In manufacturing processes, a multitude of sensors are typically deployed to collect data of process parameters. While this provides an opportunity to better predict and control the quality of the final product, the relationship between the process variables and the desired final quality is often not well-understood. To establish this relationship, machine learning (ML) models can be used. However, collecting large, labeled datasets to train the ML model can be difficult, as creating such datasets typically involves costly or destructive end-of-line quality testing of the products. To overcome this challenge, we propose a novel framework called Physics-informed Weakly-supervised Learning (PWL) that integrates physics-based models with data-driven ML models. By leveraging physical knowledge and using the outputs of physics-based models as weak labels, PWL offers an alternative to traditional methods that require large, labeled datasets. Our approach simultaneously optimizes the data-driven ML model, as well as the discrepancy and calibration parameters of the physics-based model, resulting in superior predictive performance compared to either model used alone. We demonstrate the effectiveness of PWL through simulation experiments, comparisons with existing methods, and two real-world case studies, highlighting its potential for improving quality prediction in various manufacturing systems.Note to Practitioners—The proposed method in this paper, Physics-informed Weakly-supervised Learning (PWL), addresses a common problem in manufacturing processes, where obtaining large, labeled datasets can be costly or not always possible. By integrating physics-based models with data-driven ML models, PWL leverages available physical knowledge to improve the predictive performance of product quality in manufacturing systems. This enables practitioners to better understand and optimize their manufacturing processes even with limited labeled datasets, leading to improved product quality and reduced production costs.

Multistage Manufacturing Systems (MMS) are common in industries involving complex processes with multiple stages, each impacting the final product quality. Traditional product quality modeling approaches struggle with the intricate interdependencies and variable structures within these systems, further complicated by the extensive sensor-generated data. In this paper, we introduce Graph-based Variation Propagation Network (GVPNet), an innovative approach for end-to-end learning of product quality representations and their propagation through MMS stages. GVPNet utilizes a heterogeneous Graph Attention Network (hGAT) architecture that uses a graph representation specifically tailored for MMS, facilitating the effective application of graph neural networks (GNN) in this context. The network’s ability to aggregate information and learn node embeddings enables it to predict quality variables several stages ahead, thus offering a proactive tool for quality control. GVPNet was applied in two real world case studies and demonstrated its superiority in predictive accuracy and robustness over benchmarks.

Early detection of Alzheimer’s Disease (AD) is crucial for timely interventions and optimizing treatment outcomes. Integrating multimodal neuroimaging datasets can enhance the early detection of AD. However, models must address the challenge of incomplete modalities, a common issue in real-world scenarios, as not all patients have access to all modalities due to practical constraints such as cost and availability. We propose a deep learning framework employing Incomplete Cross-modal Mutual Knowledge Distillation (IC-MKD) to model different sub-cohorts of patients based on their available modalities. In IC-MKD, the multimodal model (e.g., MRI and PET) serves as a teacher, while the single-modality model (e.g., MRI only) is the student. Our IC-MKD framework features three components: a Modality-Disentangling Teacher (MDT) model designed through information disentanglement, a student model that learns from classification errors and MDT’s knowledge, and the teacher model enhanced via distilling the student’s single-modal feature extraction capabilities. Moreover, we show the effectiveness of the proposed method through theoretical analysis and validate its performance with simulation studies. In addition, our method is demonstrated through a case study with Alzheimer’s Disease Neuroimaging Initiative (ADNI) datasets, underscoring the potential of artificial intelligence in addressing incomplete multimodal neuroimaging datasets and advancing early AD detection. Note to Practitioners—This paper was motivated by the challenge of early AD diagnosis, particularly in scenarios when clinicians encounter varied availability of patient imaging data, such as MRI and PET scans, often constrained by cost or accessibility issues. We propose an incomplete multimodal learning framework that produces tailored models for patients with only MRI and patients with both MRI and PET. This approach improves the accuracy and effectiveness of early AD diagnosis, especially when imaging resources are limited, via bi-directional knowledge transfer. We introduced a teacher model that prioritizes extracting common information between different modalities, significantly enhancing the student model’s learning process. This paper includes theoretical analysis, simulation study, and real-world case study to illustrate the method’s promising potential in early AD detection. However, practitioners should be mindful of the complexities involved in model tuning. Future work will focus on improving model interpretability and expanding its application. This includes developing methods to discover the key brain regions for predictions, enhancing clinical trust, and extending the framework to incorporate a broader range of imaging modalities, demographic information, and clinical data. These advancements aim to provide a more comprehensive view of patient health and improve diagnostic accuracy across various neurodegenerative diseases.

