Deep Neural Networks (DNNs), including architectures such as Vision Transformers (ViTs), have achieved remarkable success in medical imaging tasks. However, their performance typically hinges on the availability of large-scale, high-quality labeled datasets-resources that are often scarce or infeasible to obtain in medical domains. Generative Data Augmentation (GDA) offers a promising remedy by supplementing training sets with synthetic data generated via generative models like Diffusion Models (DMs). Yet, this approach introduces a critical challenge: synthetic data often contains significant noise, which can degrade the performance of classifiers trained on such augmented datasets. Prior solutions, including data selection and re-weighting techniques, often rely on access to clean metadata or pretrained external classifiers. In this work, we propose \emph{Informative Data Selection} (IDS), a principled sample re-weighting framework grounded in the Information Bottleneck (IB) principle. IDS assigns higher weights to more informative synthetic samples, thereby improving classifier performance in GDA-enhanced training for thorax disease classification. Extensive experiments demonstrate that IDS significantly outperforms existing data selection and re-weighting baselines. Our code is publicly available at \url{https://github.com/Statistical-Deep-Learning/IDS}.

In biomedical datasets pertaining to disease detection, data typically falls into two classes: healthy and diseased. The diseased cohort often exhibits inherent heterogeneity due to clinical subtyping. Although the healthy cohort is presumed to be homogeneous, it contains heterogeneities arising from inter-subject variation, which affects the effectiveness of classification. To address this issue, we propose a novel methodology for multivariate time series data that discerns a homogeneous sub-cohort of healthy samples, referred to as the 'Healthy Bio-Core' (HBC). The employment of HBC augments the discriminative capacity of classification models. The selection process for HBC integrates dynamic time warping (DTW), and the accuracy of the ROCKET (RandOm Convolutional KErnel Transform) classifier, treating the entire time series as a single instance. Empirical results indicate that utilizing HBC enhances classification performance in comparison to utilizing the complete healthy dataset. We substantiate this approach with three classifiers: HIVE-COTE (Hierarchical Vote Collective of Transformation-based Ensembles), MUSE (Multi-variate Unsupervised Symbols and Derivatives), and DTW-NN (DTW with Nearest Neighbor), assessing metrics, such as accuracy, precision, recall, and F1-score. Although our approach relies on DTW, it is limited to cases where a DTW path can be identified; otherwise, another distance metric must be used. Currently, the efficiency depends on the classifier used. Future studies might investigate combining different classifiers for HBC sample selection and devise a method to synthesize their outcomes. Moreover, assuming that the dataset is predominantly healthy may not hold true in contexts with significant noise. Notwithstanding these limitations, our approach results in significant improvements in classification, with average accuracy increases of 5.49%, 14.28%, and 6.16% for the sepsis, gait, and EMO pain datasets, respectively.

Deep neural networks, including Convolutional Neural Networks (CNNs) and Visual Transformers (ViT), have achieved stunning success in the medical image domain. We study thorax disease classification in this paper. Effective extraction of features for the disease areas is crucial for disease classification on radiographic images. While various neural architectures and training techniques, such as self-supervised learning with contrastive/restorative learning, have been employed for disease classification on radiographic images, there are no principled methods that can effectively reduce the adverse effect of noise and background or non-disease areas on the radiographic images for disease classification. To address this challenge, we propose a novel Low-Rank Feature Learning (LRFL) method in this paper, which is universally applicable to the training of all neural networks. The LRFL method is both empirically motivated by a Low Frequency Property (LFP) and theoretically motivated by our sharp generalization bound for neural networks with low-rank features. LFP not only widely exists in deep neural networks for generic machine learning but also exists in all the thorax medical datasets studied in this paper. In the empirical study, using a neural network such as a ViT or a CNN pre-trained on unlabeled chest X-rays by Masked Autoencoders (MAE), our novel LRFL method is applied on the pre-trained neural network and demonstrates better classification results in terms of both multi-class area under the receiver operating curve (mAUC) and classification accuracy than the current state-of-the-art. The code of LRFL is available at \url{https://github.com/Statistical-Deep-Learning/LRFL}.

