Search

Explainability is key to enhancing artificial intelligence's trustworthiness in medicine. However, several issues remain concerning the actual benefit of explainable models for clinical decision-making. Firstly, there is a lack of consensus on an evaluation framework for quantitatively assessing the practical benefits that effective explainability should provide to practitioners. Secondly, physician-centered evaluations of explainability are limited. Thirdly, the utility of built-in attention mechanisms in transformer-based models as an explainability technique is unclear. We hypothesize that superior attention maps should align with the information that physicians focus on, potentially reducing prediction uncertainty and increasing model reliability. We employed a multimodal transformer to predict lymph node metastasis in rectal cancer using clinical data and magnetic resonance imaging, exploring how well attention maps, visualized through a state-of-the-art technique, can achieve agreement with physician understanding. We estimated the model's uncertainty using meta-level information like prediction probability variance and quantified agreement. Our assessment of whether this agreement reduces uncertainty found no significant effect. In conclusion, this case study did not confirm the anticipated benefit of attention maps in enhancing model reliability. Superficial explanations could do more harm than good by misleading physicians into relying on uncertain predictions, suggesting that the current state of attention mechanisms in explainability should not be overestimated. Identifying explainability mechanisms truly beneficial for clinical decision-making remains essential.

Patients undergoing chest X-rays (CXR) often endure multiple lung diseases. When evaluating a patient's condition, due to the complex pathologies, subtle texture changes of different lung lesions in images, and patient condition differences, radiologists may make uncertain even when they have experienced long-term clinical training and professional guidance, which makes much noise in extracting disease labels based on CXR reports. In this paper, we re-extract disease labels from CXR reports to make them more realistic by considering disease severity and uncertainty in classification. Our contributions are as follows: 1. We re-extracted the disease labels with severity and uncertainty by a rule-based approach with keywords discussed with clinical experts. 2. To further improve the explainability of chest X-ray diagnosis, we designed a multi-relationship graph learning method with an expert uncertainty-aware loss function. 3. Our multi-relationship graph learning method can also interpret the disease classification results. Our experimental results show that models considering disease severity and uncertainty outperform previous state-of-the-art methods.

This paper focuses on the feasibility of Deep Neural Operator (DeepONet) as a robust surrogate modeling method within the context of digital twin (DT) for nuclear energy systems. Through benchmarking and evaluation, this study showcases the generalizability and computational efficiency of DeepONet in solving a challenging particle transport problem. DeepONet also exhibits remarkable prediction accuracy and speed, outperforming traditional ML methods, making it a suitable algorithm for real-time DT inference. However, the application of DeepONet also reveals challenges related to optimal sensor placement and model evaluation, critical aspects of real-world implementation. Addressing these challenges will further enhance the method's practicality and reliability. Overall, DeepONet presents a promising and transformative nuclear engineering research and applications tool. Its accurate prediction and computational efficiency capabilities can revolutionize DT systems, advancing nuclear engineering research. This study marks an important step towards harnessing the power of surrogate modeling techniques in critical engineering domains.

To contribute to automating the medical vision-language model, we propose a novel Chest-Xray Difference Visual Question Answering (VQA) task. Given a pair of main and reference images, this task attempts to answer several questions on both diseases and, more importantly, the differences between them. This is consistent with the radiologist's diagnosis practice that compares the current image with the reference before concluding the report. We collect a new dataset, namely MIMIC-Diff-VQA, including 700,703 QA pairs from 164,324 pairs of main and reference images. Compared to existing medical VQA datasets, our questions are tailored to the Assessment-Diagnosis-Intervention-Evaluation treatment procedure used by clinical professionals. Meanwhile, we also propose a novel expert knowledge-aware graph representation learning model to address this task. The proposed baseline model leverages expert knowledge such as anatomical structure prior, semantic, and spatial knowledge to construct a multi-relationship graph, representing the image differences between two images for the image difference VQA task. The dataset and code can be found at https://github.com/Holipori/MIMIC-Diff-VQA. We believe this work would further push forward the medical vision language model.