Cone beam computed tomography (CBCT) is a widely-used imaging modality in dental healthcare. It is an important task to segment each 3D CBCT image, which involves labeling lesions, bones, teeth, and restorative materials on a voxel-by-voxel basis, as it aids in lesion detection, diagnosis, and treatment planning. The current clinical practice relies on manual segmentation, which is labor-intensive and demands considerable expertise. Leveraging Artificial Intelligence (AI) to fully automate the segmentation process could tremendously improve the quality and efficiency of dental healthcare. The main hurdle in this advancement is reducing AI’s reliance on a large quantity of manually labeled images to train robust, accurate, and generalizable algorithms. To tackle this challenge, we propose a novel Oral-Anatomical Knowledge-informed Semi-Supervised Learning (OAK-SSL) model for 3D CBCT image segmentation and lesion detection. The uniqueness of OAK-SSL is its capability of integrating qualitative oral-anatomical knowledge of plausible lesion locations into the deep learning design. Specifically, the unique design of OAK-SSL includes three key elements, including transformation of qualitative knowledge into quantitative representation, knowledge-informed dual-task learning architecture, and knowledge-informed semi-supervised loss function. We apply OAK-SSL to a real-world dataset, focusing on segmenting CBCT images that contain small lesions. This task is inherently challenging yet holds significant clinical value as treating lesions at their early stages lead to excellent prognosis. OAK-SSL demonstrated significantly better performance than a range of existing methods. Note to Practitioners—This study tackles the challenges arising from a limited amount of labeled data due to the time-consuming manual segmentation of 3D dental cone beam computed tomography (CBCT) images. The scarcity of labeled data often impedes AI models from accurately segmenting periapical lesions. To overcome this, we introduce a novel semi-supervised learning algorithm that integrates the oral-anatomical knowledge about lesion location for 3D CBCT image segmentation. Our method effectively segments periapical lesions, including even small-sized periapical lesions, without solely relying on labeled data. The proposed method offers two significant benefits to clinicians. First, it reduces the necessity for large amounts of labeled data, particularly easing the burden of manually segmenting early-stage periapical lesions. Second, it helps reduce intra- and inter-observer disagreements and human errors by providing consistent and automated segmentation maps. These benefits not only simplify the segmentation process in dental imaging but also improve its reliability. As a result, our automated algorithm makes it easier and more trustworthy for practitioners. However, practitioners should be aware that the effectiveness of our method relies on the assumption of consistency between labeled and unlabeled data. When applying this method, it is crucial to carefully consider the characteristics of the unlabeled dataset. Significant differences in image quality, patient demographics, or acquisition parameters between labeled and unlabeled data might affect model performance.

