerging challenges in using EHRC datasets for gaining insights into the state of contemporary healthcare delivery and practice in the US.

BACKGROUND: The digitization of health-related information through electronic health records (EHR) and electronic healthcare reimbursement claims and the continued growth of self-reported health information through social media provides both tremendous opportunities and challenges in developing effective biosurveillance tools. With novel emerging infectious diseases being reported across different parts of the world, there is a need to build systems that can track, monitor and report such events in a timely manner. Further, it is also important to identify susceptible geographic regions and populations where emerging diseases may have a significant impact. METHODS: In this paper, we present an overview of Oak Ridge Biosurveillance Toolkit (ORBiT), which we have developed specifically to address data analytic challenges in the realm of public health surveillance. In particular, ORBiT provides an extensible environment to pull together diverse, large-scale datasets and analyze them to identify spatial and temporal patterns for various biosurveillance-related tasks. RESULTS: We demonstrate the utility of ORBiT in automatically extracting a small number of spatial and temporal patterns during the 2009-2010 pandemic H1N1 flu season using claims data. These patterns provide quantitative insights into the dynamics of how the pandemic flu spread across different parts of the country. We discovered that the claims data exhibits multi-scale patterns from which we could identify a small number of states in the United States (US) that act as "bridge regions" contributing to one or more specific influenza spread patterns. Similar to previous studies, the patterns show that the south-eastern regions of the US were widely affected by the H1N1 flu pandemic. Several of these south-eastern states act as bridge regions, which connect the north-east and central US in terms of flu occurrences. CONCLUSIONS: These quantitative insights show how the claims data combined with novel analytical techniques can provide important information to decision makers when an epidemic spreads throughout the country. Taken together ORBiT provides a scalable and extensible platform for public health surveillance.

Machine learning is used in many applications, from machine vision to speech recognition to decision support systems, and it is used to test applications. However, though much has been done to evaluate the performance of machine learning algorithms, little has been done to verify the algorithms or examine their failure modes. Moreover, complex learning frameworks often require stepping beyond black box evaluation to distinguish between errors based on natural limits on learning and errors that arise from mistakes in implementation. We present a conceptual architecture, failure model and taxonomy, and failure modes and effects analysis (FMEA) of a semi-supervised, multi-modal learning system, and provide specific examples from its use in a radiological analysis assistant system. The goal of the research described in this paper is to provide a foundation from which dependability analysis of systems using semi-supervised, multi-modal learning can be conducted. The methods presented provide a first step towards that overall goal.