The amount of medical images stored in hospitals is increasing faster than ever; however, utilizing the accumulated medical images has been limited. This is because existing content-based medical image retrieval (CBMIR) systems usually require example images to construct query vectors; nevertheless, example images cannot always be prepared. Besides, there can be images with rare characteristics that make it difficult to find similar example images, which we call isolated samples. Here, we introduce a novel sketch-based medical image retrieval (SBMIR) system that enables users to find images of interest without example images. The key idea lies in feature decomposition of medical images, whereby the entire feature of a medical image can be decomposed into and reconstructed from normal and abnormal features. By extending this idea, our SBMIR system provides an easy-to-use two-step graphical user interface: users first select a template image to specify a normal feature and then draw a semantic sketch of the disease on the template image to represent an abnormal feature. Subsequently, it integrates the two kinds of input to construct a query vector and retrieves reference images with the closest reference vectors. Using two datasets, ten healthcare professionals with various clinical backgrounds participated in the user test for evaluation. As a result, our SBMIR system enabled users to overcome previous challenges, including image retrieval based on fine-grained image characteristics, image retrieval without example images, and image retrieval for isolated samples. Our SBMIR system achieves flexible medical image retrieval on demand, thereby expanding the utility of medical image databases.

Medical visual question answering (VQA) aims to answer clinically relevant questions regarding input medical images. This technique has the potential to improve the efficiency of medical professionals while relieving the burden on the public health system, particularly in resource-poor countries. Existing medical VQA methods tend to encode medical images and learn the correspondence between visual features and questions without exploiting the spatial, semantic, or medical knowledge behind them. This is partially because of the small size of the current medical VQA dataset, which often includes simple questions. Therefore, we first collected a comprehensive and large-scale medical VQA dataset, focusing on chest X-ray images. The questions involved detailed relationships, such as disease names, locations, levels, and types in our dataset. Based on this dataset, we also propose a novel baseline method by constructing three different relationship graphs: spatial relationship, semantic relationship, and implicit relationship graphs on the image regions, questions, and semantic labels. The answer and graph reasoning paths are learned for different questions.

Neural Operator Networks (ONets) represent a novel advancement in machine learning algorithms, offering a robust and generalizable alternative for approximating partial differential equations (PDEs) solutions. Unlike traditional Neural Networks (NN), which directly approximate functions, ONets specialize in approximating mathematical operators, enhancing their efficacy in addressing complex PDEs. In this work, we evaluate the capabilities of Deep Operator Networks (DeepONets), an ONets implementation using a branch/trunk architecture. Three test cases are studied: a system of ODEs, a general diffusion system, and the convection/diffusion Burgers equation. It is demonstrated that DeepONets can accurately learn the solution operators, achieving prediction accuracy scores above 0.96 for the ODE and diffusion problems over the observed domain while achieving zero shot (without retraining) capability. More importantly, when evaluated on unseen scenarios (zero shot feature), the trained models exhibit excellent generalization ability. This underscores ONets vital niche for surrogate modeling and digital twin development across physical systems. While convection-diffusion poses a greater challenge, the results confirm the promise of ONets and motivate further enhancements to the DeepONet algorithm. This work represents an important step towards unlocking the potential of digital twins through robust and generalizable surrogates.

Artificial intelligence (AI) and Machine learning (ML) are increasingly used in energy and engineering systems, but these models must be fair, unbiased, and explainable. It is critical to have confidence in AI's trustworthiness. ML techniques have been useful in predicting important parameters and in improving model performance. However, for these AI techniques to be useful for making decisions, they need to be audited, accounted for, and easy to understand. Therefore, the use of explainable AI (XAI) and interpretable machine learning (IML) is crucial for the accurate prediction of prognostics, such as remaining useful life (RUL), in a digital twin system, to make it intelligent while ensuring that the AI model is transparent in its decision-making processes and that the predictions it generates can be understood and trusted by users. By using AI that is explainable, interpretable, and trustworthy, intelligent digital twin systems can make more accurate predictions of RUL, leading to better maintenance and repair planning, and ultimately, improved system performance. The objective of this paper is to explain the ideas of XAI and IML and to justify the important role of AI/ML in the digital twin framework and components, which requires XAI to understand the prediction better. This paper explains the importance of XAI and IML in both local and global aspects to ensure the use of trustworthy AI/ML applications for RUL prediction. We used the RUL prediction for the XAI and IML studies and leveraged the integrated Python toolbox for interpretable machine learning~(PiML).