Cancer remains one of the most challenging diseases to treat in the medical field. Machine learning (ML) has enabled in-depth analysis of complex patterns from large, diverse datasets, greatly facilitating “healthcare automation” in cancer diagnosis and prognosis. Despite these advancements, ML models face challenges stemming from limited labeled sample sizes, the intricate interplay of high-dimensionality data types, the inherent heterogeneity observed among patients and within tumors, and concerns about interpretability and consistency with existing biomedical knowledge. One approach to address these challenges is to integrate biomedical knowledge into data-driven models, which has proven potential to improve the accuracy, robustness, and interpretability of model results. Here, we review the state-of-the-art ML studies that leverage the fusion of biomedical knowledge and data, termed knowledge-informed machine learning (KIML), to advance cancer diagnosis and prognosis. We provide an overview of diverse forms of knowledge representation and current strategies of knowledge integration into machine learning pipelines with concrete examples. We conclude the review article by discussing future directions aimed at leveraging KIML to advance cancer research and healthcare automation. A live summary of the review is hosted athttps://lingchm.github.io/kinformed-machine-learning-cancer/offering an evolving resource to support research in this field.Note to Practitioners—Recognizing the challenges posed by inter-patient and intratumoral heterogeneity, constrained sample size, and interpretability requirements in cancer applications, practitioners should consider integration of existing biomedical knowledge into their modeling frameworks. This strategy holds promise for enhancing model performance, robustness, and interpretability. We review classic machine learning and deep learning models that incorporated domain knowledge in their cancer diagnosis and prognosis models spanning models that use clinical, imaging, molecular, and treatment data. Pros and cons of each integration approach are discussed. Key design questions such as which knowledge to leverage, how to represent it effectively, and how to seamlessly integrate it into their models need be examined for each case. Collaboration between modeling scientists and medical experts is essential in this endeavor.

Ordinal learning (OL) is a type of machine learning models with broad utility in health care applications such as diagnosis of different grades of a disease (e.g., mild, modest, severe) and prediction of the speed of disease progression (e.g., very fast, fast, moderate, slow). This paper aims to tackle a situation when precisely labeled samples are limited in the training set due to cost or availability constraints, whereas there could be an abundance of samples with imprecise labels. We focus on imprecise labels that are intervals, i.e., one can know that the a sample belongs to an interval of labels but cannot know which unique label it has. This situation is quite common in health care datasets due to limitations of the diagnostic instrument, sparse clinical visits, or/and patient dropout. Limited research has been done to develop OL models with imprecise/interval labels. We propose a new Hybrid Ordinal Learner (HOL) to integrate samples with both precise and interval labels to train a robust OL model. We also develop a tractable and efficient optimization algorithm to solve the HOL formulation. We compare HOL with several recently developed OL methods on four benchmarking datasets, which demonstrate the superior performance of HOL. Finally, we apply HOL to a real-world dataset for predicting the speed of progressing to Alzheimer’s Disease (AD) for individuals with Mild Cognitive Impairment (MCI) based on a combination of multi-modality neuroimaging and demographic/clinical datasets. HOL achieves high accuracy in the prediction and outperforms existing methods. The capability of accurately predicting the speed of progression to AD for each individual with MCI has the potential for helping facilitate more individually-optimized interventional strategies.Note to Practitioners—Machine learning (ML) algorithms have been widely adopted to support disease diagnosis and prognosis. In some situations, the outcome variable of interest is on an ordinal scale, i.e., it includes several classes with a natural order. For example, the variable of interest can be the grade of a disease as mild, moderate, or severe; or it can be the progression speed of a disease as very fast, fast, moderate, or slow. Ordinal learning (OL) is the type of ML algorithms for ordinal variable prediction. Most existing OL algorithms can only include samples with precise labels in training. However, it is common to have samples with imprecise/interval labels, i.e., we know that a sample belongs to a range of classes/labels but do not know which specific class/label it belongs to. This situation can happen due to a variety of different reasons such as use of less accurate diagnostic instrument under cost or availability constraints, sparse clinical assessment, and patient dropout. We propose a Hybrid Ordinal Learner (HOL) to integrate samples with both precise and interval labels to train a robust OL model. HOL is evaluated using four public benchmarking datasets and shows superior performance compared to existing methods. Also, we apply HOL to a real-world dataset for predicting the speed of progressing to Alzheimer’s Disease (AD) for individuals with Mild Cognitive Impairment (MCI). MCI is the prodromal stage of AD. Individuals with MCI show noticeable signs of memory loss and cognitive declines, but these symptoms are not severe enough to interfere their independent living. HOL achieves high accuracy in predicting the speed of progressing to AD for each MCI subject (e.g., the speed of ‘very fast’‘, fast’‘, moderate’, or ‘slow), which could potentially help facilitate the development of more individually-optimized interventional strategies.