Age is one of the major known risk factors for Alzheimer's Disease (AD). Detecting AD early is crucial for effective treatment and preventing irreversible brain damage. Brain age, a measure derived from brain imaging reflecting structural changes due to aging, may have the potential to identify AD onset, assess disease risk, and plan targeted interventions. Deep learning-based regression techniques to predict brain age from magnetic resonance imaging (MRI) scans have shown great accuracy recently. However, these methods are subject to an inherent regression to the mean effect, which causes a systematic bias resulting in an overestimation of brain age in young subjects and underestimation in old subjects. This weakens the reliability of predicted brain age as a valid biomarker for downstream clinical applications. Here, we reformulate the brain age prediction task from regression to classification to address the issue of systematic bias. Recognizing the importance of preserving ordinal information from ages to understand aging trajectory and monitor aging longitudinally, we propose a novel ORdinal Distance Encoded Regularization (ORDER) loss that incorporates the order of age labels, enhancing the model's ability to capture age-related patterns. Extensive experiments and ablation studies demonstrate that this framework reduces systematic bias, outperforms state-of-art methods by statistically significant margins, and can better capture subtle differences between clinical groups in an independent AD dataset. Our implementation is publicly available at https://github.com/jaygshah/Robust-Brain-Age-Prediction.

Harnessing the power of deep neural networks in the medical imaging domain is challenging due to the difficulties in acquiring large annotated datasets, especially for rare diseases, which involve high costs, time, and effort for annotation. Unsupervised disease detection methods, such as anomaly detection, can significantly reduce human effort in these scenarios. While anomaly detection typically focuses on learning from images of healthy subjects only, real-world situations often present unannotated datasets with a mixture of healthy and diseased subjects. Recent studies have demonstrated that utilizing such unannotated images can improve unsupervised disease and anomaly detection. However, these methods do not utilize knowledge specific to registered neuroimages, resulting in a subpar performance in neurologic disease detection. To address this limitation, we propose Brainomaly, a GAN-based image-to-image translation method specifically designed for neurologic disease detection. Brainomaly not only offers tailored image-to-image translation suitable for neuroimages but also leverages unannotated mixed images to achieve superior neurologic disease detection. Additionally, we address the issue of model selection for inference without annotated samples by proposing a pseudo-AUC metric, further enhancing Brainomaly's detection performance. Extensive experiments and ablation studies demonstrate that Brainomaly outperforms existing state-of-the-art unsupervised disease and anomaly detection methods by significant margins in Alzheimer's disease detection using a publicly available dataset and headache detection using an institutional dataset. The code is available from https://github.com/mahfuzmohammad/Brainomaly.

BACKGROUND: The active functionalities of RNA are recognized to be heavily dependent on the structure and sequence. Therefore, a model that can accurately evaluate a design by giving RNA sequence-structure pairs would be a valuable tool for many researchers. Machine learning methods have been explored to develop such tools, showing promising results. However, two key issues remain. Firstly, the performance of machine learning models is affected by the features used to characterize RNA. Currently, there is no consensus on which features are the most effective for characterizing RNA sequence-structure pairs. Secondly, most existing machine learning methods extract features describing entire RNA molecule. We argue that it is essential to define additional features that characterize nucleotides and specific sections of RNA structure to enhance the overall efficacy of the RNA design process. RESULTS: We develop two deep learning models for evaluating RNA sequence-secondary structure pairs. The first model, NU-ResNet, uses a convolutional neural network architecture that solves the aforementioned problems by explicitly encoding RNA sequence-structure information into a 3D matrix. Building upon NU-ResNet, our second model, NUMO-ResNet, incorporates additional information derived from the characterizations of RNA, specifically the 2D folding motifs. In this work, we introduce an automated method to extract these motifs based on fundamental secondary structure descriptions. We evaluate the performance of both models on an independent testing dataset. Our proposed models outperform the models from literatures in this independent testing dataset. To assess the robustness of our models, we conduct 10-fold cross validation. To evaluate the generalization ability of NU-ResNet and NUMO-ResNet across different RNA families, we train and test our proposed models in different RNA families. Our proposed models show superior performance compared to the models from literatures when being tested across different independent RNA families. CONCLUSIONS: In this study, we propose two deep learning models, NU-ResNet and NUMO-ResNet, to evaluate RNA sequence-secondary structure pairs. These two models expand the field of data-driven approaches for learning RNA. Furthermore, these two models provide the new method to encode RNA sequence-secondary structure pairs.

Existing approaches to automatic data transformation are insufficient to meet the requirements in many real-world scenarios, such as the building sector. First, there is no convenient interface for domain experts to provide domain knowledge easily. Second, they require significant training data collection overheads. Third, the accuracy suffers from complicated schema changes. To address these shortcomings, we present a novel approach that leverages the unique capabilities of large language models (LLMs) in coding, complex reasoning, and zero-shot learning to generate SQL code that transforms the source datasets into the target datasets. We demonstrate the viability of this approach by designing an LLM-based framework, termed SQLMorpher, which comprises a prompt generator that integrates the initial prompt with optional domain knowledge and historical patterns in external databases. It also implements an iterative prompt optimization mechanism that automatically improves the prompt based on flaw detection. The key contributions of this work include (1) pioneering an end-to-end LLM-based solution for data transformation, (2) developing a benchmark dataset of 105 real-world building energy data transformation problems, and (3) conducting an extensive empirical evaluation where our approach achieved 96% accuracy in all 105 problems. SQLMorpher demonstrates the effectiveness of utilizing LLMs in complex, domain-specific challenges, highlighting the potential of their potential to drive sustainable solutions.

Image synthesis is a novel solution in precision medicine for scenarios where important medical imaging is not otherwise available. The convolutional neural network (CNN) is an ideal model for this task because of its powerful learning capabilities through the large number of layers and trainable parameters. In this research, we propose a new architecture of residual inception encoder-decoder neural network (RIED-Net) to learn the nonlinear mapping between the input images and targeting output images. To evaluate the validity of the proposed approach, it is compared with two models from the literature: synthetic CT deep convolutional neural network (sCT-DCNN) and shallow CNN, using both an institutional mammogram dataset from Mayo Clinic Arizona and a public neuroimaging dataset from the Alzheimer's Disease Neuroimaging Initiative. Experimental results show that the proposed RIED-Net outperforms the two models on both datasets significantly in terms of structural similarity index, mean absolute percent error, and peak signal-to-noise ratio.

BACKGROUND: RNA molecules play many crucial roles in living systems. The spatial complexity that exists in RNA structures determines their cellular functions. Therefore, understanding RNA folding conformations, in particular, RNA secondary structures, is critical for elucidating biological functions. Existing literature has focused on RNA design as either an RNA structure prediction problem or an RNA inverse folding problem where free energy has played a key role. RESULTS: In this research, we propose a Positive-Unlabeled data- driven framework termed ENTRNA. Other than free energy and commonly studied sequence and structural features, we propose a new feature, Sequence Segment Entropy (SSE), to measure the diversity of RNA sequences. ENTRNA is trained and cross-validated using 1024 pseudoknot-free RNAs and 1060 pseudoknotted RNAs from the RNASTRAND database respectively. To test the robustness of the ENTRNA, the models are further blind tested on 206 pseudoknot-free and 93 pseudoknotted RNAs from the PDB database. For pseudoknot-free RNAs, ENTRNA has 86.5% sensitivity on the training dataset and 80.6% sensitivity on the testing dataset. For pseudoknotted RNAs, ENTRNA shows 81.5% sensitivity on the training dataset and 71.0% on the testing dataset. To test the applicability of ENTRNA to long structural-complex RNA, we collect 5 laboratory synthetic RNAs ranging from 1618 to 1790 nucleotides. ENTRNA is able to predict the foldability of 4 RNAs. CONCLUSION: In this article, we reformulate the RNA design problem as a foldability prediction problem which is to predict the likelihood of the co-existence of a sequence-structure pair. This new construct has the potential for both RNA structure prediction and the inverse folding problem. In addition, this new construct enables us to explore data-driven approaches in RNA research.

This paper presents a Coevolutionary Structure-Redesigned-Based Bacteria Foraging Optimization (CSRBFO) based on the natural phenomenon that most living creatures tend to cooperate with each other so as to fulfill tasks more effectively. Aiming at lowering computational complexity while maintaining the critical search capability of standard bacterial foraging optimization (BFO), we employ a general loop to replace the nested loop and eliminate the reproduction step of BFO. Hence, the proposed CSRBFO only consists of two main steps: (1) chemotaxis and (2) elimination & dispersal. A coevolutionary strategy by which all bacteria can learn from each other and search for optima cooperatively is incorporated into the chemotactic step to accelerate convergence and facilitate accurate search. In the elimination & dispersal step, the three-stage evolutionary strategy with different learning methods for maintaining diversity is studied. An evaluation of the convergence status is then added to determine whether bacteria should move on to the next stage or not. The combination of coevolutionary strategy and convergence status evaluation is expected to balance exploration and exploitation. Experimental results comparing seven well-known heuristic algorithms on 24 benchmark functions demonstrate that the proposed CSRBFO outperforms the comparison algorithms significantly in most of the cases.

The identification of small structures (blobs) from medical images to quantify clinically relevant features, such as size and shape, is important in many medical applications. One particular application explored here is the automated detection of kidney glomeruli after targeted contrast enhancement and magnetic resonance imaging. We propose a computationally efficient algorithm, termed the Hessian-based Difference of Gaussians (HDoG), to segment small blobs (e.g., glomeruli from kidney) from 3D medical images based on local convexity, intensity and shape information. The image is first smoothed and pre-segmented into small blob candidate regions based on local convexity. Two novel 3D regional features (regional blobness and regional flatness) are then extracted from the candidate regions. Together with regional intensity, the three features are used in an unsupervised learning algorithm for auto post-pruning. HDoG is first validated in a 2D form and compared with other three blob detectors from literature, which are generally for 2D images only. To test the detectability of blobs from 3D images, 240 sets of simulated images are rendered for scenarios mimicking the renal nephron distribution observed in contrast-enhanced, 3D MRI. The results show a satisfactory performance of HDoG in detecting large numbers of small blobs. Two sets of real kidney 3D MR images (6 rats, 3 human) are then used to validate the applicability of HDoG for glomeruli detection. By comparing MRI to stereological measurements, we verify that HDoG is a robust and efficient unsupervised technique for 3D blobs segmentation.

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.

Particle swarm optimization (PSO) has attracted much attention and has been applied to many scientific and engineering applications in the last decade. Most recently, an intelligent augmented particle swarm optimization with multiple adaptive methods (PSO-MAM) was proposed and was demonstrated to be effective for diverse functions. However, inherited from PSO, the performance of PSO-MAM heavily depends on the settings of three parameters: the two learning factors and the inertia weight. In this paper, we propose a parameter control mechanism to adaptively change the parameters and thus improve the robustness of PSO-MAM. A new method, adaptive PSO-MAM (APSO-MAM) is developed that is expected to be more robust than PSO-MAM. We comprehensively evaluate the performance of APSO-MAM by comparing it with PSO-MAM and several state-of-the-art PSO algorithms and evolutionary algorithms. The proposed parameter control method is also compared with several existing parameter control methods. The experimental results demonstrate that APSO-MAM outperforms the compared PSO algorithms and evolutionary algorithms, and is more robust than PSO-MAM.

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.

Support Vector Machine (SVM) is a classifier designed to achieve optimized classification accuracy . It has been applied to numerous applications associated with images. Yet challenges remain when applying SVM on segmenting mouse brain images. This is due to the fact that each high-resolution mouse brain image is a very large data set and it is a multiclass classification problem with extremely imbalanced data size for different classes. To address these issues, a mix-ratio sampling approach for SVM is proposed which determines various over-sampling ratios for different minority classes. In addition, to improve the imaging classification accuracy , spatial information is incorporated into the classification problem. Five mouse Magnetic Resonance Microscopy (MRM) images are collected to test the accuracy of classifying 21 brain structures. The first comparison experiment demonstrates the SVM with mix-ratio sampling method relieves the imbalance problem for multiclass more effectively and efficiently than the SVM with simple over-sampling method. In the second comparison experiment, another classifier, Artificial Neural Network (ANN) is used to compare against SVM based on the same mix-ratio sampled data and the results indicate that SVM shows better classification performance than ANN. Thirdly, the cross validation is conducted to demonstrate SVM with mix-ration sampling can classify multiclass imbalanced data with high accuracy.

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.

Electronic messaging protocols such as RosettaNet (RN) automate the asynchronous exchange of business documents between collaborating trading partners over the Internet. Such protocols commonly employ mechanisms such as retries, time-outs, and fault handling to overcome uncertainty in the timing and reliability of message transmission, receipt, and processing. To determine the reliability and performance of such a protocol under varying network and message processing conditions, we have developed reusable patterns for Petri net modeling of these common mechanisms and used them to assemble a timed Petri net simulation that represents the RN standard's behavior. The reusable patterns are derived through translation of the abstract representations provided in the RN standard into an executable model, paving the way for a multimodeling approach for supply chain communications. The resulting stochastic Petri net is simulated to generate performance curves guiding improved protocol reliability.

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.