In response to the urgent need to establish AI/ML-integrated Digital Twin (DT) technology within next-generation nuclear systems, advancements in modeling methods and simulation codes are necessary. The increased complexity of models demands significant computational resources to quantify their uncertainties. To address this challenge, a data-driven non-intrusive uncertainty quantification method via polynomial chaos expansion is introduced as an efficient strategy within the finite element analysis-based fuel performance code BISON. Models of and fuels, alongside SiC/SiC cladding material, were prepared to demonstrate the proposed method. The impact of four independent uncertain input variables on the system output was quantified, requiring fewer than 100 BISON simulations for each model. This approach not only accelerates the modeling and simulation task but also enhances the reliability in the development of DT-enabling technologies.

Computationally efficient and trustworthy machine learning algorithms are necessary for Digital Twin (DT) framework development. Generally speaking, DT-enabling technologies consist of five major components: (i) Machine learning (ML)-driven prediction algorithm, (ii) Temporal synchronization between physics and digital assets utilizing advanced sensors/instrumentation, (iii) uncertainty propagation, and (iv) DT operational framework. Unfortunately, there is still a significant gap in developing those components for nuclear plant operation. In order to address this gap, this study specifically focuses on the "ML-driven prediction algorithms" as a viable component for the nuclear reactor operation while assessing the reliability and efficacy of the proposed model. Therefore, as a DT prediction component, this study develops a multi-stage predictive model consisting of two feedforward Deep Learning using Neural Networks (DNNs) to determine the final steady-state power of a reactor transient for a nuclear reactor/plant. The goal of the multi-stage model architecture is to convert probabilistic classification to continuous output variables to improve reliability and ease of analysis. Four regression models are developed and tested with input from the first stage model to predict a single value representing the reactor power output. The combined model yields 96% classification accuracy for the first stage and 92% absolute prediction accuracy for the second stage. The development procedure is discussed so that the method can be applied generally to similar systems. An analysis of the role similar models would fill in DTs is performed.

To understand the potential of intelligent confirmatory tools, the U.S. Nuclear Regulatory Committee (NRC) initiated a future-focused research project to assess the regulatory viability of machine learning (ML) and artificial intelligence (AI)-driven Digital Twins (DTs) for nuclear power applications. Advanced accident tolerant fuel (ATF) is one of the priority focus areas of the U.S. Department of Energy (DOE). A DT framework can offer game-changing yet practical and informed solutions to the complex problem of qualifying advanced ATFs. Considering the regulatory standpoint of the modeling and simulation (M&S) aspect of DT, uncertainty quantification and sensitivity analysis are paramount to the DT framework's success in terms of multi-criteria and risk-informed decision-making. This chapter introduces the ML-based uncertainty quantification and sensitivity analysis methods while exhibiting actual applications to the finite element-based nuclear fuel performance code BISON.

Robust optimization is a method for optimization under uncertainties in engineering systems and designs for applications ranging from aeronautics to nuclear. In a robust design process, parameter variability (or uncertainty) is incorporated into the engineering systems' optimization process to assure the systems' quality and reliability. This chapter focuses on a robust optimization approach for developing robust and reliable advanced systems and explains the framework for using uncertainty quantification and optimization techniques. For the uncertainty analysis, a polynomial chaos-based approach is combined with the optimization algorithms MOSA (Multi-Objective Simulated Annealing), and the process is discussed with a simplified test function. For the optimization process, gradient-free genetic algorithms are considered as the optimizer scans the whole design space, and the optimal values are not always dependent on the initial values.