Unlocking the power of 3D point cloud machine learning models can be a challenge due to the need for extensive labeled datasets, which presents a challenge when applying these models to new domains. Transfer learning can help overcome this challenge by utilizing data from related tasks to enhance model performance. However, traditional (2D) transfer learning methods struggle with 3D point cloud domain adaptation, due to differences in physical environments and sensor configurations. To address this issue, we propose PLURAL, a novel 3D point cloud transfer learning methodology based on contrastive learning with augmentations. Our approach is inspired by the notion that high-level shape features are more transferable than low-level geometry features. We propose a co-training architecture that includes separate 3D point cloud models with domain-specific parameters, as well as a module for learning domain-invariant features. Additionally, PLURAL extends the approach of contrastive instance alignment to 3D point cloud modeling by considering physics-informed hard sample mining. Our experiments on simulation and real-world datasets demonstrate that PLURAL outperforms state-of-the-art transfer learning methods by a significant margin, effectively reducing the domain gap.Note to Practitioners—The usage of 3D point cloud machine learning models is currently limited by the need for extensive labeled data. With our proposed framework, data from related tasks can be utilized to enhance the model performance on new applications or domains. PLURAL explicitly considers the acquisition of 3D point clouds by diverse sensors and in diverse environments. The method is highly adaptable and includes separate models with domain-specific parameters, making it applicable to a wide range of applications and domains.

Telemonitoring is the use of electronic devices such as smartphones to remotely monitor patients. It provides great convenience and enables timely medical decisions. To facilitate the decision making for each patient, a model is needed to translate the data collected by the patient’s smartphone into a predicted score for his/her disease severity. To train a robust predictive model, semi-supervised learning (SSL) provides a viable approach by integrating both labeled and unlabeled samples to leverage all the available data from each patient. There are two challenging issues that need to be simultaneously addressed in using SSL for this problem: (1) feature selection from high-dimensional noisy telemonitoring data; and (2) instance selection from many, possibly redundant unlabeled samples. We propose a novel SSL model allowing for simultaneous feature and instance selection, namely the S2SSL model. We present a real-data application of telemonitoring for patients with Parkinson’s Disease using their smartphone-collected activity data such as tapping and speaking. A total of 382 features were extracted from the activity data of each patient. 74 labeled and 563 unlabeled instances from 37 patients were used to train S2SSL. The trained model achieved a high accuracy of 0.828 correlation between the true and predicted disease severity scores on a validation dataset. Note to Practitioners—Telemonitoring is an emerging health care platform enabled by smartphones and wearables. Because it allows for health data to be collected anytime and anywhere, patients can be frequently monitored and medical decisions can be made more timely and effectively. This paper addresses the data science challenges in leveraging the telemonitoring platform to benefit patient care. Specifically, we propose a new model, S2SSL, to tackle these challenges and provide better robustness, accuracy, and efficiency. This paper may be interesting to health care practitioners seeking advanced analytics capabilities to model and integrate the data collected through telemonitoring devices, with ultimate purposes of improving the decision in treating each patient and increasing patient access to specialized care.

Precision medicine aims to provide diagnosis and treatment accounting for individual differences. To develop machine learning models in support of precision medicine, personalized models are expected to have better performance than one-model-fits-all approaches. A significant challenge, however, is the limited number of labeled samples that can be collected from each individual due to practical constraints. Transfer Learning (TL) addresses this challenge by leveraging the information of other patients with the same disease (i.e., the source domain) when building a personalized model for each patient (i.e., the target domain). We propose Weakly-Supervised Transfer Learning (WS-TL) to tackle two challenges that existing TL algorithms do not address well: (i) the target domain has only a few or even no labeled samples; (ii) how to integrate domain knowledge into the TL design. We design a novel mathematical framework of WS-TL to learn a model for the target domain based on paired samples whose order relationships are inferred from domain knowledge, while at the same time integrating labeled samples in the source domain for transfer learning. Also, we propose an efficient active sampling strategy to select informative paired samples. Theoretical properties were investigated. Finally, we present a real-world application in precision medicine of brain cancer, where WS-TL is used to build personalized patient models to predict Tumor Cell Density (TCD) distribution across the brain based on MRI images. WS-TL has the highest accuracy compared to a variety of existing TL algorithms. The predicted TCD map for each patient can help facilitate individually optimized treatment.Note to Practitioners—This work was motivated by Precision Medicine applications that need to build personalized machine learning models to account for individual differences. Due to limited data from each person, Transfer Learning (TL) provides a promising approach, which can leverage the information of other patients with the same disease (i.e., the source domain) when building a personalized model for each patient (i.e., the target domain). The proposed WS-TL model addresses the application scenarios with two unique properties: (i) the target domain has a few and even no labeled samples, which is a challenging situation that most existing TL methods do not address well; (ii) there is domain knowledge to provide weak labels for a large number of unlabeled samples in the form of order relationships, which provides an opportunity to integrate the domain knowledge into the TL design. We demonstrate WS-TL in a Precision Medicine application for brain cancer and show promising results. WS-TL has the potential of addressing a broad range of other application areas in building personalized models.

The automated capability of generating spatial prediction for a variable of interest is desirable in various science and engineering domains. Take precision medicine of cancer as an example, in which the goal is to match patients with treatments based on molecular markers identified in each patient’s tumor. A substantial challenge, however, is that the molecular markers can vary significantly at different spatial locations of a tumor. If this spatial distribution could be predicted, the precision of cancer treatment could be greatly improved by adapting treatment to the spatial molecular heterogeneity. This is a challenging task because no technology is available to measure the molecular markers at each spatial location within a tumor. Biopsy samples provide direct measurement, but they are scarce/local. Imaging, such as MRI, is global, but it only provides proxy/indirect measurement. Also available are mechanistic models or domain knowledge, which are often approximate or incomplete. This article proposes a novel machine learning framework to fuse the three sources of data/information to generate a spatial prediction, namely, the knowledge-infused global-local (KGL) data fusion model. A novel mathematical formulation is proposed and solved with theoretical study. We present a real-data application of predicting the spatial distribution of tumor cell density (TCD)—an important molecular marker for brain cancer. A total of 82 biopsy samples were acquired from 18 patients with glioblastoma, together with six MRI contrast images from each patient and biological knowledge encoded by a PDE simulator-based mechanistic model called proliferation-invasion (PI). KGL achieved the highest prediction accuracy and minimum prediction uncertainty compared with a variety of competing methods. The result has important implications for providing individualized, spatially optimized treatment for each patient.Note to Practitioners—This article proposes a machine learning framework to fuse local data, global imaging, and domain knowledge to generate a spatial prediction for a variable of interest. This methodology is relevant to multiple application domains. In precision medicine, it will allow for mapping the spatial distribution of important, treatment-informing molecular markers across each tumor by integrating biopsy data, MRI, and biological knowledge. This capability can help resolve the spatial heterogeneity of molecular characteristics and greatly improve the precision of cancer treatment. Other applications include early detection of regional fire risk across a forest by integrating ground/aerial survey data, satellite imagery, and fire simulator output, as well as regional poverty estimation for resource allocation.

Machine learning research on medical images is becoming popular as advanced imaging technologies and equipment in medicine become more and more available. Dental Cone-beam Computed Tomography (Dental CBCT), a frequently-used visualization tool for oral diagnosis, provides valuable three-dimensional information, whose development for automation of Dental CBCT analysis, on the other hand, is relatively preliminary. Generally, there are three important characteristics for analyzing Dental CBCT with noisy labels and limited labeled sample size, and availability of oral medicine knowledge. Based on those characteristics, we develop an image segmentation method for Dental CBCT by integrating domain knowledge into deep U-Net for the 3D segmentation. Finally, depending on whether the knowledge can be decomposed into each pixel, the knowledge constraints are classified into two types: separable and non-separable constraints. All knowledge constraints can be represented as a posterior regularization term and solved in different ways in accordance with related types. For separable constraints, the mean-field theory is employed to solve an optimization problem with the independence assumption about the distributions of output variables on each pixel. For non-separable constraints, we propose to combine the importance sampling based approach and the stochastic optimization algorithm. Finally, we propose to formulate the domain knowledge to the learning stage to improve the accuracy and efficiency of automation of Dental CBCT segmentation. Finally, we will apply the proposed methods into the real datasets collected and manually labeled by the doctors at the University of Pennsylvania.

Compared with the rapidly growing artificial intelligence (AI) research in other branches of healthcare, the pace of developing AI capacities in dental care is relatively slow. Dental care automation, especially the automated capability for dental cone beam computed tomography (CBCT) segmentation and lesion detection, is highly needed. CBCT is an important imaging modality that is experiencing ever-growing utilization in various dental specialties. However, little research has been done for segmenting different structures, restorative materials, and lesions using deep learning. This is due to multifold challenges such as content-rich oral cavity and significant within-label variation on each CBCT image as well as the inherent difficulty of obtaining many high-quality labeled images for training. On the other hand, oral-anatomical knowledge exists in dentistry, which shall be leveraged and integrated into the deep learning design. In this article, we propose a novel anatomically constrained Dense U-Net for integrating oral-anatomical knowledge with data-driven Dense U-Net. The proposed algorithm is formulated as a regularized or constrained optimization and solved using mean-field variational approximation to achieve computational efficiency. Mathematical encoding for transforming descriptive knowledge into a quantitative form is also proposed. Our experiment demonstrates that the proposed algorithm outperforms the standard Dense U-Net in both lesion detection accuracy and dice coefficient (DICE) indices in multilabel segmentation. Benefited from the integration with anatomical domain knowledge, our algorithm performs well with data from a small number of patients included in the training. Note to Practitioners-This article proposes a novel deep learning algorithm to enable the automated capability for cone beam computed tomography (CBCT) segmentation and lesion detection. Despite the growing adoption of CBCT in various dental specialties, such capability is currently lacking. The proposed work will provide tools to help reduce subjectivity and human errors, as well as streamline and expedite the clinical workflow. This will greatly facilitate dental care automation. Furthermore, due to the capacity of integrating oral-anatomical knowledge into the deep learning design, the proposed algorithm does not require many high-quality labeled images to train. The algorithm can provide good accuracy under limited training samples. This ability is highly desirable for practitioners by saving labor-intensive, costly labeling efforts, and enjoying the benefits provided by AI.

Telemonitoring is the use of electronic devices to remotely monitor patients. Taking the Parkinson's disease (PD) as an example, the use of at-home testing device (AHTD) enables remote, internet-based measurement of PD vocal symptoms. Translating AHTD measurement into a unified PD rating scale (UPDRS) through predictive analytics enables cost-effective, convenient, and close tracking of PD progression. Building a predictive model between AHTD measurement and UPDRS is not straightforward because PD patients are highly heterogeneous, which requires patient-specific models. Learning a patient-specific model faces the challenge of limited data. Transfer learning (TL) tackles this challenge by leveraging other patients' information to make up the data shortage when modeling a target patient. Among different TL methods, the category of parameter transfer methods is more appropriate for the telemonitoring application because it transfers patient-specific model parameters but not patients' data. However, existing parameter transfer methods fall short because not every other patient's information is helpful and blind transfer causes the problem of negative transfer. To tackle this limitation, we propose a positive TL (PTL) method. We provide an in-depth theoretical study on the risk and condition for negative transfer to happen, which further drive the development of novel PTL algorithms that are robust to negative transfer. We apply PTL to predict UPDRS of 42 PD patients using their AHTD vocal measurement. PTL achieves significantly better accuracy compared with single learning and one-model-fits-all approaches.

Structure learning of Bayesian Networks (BNs) is an important topic in machine learning. Driven by modern applications in genetics and brain sciences, accurate and efficient learning of large-scale BN structures from high-dimensional data becomes a challenging problem. To tackle this challenge, we propose a Sparse Bayesian Network (SBN) structure learning algorithm that employs a novel formulation involving one L1-norm penalty term to impose sparsity and another penalty term to ensure that the learned BN is a Directed Acyclic Graph--a required property of BNs. Through both theoretical analysis and extensive experiments on 11 moderate and large benchmark networks with various sample sizes, we show that SBN leads to improved learning accuracy, scalability, and efficiency as compared with 10 existing popular BN learning algorithms. We apply SBN to a real-world application of brain connectivity modeling for Alzheimer's disease (AD) and reveal findings that could lead to advancements in AD research.

Modeling the reliability of multi-level engineering systems is critically important yet technically challenging. Existing methods have limited consideration of the failure interdependency caused by the interactions among subsystems and components. In this paper, a new Bayesian Network (BN) representation of system structure and component interactions is proposed. Based on the BN representation, a Bayesian framework is developed to fuse the information from different levels of a system for modeling the reliability of the components, subsystems and the system as a whole. A case study is conducted to demonstrate the effectiveness of the proposed methodology.

Recent studies have shown that Alzheimer's disease (AD) is related to alteration in brain connectivity networks. One type of connectivity, called effective connectivity, defined as the directional relationship between brain regions, is essential to brain function. However, there have been few studies on modeling the effective connectivity of AD and characterizing its difference from normal controls (NC). In this paper, we investigate the sparse Bayesian Network (BN) for effective connectivity modeling. Specifically, we propose a novel formulation for the structure learning of BNs, which involves one L1-norm penalty term to impose sparsity and another penalty to ensure the learned BN to be a directed acyclic graph - a required property of BNs. We show, through both theoretical analysis and extensive experiments on eleven moderate and large benchmark networks with various sample sizes, that the proposed method has much improved learning accuracy and scalability compared with ten competing algorithms. We apply the proposed method to FDG-PET images of 42 AD and 67 NC subjects, and identify the effective connectivity models for AD and NC, respectively. Our study reveals that the effective connectivity of AD is different from that of NC in many ways, including the global-scale effective connectivity, intra-lobe, interlobe, and inter-hemispheric effective connectivity distributions, as well as the effective connectivity associated with specific brain regions. These findings are consistent with known pathology and clinical progression of AD, and will contribute to AD knowledge discovery.

Diagnosis of Alzheimer's disease (AD) at the early stage of the disease development is of great clinical importance. Current clinical assessment that relies primarily on cognitive measures proves low sensitivity and specificity. The fast growing neuroimaging techniques hold great promise. Research so far has focused on single neuroimaging modalities. However, as different modalities provide complementary measures for the same disease pathology, fusion of multi-modality data may increase the statistical power in identification of disease-related brain regions. This is especially true for early AD, at which stage the disease-related regions are most likely to be weak-effect regions that are difficult to be detected from a single modality alone. We propose a sparse composite linear discriminant analysis model (SCLDA) for identification of disease-related brain regions of early AD from multi-modality data. SCLDA uses a novel formulation that decomposes each LDA parameter into a product of a common parameter shared by all the modalities and a parameter specific to each modality, which enables joint analysis of all the modalities and borrowing strength from one another. We prove that this formulation is equivalent to a penalized likelihood with non-convex regularization, which can be solved by the DC ((difference of convex functions) programming. We show that in using the DC programming, the property of the non-convex regularization in terms of preserving weak-effect features can be nicely revealed. We perform extensive simulations to show that SCLDA outperforms existing competing algorithms on feature selection, especially on the ability for identifying weak-effect features. We apply SCLDA to the Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET) images of 49 AD patients and 67 normal controls (NC). Our study identifies disease-related brain regions consistent with findings in the AD literature.

Field failure data often exhibit extra heterogeneity as early failure data may have quite different distribution characteristics from later failure data. These infant failures may come from a defective subpopulation instead of the normal product population. Many exiting methods for field failure analyses focus only on the estimation for a hypothesized mixture model, while the model identification is ignored. This paper aims to develop efficient, accurate methods for both detecting data heterogeneity, and estimating mixture model parameters. Mixture distribution detection is achieved by applying a mixture detection plot (MDP) on field failure observations. The penalized likelihood method, and the expectation-maximization (EM) algorithm are then used for estimating the components in the mixture model. Two field datasets are employed to demonstrate and validate the proposed approach.

Effective diagnosis of Alzheimer's disease (AD), the most common type of dementia in elderly patients, is of primary importance in biomedical research. Recent studies have demonstrated that AD is closely related to the structure change of the brain network, i.e., the connectivity among different brain regions. The connectivity patterns will provide useful imaging-based biomarkers to distinguish Normal Controls (NC), patients with Mild Cognitive Impairment (MCI), and patients with AD. In this paper, we investigate the sparse inverse covariance estimation technique for identifying the connectivity among different brain regions. In particular, a novel algorithm based on the block coordinate descent approach is proposed for the direct estimation of the inverse covariance matrix. One appealing feature of the proposed algorithm is that it allows the user feedback (e.g., prior domain knowledge) to be incorporated into the estimation process, while the connectivity patterns can be discovered automatically. We apply the proposed algorithm to a collection of FDG-PET images from 232 NC, MCI, and AD subjects. Our experimental results demonstrate that the proposed algorithm is promising in revealing the brain region connectivity differences among these groups.

Recent advances in neuroimaging techniques provide great potentials for effective diagnosis of Alzheimer’s disease (AD), the most common form of dementia. Previous studies have shown that AD is closely related to alternation in the functional brain network, i.e., the functional connectivity among different brain regions. In this paper, we consider the problem of learning functional brain connectivity from neuroimaging, which holds great promise for identifying image-based markers used to distinguish Normal Controls (NC), patients with Mild Cognitive Impairment (MCI), and patients with AD. More specifically, we study sparse inverse covariance estimation (SICE), also known as exploratory Gaussian graphical models, for brain connectivity modeling. In particular, we apply SICE to learn and analyze functional brain connectivity patterns from different subject groups, based on a key property of SICE, called the “monotone property” we established in this paper. Our experimental results on neuroimaging PET data of 42 AD, 116 MCI, and 67 NC subjects reveal several interesting connectivity patterns consistent with literature findings, and also some new patterns that can help the knowledge discovery of AD.

Effective diagnosis of Alzheimer's disease (AD) is of primary importance in biomedical research. Recent studies have demonstrated that neuroimaging parameters are sensitive and consistent measures of AD. In addition, genetic and demographic information have also been successfully used for detecting the onset and progression of AD. The research so far has mainly focused on studying one type of data source only. It is expected that the integration of heterogeneous data (neuroimages, demographic, and genetic measures) will improve the prediction accuracy and enhance knowledge discovery from the data, such as the detection of biomarkers. In this paper, we propose to integrate heterogeneous data for AD prediction based on a kernel method. We further extend the kernel framework for selecting features (biomarkers) from heterogeneous data sources. The proposed method is applied to a collection of MRI data from 59 normal healthy controls and 59 AD patients. The MRI data are pre-processed using tensor factorization. In this study, we treat the complementary voxel-based data and region of interest (ROI) data from MRI as two data sources, and attempt to integrate the complementary information by the proposed method. Experimental results show that the integration of multiple data sources leads to a considerable improvement in the prediction accuracy. Results also show that the proposed algorithm identifies biomarkers that play more significant roles than others in AD diagnosis.