It is important to accurately model materials' properties at lower length scales (micro-level) while translating the effects to the components and/or system level (macro-level) can significantly reduce the amount of experimentation required to develop new technologies. Robustness analysis of fuel and structural performance for harsh environments (such as power uprated reactor systems or aerospace applications) using machine learning-based multi-scale modeling and robust optimization under uncertainties are required. The fiber and matrix material characteristics are potential sources of uncertainty at the microscale. The stacking sequence (angles of stacking and thickness of layers) of composite layers causes meso-scale uncertainties. It is also possible for macro-scale uncertainties to arise from system properties, like the load or the initial conditions. This chapter demonstrates advanced data-driven methods and outlines the specific capability that must be developed/added for the multi-scale modeling of advanced composite materials. This chapter proposes a multi-scale modeling method for composite structures based on a finite element method (FEM) simulation driven by surrogate models/emulators based on microstructurally informed meso-scale materials models to study the impact of operational parameters/uncertainties using machine learning approaches. To ensure optimal composite materials, composite properties are optimized with respect to initial materials volume fraction using data-driven numerical algorithms.

The application of digital twin (DT) technology to the nuclear field is one of the challenges in the future development of nuclear energy. Possible applications of DT technology in the nuclear field are expected to be very wide: operate commercial nuclear reactors, monitor spent fuel storage and disposal facilities, and develop new nuclear systems. As U.S. Nuclear Regulatory Committee (NRC) recently announced, machine learning (ML) and artificial intelligence (AI) will be new domains in the nuclear field. Considering the data science perspective, Gaussian Process (GP) has proven to be an ML algorithm for modeling and simulation components of the digital twin framework, specifically for the accident tolerant fuel (ATF) concepts. ATF is one of the high-priority areas for both the U.S. Department of Energy (DOE) and NRC. GP's inherent treatment of lack of data, missing data, and data inconsistencies (noisy/erroneous data) present in the ATF concepts make it an attractive machine learning algorithm for implementation in the DT framework. This chapter focuses on the practical demonstration of GP and its applicability to the DT framework for predicting ATF.

The Gaussian Process (GP)-based surrogate model has the inherent capability of capturing the anomaly arising from limited data, lack of data, missing data, and data inconsistencies (noisy/erroneous data) present in the modeling and simulation component of the digital twin framework, specifically for the accident tolerant fuel (ATF) concepts. However, GP will not be very accurate when we have limited high-fidelity (experimental) data. In addition, it is challenging to apply higher dimensional functions (>20-dimensional function) to approximate predictions with the GP. Furthermore, noisy data or data containing erroneous observations and outliers are major challenges for advanced ATF concepts. Also, the governing differential equation is empirical for longer-term ATF candidates, and data availability is an issue. Physics-informed multi-fidelity Kriging (MFK) can be useful for identifying and predicting the required material properties. MFK is particularly useful with low-fidelity physics (approximating physics) and limited high-fidelity data - which is the case for ATF candidates since there is limited data availability. This chapter explores the method and presents its application to experimental thermal conductivity measurement data for ATF. The MFK method showed its significance for a small number of data that could not be modeled by the conventional Kriging method. Mathematical models constructed with this method can be easily connected to later-stage analysis such as uncertainty quantification and sensitivity analysis and are expected to be applied to fundamental research and a wide range of product development fields. The overarching objective of this chapter is to show the capability of MFK surrogates that can be embedded in a digital twin system for ATF.

Industry 4.0 targets the conversion of the traditional industries into intelligent ones through technological revolution. This revolution is only possible through innovation, optimization, interconnection, and rapid decision-making capability. Numerical models are believed to be the key components of Industry 4.0, facilitating quick decision-making through simulations instead of costly experiments. However, numerical investigation of precise, high-fidelity models for optimization or decision-making is usually time-consuming and computationally expensive. In such instances, data-driven surrogate models are excellent substitutes for fast computational analysis and the probabilistic prediction of the output parameter for new input parameters. The emergence of Internet of Things (IoT) and Machine Learning (ML) has made the concept of surrogate modeling even more viable. However, these surrogate models contain intrinsic uncertainties, originate from modeling defects, or both. These uncertainties, if not quantified and minimized, can produce a skewed result. Therefore, proper implementation of uncertainty quantification techniques is crucial during optimization, cost reduction, or safety enhancement processes analysis. This chapter begins with a brief overview of the concept of surrogate modeling, transfer learning, IoT and digital twins. After that, a detailed overview of uncertainties, uncertainty quantification frameworks, and specifics of uncertainty quantification methodologies for a surrogate model linked to a digital twin is presented. Finally, the use of uncertainty quantification approaches in the nuclear industry has been addressed.

The concept of small modular reactor has changed the outlook for tackling future energy crises. This new reactor technology is very promising considering its lower investment requirements, modularity, design simplicity, and enhanced safety features. The application of artificial intelligence-driven multi-scale modeling (neutronics, thermal hydraulics, fuel performance, etc.) incorporating Digital Twin and associated uncertainties in the research of small modular reactors is a recent concept. In this work, a comprehensive study is conducted on the multiscale modeling of accident-tolerant fuels. The application of these fuels in the light water-based small modular reactors is explored. This chapter also focuses on the application of machine learning and artificial intelligence in the design optimization, control, and monitoring of small modular reactors. Finally, a brief assessment of the research gap on the application of artificial intelligence to the development of high burnup composite accident-tolerant fuels is provided. Necessary actions to fulfill these gaps are also discussed.

In medical imaging, the characteristics purely derived from a disease should reflect the extent to which abnormal findings deviate from the normal features. Indeed, physicians often need corresponding images without abnormal findings of interest or, conversely, images that contain similar abnormal findings regardless of normal anatomical context. This is called comparative diagnostic reading of medical images, which is essential for a correct diagnosis. To support comparative diagnostic reading, content-based image retrieval (CBIR), which can selectively utilize normal and abnormal features in medical images as two separable semantic components, will be useful. Therefore, we propose a neural network architecture to decompose the semantic components of medical images into two latent codes: normal anatomy code and abnormal anatomy code. The normal anatomy code represents normal anatomies that should have existed if the sample is healthy, whereas the abnormal anatomy code attributes to abnormal changes that reflect deviation from the normal baseline. These latent codes are discretized through vector quantization to enable binary hashing, which can reduce the computational burden at the time of similarity search. By calculating the similarity based on either normal or abnormal anatomy codes or the combination of the two codes, our algorithm can retrieve images according to the selected semantic component from a dataset consisting of brain magnetic resonance images of gliomas. Our CBIR system qualitatively and quantitatively achieves remarkable results.

Medical images can be decomposed into normal and abnormal features, which is considered as the compositionality. Based on this idea, we propose an encoder-decoder network to decompose a medical image into two discrete latent codes: a normal anatomy code and an abnormal anatomy code. Using these latent codes, we demonstrate a similarity retrieval by focusing on either normal or abnormal features of medical images., Comment: Machine Learning for Health (ML4H) at NeurIPS 2020 - Extended Abstract

In real-world clinical practice, overlooking unanticipated findings can result in serious consequences. However, supervised learning, which is the foundation for the current success of deep learning, only encourages models to identify abnormalities that are defined in datasets in advance. Therefore, abnormality detection must be implemented in medical images that are not limited to a specific disease category. In this study, we demonstrate an unsupervised learning framework for pixel-wise abnormality detection in brain magnetic resonance imaging captured from a patient population with metastatic brain tumor. Our concept is as follows: If an image reconstruction network can faithfully reproduce the global features of normal anatomy, then the abnormal lesions in unseen images can be identified based on the local difference from those reconstructed as normal by a discriminative network. Both networks are trained on a dataset comprising only normal images without labels. In addition, we devise a metric to evaluate the anatomical fidelity of the reconstructed images and confirm that the overall detection performance is improved when the image reconstruction network achieves a higher score. For evaluation, clinically significant abnormalities are comprehensively segmented. The results show that the area under the receiver operating characteristics curve values for metastatic brain tumors, extracranial metastatic tumors, postoperative cavities, and structural changes are 0.78, 0.61, 0.91, and 0.60, respectively., Comment: This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible