OBJECTIVE: Extracting social determinants of health (SDoHs) from medical notes depends heavily on labor-intensive annotations, which are typically task-specific, hampering reusability and limiting sharing. Here, we introduce SDoH-GPT, a novel framework leveraging few-shot learning large language models (LLMs) to automate the extraction of SDoH from unstructured text, aiming to improve both efficiency and generalizability. MATERIALS AND METHODS: SDoH-GPT is a framework including the few-shot learning LLM methods to extract the SDoH from medical notes and the XGBoost classifiers which continue to classify SDoH using the annotations generated by the few-shot learning LLM methods as training datasets. The unique combination of the few-shot learning LLM methods with XGBoost utilizes the strength of LLMs as great few shot learners and the efficiency of XGBoost when the training dataset is sufficient. Therefore, SDoH-GPT can extract SDoH without relying on extensive medical annotations or costly human intervention. RESULTS: Our approach achieved tenfold and twentyfold reductions in time and cost, respectively, and superior consistency with human annotators measured by Cohen's kappa of up to 0.92. The innovative combination of LLM and XGBoost can ensure high accuracy and computational efficiency while consistently maintaining 0.90+ AUROC scores. DISCUSSION: This study has verified SDoH-GPT on three datasets and highlights the potential of leveraging LLM and XGBoost to revolutionize medical note classification, demonstrating its capability to achieve highly accurate classifications with significantly reduced time and cost. CONCLUSION: The key contribution of this study is the integration of LLM with XGBoost, which enables cost-effective and high quality annotations of SDoH. This research sets the stage for SDoH can be more accessible, scalable, and impactful in driving future healthcare solutions.

While large-scale human motion capture datasets have advanced human motion generation, modeling and generating dynamic 3D human-object interactions (HOIs) remain challenging due to dataset limitations. Existing datasets often lack extensive, high-quality motion and annotation and exhibit artifacts such as contact penetration, floating, and incorrect hand motions. To address these issues, we introduce InterAct, a large-scale 3D HOI benchmark featuring dataset and methodological advancements. First, we consolidate and standardize 21.81 hours of HOI data from diverse sources, enriching it with detailed textual annotations. Second, we propose a unified optimization framework to enhance data quality by reducing artifacts and correcting hand motions. Leveraging the principle of contact invariance, we maintain human-object relationships while introducing motion variations, expanding the dataset to 30.70 hours. Third, we define six benchmarking tasks and develop a unified HOI generative modeling perspective, achieving state-of-the-art performance. Extensive experiments validate the utility of our dataset as a foundational resource for advancing 3D human-object interaction generation. The dataset will be publicly accessible to support further research in the field.

Proficient multilingual speakers often intentionally switch languages in the middle of a conversation.Similarly, recent reasoning-focused bilingual large language models (LLMs) with strong capabilities in both languages exhibit language mixing-alternating languages within their chain of thought.Discouraging this behavior in DeepSeek-R1 was found to degrade accuracy, suggesting that language mixing may benefit reasoning.In this work, we study language switching in Chinese-English bilingual reasoning models.We identify reinforcement learning with verifiable rewards (RLVR) as the critical training stage that leads to language mixing.We show that language mixing can enhance reasoning: enforcing monolingual decoding reduces accuracy by 5.6 percentage points on MATH500.Additionally, a lightweight probe can be trained to predict whether a potential language switch would benefit or harm reasoning, and when used to guide decoding, increases accuracy by 2.92 percentage points.Our findings suggest that language mixing is not merely a byproduct of multilingual training, but is a strategic reasoning behavior.1

Medical multi-modal learning requires an effective fusion capability of various heterogeneous modalities. One vital challenge is how to effectively fuse modalities when their data quality varies across different modalities and patients. For example, in the TCGA benchmark, the performance of the same modality can differ between types of cancer. Moreover, data collected at different times, locations, and with varying reagents can introduce inter-modal data quality differences ($i.e.$, $\textbf{Modality Batch Effect}$). In response, we propose ${\textbf{A}}$daptive ${\textbf{M}}$odality Token Re-Balan${\textbf{C}}$ing ($\texttt{AMC}$), a novel top-down dynamic multi-modal fusion approach. The core of $\texttt{AMC}$ is to quantify the significance of each modality (Top) and then fuse them according to the modality importance (Down). Specifically, we access the quality of each input modality and then replace uninformative tokens with inter-modal tokens, accordingly. The more important a modality is, the more informative tokens are retained from that modality. The self-attention will further integrate these mixed tokens to fuse multi-modal knowledge. Comprehensive experiments on both medical and general multi-modal datasets demonstrate the effectiveness and generalizability of $\texttt{AMC}$.

One of the key technologies for the success of Large Language Models (LLMs) is preference alignment. However, a notable side effect of preference alignment is poor calibration: while the pre-trained models are typically well-calibrated, LLMs tend to become poorly calibrated after alignment with human preferences. In this paper, we investigate why preference alignment affects calibration and how to address this issue. For the first question, we observe that the preference collapse issue in alignment undesirably generalizes to the calibration scenario, causing LLMs to exhibit overconfidence and poor calibration. To address this, we demonstrate the importance of fine-tuning with domain-specific knowledge to alleviate the overconfidence issue. To further analyze whether this affects the model’s performance, we categorize models into two regimes: calibratable and non-calibratable, defined by bounds of Expected Calibration Error (ECE). In the calibratable regime, we propose a calibration-aware fine-tuning approach to achieve proper calibration without compromising LLMs’ performance. However, as models are further fine-tuned for better performance, they enter the non-calibratable regime. For this case, we develop an EM-algorithm-based ECE regularization for the fine-tuning loss to maintain low calibration error. Extensive experiments validate the effectiveness of the proposed methods.

As enterprise codebases continue to grow in scale and complexity, the volume of lint errors far exceeds engineers' manual remediation capacity, leading to continuous accumulation of technical debt and hindered development efficiency. This paper presents BitsAI-Fix, an automated lint error remediation workflow based on Large Language Models (LLMs), designed to address this critical challenge in industrial-scale environments. BitsAI-Fix employs tree-sitter for context expansion and generates search-and-replace format patches through specially trained LLMs, followed by lint scan re-verification to output final remediation results. Additionally, our approach introduces an innovative progressive reinforcement learning (RL) training strategy that can automatically acquire verifiable training data during the project cold-start phase and continuously iterate the model by collecting online samples through feedback after system deployment. Furthermore, we designed a targeted rule-based reward mechanism that combines format rewards and correctness rewards while penalizing redundant modifications. We also propose a "code diff matching" methodology to continuously track online effectiveness. In production deployment at ByteDance, our solution has supported over 5,000 engineers, resolved more than 12,000 static analysis issues, achieved approximately 85% remediation accuracy, with around 1,000 weekly active adopters. This work demonstrates the practical feasibility of LLM-based code remediation solutions in enterprise environments and serves as a reference for automated code fix in large-scale industrial scenarios.

Jinhao Duan, Xinyu Zhao, Zhuoxuan Zhang, Eunhye Grace Ko, Lily Boddy, Chenan Wang, Tianhao Li, Alexander Rasgon, Junyuan Hong, Min Kyung Lee, Chenxi Yuan, Qi Long, Ying Ding, Tianlong Chen, Kaidi Xu. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Large language models (LLMs) raise concerns about content authenticity and integrity because they can generate human-like text at scale. Text watermarks, which embed detectable statistical signals into generated text, offer a provable way to verify content origin. Many detection methods rely on pivotal statistics that are i.i.d. under human-written text, making goodness-of-fit (GoF) tests a natural tool for watermark detection. However, GoF tests remain largely underexplored in this setting. In this paper, we systematically evaluate eight GoF tests across three popular watermarking schemes, using three open-source LLMs, two datasets, various generation temperatures, and multiple post-editing methods. We find that general GoF tests can improve both the detection power and robustness of watermark detectors. Notably, we observe that text repetition, common in low-temperature settings, gives GoF tests a unique advantage not exploited by existing methods. Our results highlight that classic GoF tests are a simple yet powerful and underused tool for watermark detection in LLMs.

Differentially private (DP) decentralized Federated Learning (FL) allows local users to collaborate without sharing their data with a central server. However, accurately quantifying the privacy budget of private FL algorithms is challenging due to the co-existence of complex algorithmic components such as decentralized communication and local updates. This paper addresses privacy accounting for two decentralized FL algorithms within the $f$-differential privacy ($f$-DP) framework. We develop two new $f$-DP–based accounting methods tailored to decentralized settings: Pairwise Network $f$-DP (PN-$f$-DP), which quantifies privacy leakage between user pairs under random-walk communication, and Secret-based $f$-Local DP (Sec-$f$-LDP), which supports structured noise injection via shared secrets. By combining tools from $f$-DP theory and Markov chain concentration, our accounting framework captures privacy amplification arising from sparse communication, local iterations, and correlated noise. Experiments on synthetic and real datasets demonstrate that our methods yield consistently tighter $(\epsilon, \delta)$ bounds and improved utility compared to Rényi DP–based approaches, illustrating the benefits of $f$-DP in decentralized privacy accounting.

Tensor Canonical Correlation Analysis (TCCA) is a commonly employed statistical method utilized to examine linear associations between two sets of tensor datasets. However, the existing TCCA models fail to adequately address the heterogeneity present in real-world tensor data, such as brain imaging data collected from diverse groups characterized by factors like sex and race. Consequently, these models may yield biased outcomes. In order to surmount this constraint, we propose a novel approach called Multi-Group TCCA (MG-TCCA), which enables the joint analysis of multiple subgroups. By incorporating a dual sparsity structure and a block coordinate ascent algorithm, our MG-TCCA method effectively addresses heterogeneity and leverages information across different groups to identify consistent signals. This novel approach facilitates the quantification of shared and individual structures, reduces data dimensionality, and enables visual exploration. To empirically validate our approach, we conduct a study focused on investigating correlations between two brain positron emission tomography (PET) modalities (AV-45 and FDG) within an Alzheimer's disease (AD) cohort. Our results demonstrate that MG-TCCA surpasses traditional TCCA and Sparse TCCA (STCCA) in identifying sex-specific cross-modality imaging correlations. This heightened performance of MG-TCCA provides valuable insights for the characterization of multimodal imaging biomarkers in AD.

Missing values are prevalent in temporal electronic health records (EHR) data and are known to complicate data analysis and lead to biased results. The current state-of-the-art (SOTA) models for imputing missing values in EHR primarily leverage correlations across time points and across features, which perform well when data have strong correlation across time points, such as in ICU data where high-frequency time series data are collected. However, this is often insufficient for temporal EHR data from non-ICU settings (e.g., outpatient visits for primary care or specialty care), where data are collected at substantially sparser time points, resulting in much weaker correlation across time points. To address this methodological gap, we propose the Similarity-Aware Diffusion Model-Based Imputation (SADI), a novel imputation method that leverages the diffusion model and utilizes information across dependent variables. We apply SADI to impute incomplete temporal EHR data and propose a similarity-aware denoising function, which includes a self-attention mechanism to model the correlations between time points, features, and similar patients. To the best of our knowledge, this is the first time that the information of similar patients is directly used to construct imputation for incomplete temporal EHR data. Our extensive experiments on two datasets, the Critical Path For Alzheimer’s Disease (CPAD) data and the PhysioNet Challenge 2012 data, show that SADI outperforms the current SOTA under various missing data mechanisms, including missing completely at random (MCAR), missing at random (MAR), and missing not at random (MNAR).

techniques that often evade existing detection methods. To address this challenge, we present a novel semantic-enhanced framework for detecting LLM-generated text (SEFD) that leverages a retrieval-based mechanism to fully utilize text semantics. Our framework improves upon existing detection methods by systematically integrating retrieval-based techniques with traditional detectors, employing a carefully curated retrieval mechanism that strikes a balance between comprehensive coverage and computational efficiency. We showcase the effectiveness of our approach in sequential text scenarios common in real-world applications, such as online forums and Q&A platforms. Through comprehensive experiments across various LLM-generated texts and detection methods, we demonstrate that our framework substantially enhances detection accuracy in paraphrasing scenarios while maintaining robustness for standard LLM-generated content. This work contributes significantly to ongoing efforts to safeguard information integrity in an era where AI-generated content is increasingly prevalent.

Training Deep Neural Networks (DNNs) with adversarial examples often results in poor generalization to test-time adversarial data. This paper investigates this issue, known as adversarially robust generalization, through the lens of Rademacher complexity. Building upon the studies by Khim and Loh (2018); Yin et al. (2019), numerous works have been dedicated to this problem, yet achieving a satisfactory bound remains an elusive goal. Existing works on DNNs either apply to a surrogate loss instead of the robust loss or yield bounds that are notably looser compared to their standard counterparts. In the latter case, the bounds have a higher dependency on the width $m$ of the DNNs or the dimension $d$ of the data, with an extra factor of at least $\mathcal{O}(\sqrt{m})$ or $\mathcal{O}(\sqrt{d})$. This paper presents upper bounds for adversarial Rademacher complexity of DNNs that match the best-known upper bounds in standard settings, as established in the work of Bartlett et al. (2017), with the dependency on width and dimension being $\mathcal{O}(\ln(dm))$. The central challenge addressed is calculating the covering number of adversarial function classes. We aim to construct a new cover that possesses two properties: 1) compatibility with adversarial examples, and 2) precision comparable to covers used in standard settings. To this end, we introduce a new variant of covering number called the \emph{uniform covering number}, specifically designed and proven to reconcile these two properties. Consequently, our method effectively bridges the gap between Rademacher complexity in robust and standard generalization.

Designing faithful yet accurate AI models is challenging, particularly in the field of individual treatment effect estimation (ITE). ITE prediction models deployed in critical settings such as healthcare should ideally be (i) accurate, and (ii) provide faithful explanations. However, current solutions are inadequate: state-of-the-art black-box models do not supply explanations, post-hoc explainers for black-box models lack faithfulness guarantees, and self-interpretable models greatly compromise accuracy. To address these issues, we propose DISCRET, a self-interpretable ITE framework that synthesizes faithful, rule-based explanations for each sample. A key insight behind DISCRET is that explanations can serve dually as database queries to identify similar subgroups of samples. We provide a novel RL algorithm to efficiently synthesize these explanations from a large search space. We evaluate DISCRET on diverse tasks involving tabular, image, and text data. DISCRET outperforms the best self-interpretable models and has accuracy comparable to the best black-box models while providing faithful explanations. DISCRET is available at https://github.com/wuyinjun-1993/DISCRET-ICML2024.

Multimodal learning has gained increasing importance across various fields, offering the ability to integrate data from diverse sources such as images, text, and personalized records, which are frequently observed in medical domains. However, in scenarios where some modalities are missing, many existing frameworks struggle to accommodate arbitrary modality combinations, often relying heavily on a single modality or complete data. This oversight of potential modality combinations limits their applicability in real-world situations. To address this challenge, we propose Flex-MoE (Flexible Mixture-of-Experts), a new framework designed to flexibly incorporate arbitrary modality combinations while maintaining robustness to missing data. The core idea of Flex-MoE is to first address missing modalities using a new missing modality bank that integrates observed modality combinations with the corresponding missing ones. This is followed by a uniquely designed Sparse MoE framework. Specifically, Flex-MoE first trains experts using samples with all modalities to inject generalized knowledge through the generalized router ($\mathcal{G}$-Router). The $\mathcal{S}$-Router then specializes in handling fewer modality combinations by assigning the top-1 gate to the expert corresponding to the observed modality combination. We evaluate Flex-MoE on the ADNI dataset, which encompasses four modalities in the Alzheimer's Disease domain, as well as on the MIMIC-IV dataset. The results demonstrate the effectiveness of Flex-MoE, highlighting its ability to model arbitrary modality combinations in diverse missing modality scenarios. Code is available at: \url{https://github.com/UNITES-Lab/flex-moe}.

This paper examines the issue of fairness in the estimation of graphical models (GMs), particularly Gaussian, Covariance, and Ising models. These models play a vital role in understanding complex relationships in high-dimensional data. However, standard GMs can result in biased outcomes, especially when the underlying data involves sensitive characteristics or protected groups. To address this, we introduce a comprehensive framework designed to reduce bias in the estimation of GMs related to protected attributes. Our approach involves the integration of the pairwise graph disparity error and a tailored loss function into a nonsmooth multi-objective optimization problem, striving to achieve fairness across different sensitive groups while maintaining the effectiveness of the GMs. Experimental evaluations on synthetic and real-world datasets demonstrate that our framework effectively mitigates bias without undermining GMs' performance.

Epidemiologic and genetic studies in many complex diseases suggest subgroup disparities (e.g. by sex, race) in disease course and patient outcomes. We consider this from the standpoint of integrative analysis where we combine information from different views (e.g. genomics, proteomics, clinical data). Existing integrative analysis methods ignore the heterogeneity in subgroups, and stacking the views and accounting for subgroup heterogeneity does not model the association among the views. We propose Heterogeneity in Integration and Prediction (HIP), a statistical approach for joint association and prediction that leverages the strengths in each view to identify molecular signatures that are shared by and specific to a subgroup. We apply HIP to proteomics and gene expression data pertaining to chronic obstructive pulmonary disease (COPD) to identify proteins and genes shared by, and unique to, males and females, contributing to the variation in COPD, measured by airway wall thickness. Our COPD findings have identified proteins, genes, and pathways that are common across and specific to males and females, some implicated in COPD, while others could lead to new insights into sex differences in COPD mechanisms. HIP accounts for subgroup heterogeneity in multi-view data, ranks variables based on importance, is applicable to univariate or multivariate continuous outcomes, and incorporates covariate adjustment. With the efficient algorithms implemented using PyTorch, this method has many potential scientific applications and could enhance multiomics research in health disparities. HIP is available at https://github.com/lasandrall/HIP, a video tutorial at https://youtu.be/O6E2OLmeMDo and a Shiny Application at https://multi-viewlearn.shinyapps.io/HIP_ShinyApp/ for users with limited programming experience.

Biclustering is a useful method for simultaneously grouping samples and features and has been applied across various biomedical data types. However, most existing biclustering methods lack the ability to integratively analyze multi-modal data such as multi-omics data such as genome, transcriptome and epigenome. Moreover, the potential of leveraging biological knowledge represented by graphs, which has been demonstrated to be beneficial in various statistical tasks such as variable selection and prediction, remains largely untapped in the context of biclustering. To address both, we propose a novel Bayesian biclustering method called Bayesian graph-guided biclustering (BGB). Specifically, we introduce a new hierarchical sparsity-inducing prior to effectively incorporate biological graph information and establish a unified framework to model multi-view data. We develop an efficient Markov chain Monte Carlo algorithm to conduct posterior sampling and inference. Extensive simulations and real data analysis show that BGB outperforms other popular biclustering methods. Notably, BGB is robust in terms of utilizing biological knowledge and has the capability to reveal biologically meaningful information from heterogeneous multi-modal data.

This paper investigates fairness and bias in Canonical Correlation Analysis (CCA), a widely used statistical technique for examining the relationship between two sets of variables. We present a framework that alleviates unfairness by minimizing the correlation disparity error associated with protected attributes. Our approach enables CCA to learn global projection matrices from all data points while ensuring that these matrices yield comparable correlation levels to group-specific projection matrices. Experimental evaluation on both synthetic and real-world datasets demonstrates the efficacy of our method in reducing correlation disparity error without compromising CCA accuracy.

We present a novel Bayesian-based optimization framework that addresses the challenge of generalization in overparameterized models when dealing with imbalanced subgroups and limited samples per subgroup. Our proposed tri-level optimization framework utilizes local predictors, which are trained on a small amount of data, as well as a fair and class-balanced predictor at the middle and lower levels. To effectively overcome saddle points for minority classes, our lower-level formulation incorporates sharpness-aware minimization. Meanwhile, at the upper level, the framework dynamically adjusts the loss function based on validation loss, ensuring a close alignment between the global predictor and local predictors. Theoretical analysis demonstrates the framework’s ability to enhance classification and fairness generalization, potentially resulting in improvements in the generalization bound. Empirical results validate the superior performance of our tri-level framework compared to existing state-of-the-art approaches. The source code can be found at \url{https://github.com/PennShenLab/FACIMS}.

MOTIVATION: With the rapid development of modern technologies, massive data are available for the systematic study of Alzheimer's disease (AD). Though many existing AD studies mainly focus on single-modality omics data, multi-omics datasets can provide a more comprehensive understanding of AD. To bridge this gap, we proposed a novel structural Bayesian factor analysis framework (SBFA) to extract the information shared by multi-omics data through the aggregation of genotyping data, gene expression data, neuroimaging phenotypes and prior biological network knowledge. Our approach can extract common information shared by different modalities and encourage biologically related features to be selected, guiding future AD research in a biologically meaningful way. METHOD: Our SBFA model decomposes the mean parameters of the data into a sparse factor loading matrix and a factor matrix, where the factor matrix represents the common information extracted from multi-omics and imaging data. Our framework is designed to incorporate prior biological network information. Our simulation study demonstrated that our proposed SBFA framework could achieve the best performance compared with the other state-of-the-art factor-analysis-based integrative analysis methods. RESULTS: We apply our proposed SBFA model together with several state-of-the-art factor analysis models to extract the latent common information from genotyping, gene expression and brain imaging data simultaneously from the ADNI biobank database. The latent information is then used to predict the functional activities questionnaire score, an important measurement for diagnosis of AD quantifying subjects' abilities in daily life. Our SBFA model shows the best prediction performance compared with the other factor analysis models. AVAILABILITY: Code are publicly available at https://github.com/JingxuanBao/SBFA. CONTACT: [email protected].

A widely recognized difficulty in federated learning arises from the statistical heterogeneity among clients: local datasets often originate from distinct yet not entirely unrelated probability distributions, and personalization is, therefore, necessary to achieve optimal results from each individual’s perspective. In this paper, we show how the excess risks of personalized federated learning using a smooth, strongly convex loss depend on data heterogeneity from a minimax point of view, with a focus on the FedAvg algorithm (McMahan et al., 2017) and pure local training (i.e., clients solve empirical risk minimization problems on their local datasets without any communication). Our main result reveals an approximate alternative between these two baseline algorithms for federated learning: the former algorithm is minimax rate optimal over a collection of instances when data heterogeneity is small, whereas the latter is minimax rate optimal when data heterogeneity is large, and the threshold is sharp up to a constant. As an implication, our results show that from a worst-case point of view, a dichotomous strategy that makes a choice between the two baseline algorithms is rate-optimal. Another implication is that the popular FedAvg following by local fine tuning strategy is also minimax optimal under additional regularity conditions. Our analysis relies on a new notion of algorithmic stability that takes into account the nature of federated learning.

Investigating the relationship between genetic variation and phenotypic traits is a key issue in quantitative genetics. Specifically for Alzheimer's disease, the association between genetic markers and quantitative traits remains vague while, once identified, will provide valuable guidance for the study and development of genetic-based treatment approaches. Currently, to analyze the association of two modalities, sparse canonical correlation analysis (SCCA) is commonly used to compute one sparse linear combination of the variable features for each modality, giving a pair of linear combination vectors in total that maximizes the cross-correlation between the analyzed modalities. One drawback of the plain SCCA model is that the existing findings and knowledge cannot be integrated into the model as priors to help extract interesting correlation as well as identify biologically meaningful genetic and phenotypic markers. To bridge this gap, we introduce preference matrix guided SCCA (PM-SCCA) that not only takes priors encoded as a preference matrix but also maintains computational simplicity. A simulation study and a real-data experiment are conducted to investigate the effectiveness of the model. Both experiments demonstrate that the proposed PM-SCCA model can capture not only genotype-phenotype correlation but also relevant features effectively.

Mendelian randomization is a versatile tool to identify the possible causal relationship between an omics biomarker and disease outcome using genetic variants as instrumental variables. A key theme is the prioritization of genes whose omics readouts can be used as predictors of the disease outcome through analyzing GWAS and QTL summary data. However, there is a dearth of study of the best practice in probing the effects of multiple -omics biomarkers annotated to the same gene of interest. To bridge this gap, we propose powerful combination tests that integrate multiple correlated $P$-values without assuming the dependence structure between the exposures. Our extensive simulation experiments demonstrate the superiority of our proposed approach compared with existing methods that are adapted to the setting of our interest. The top hits of the analyses of multi-omics Alzheimer's disease datasets include genes ABCA7 and ATP1B1.

BACKGROUND: In Alzheimer's Diseases (AD) research, multimodal imaging analysis can unveil complementary information from multiple imaging modalities and further our understanding of the disease. One application is to discover disease subtypes using unsupervised clustering. However, existing clustering methods are often applied to input features directly, and could suffer from the curse of dimensionality with high-dimensional multimodal data. The purpose of our study is to identify multimodal imaging-driven subtypes in Mild Cognitive Impairment (MCI) participants using a multiview learning framework based on Deep Generalized Canonical Correlation Analysis (DGCCA), to learn shared latent representation with low dimensions from 3 neuroimaging modalities. RESULTS: DGCCA applies non-linear transformation to input views using neural networks and is able to learn correlated embeddings with low dimensions that capture more variance than its linear counterpart, generalized CCA (GCCA). We designed experiments to compare DGCCA embeddings with single modality features and GCCA embeddings by generating 2 subtypes from each feature set using unsupervised clustering. In our validation studies, we found that amyloid PET imaging has the most discriminative features compared with structural MRI and FDG PET which DGCCA learns from but not GCCA. DGCCA subtypes show differential measures in 5 cognitive assessments, 6 brain volume measures, and conversion to AD patterns. In addition, DGCCA MCI subtypes confirmed AD genetic markers with strong signals that existing late MCI group did not identify. CONCLUSION: Overall, DGCCA is able to learn effective low dimensional embeddings from multimodal data by learning non-linear projections. MCI subtypes generated from DGCCA embeddings are different from existing early and late MCI groups and show most similarity with those identified by amyloid PET features. In our validation studies, DGCCA subtypes show distinct patterns in cognitive measures, brain volumes, and are able to identify AD genetic markers. These findings indicate the promise of the imaging-driven subtypes and their power in revealing disease structures beyond early and late stage MCI.

Federated learning (FL) is a training paradigm where the clients collaboratively learn models by repeatedly sharing information without compromising much on the privacy of their local sensitive data. In this paper, we introduce \emph{federated $f$-differential privacy}, a new notion specifically tailored to the federated setting, based on the framework of Gaussian differential privacy. Federated $f$-differential privacy operates on \emph{record level}: it provides the privacy guarantee on each individual record of one client’s data against adversaries. We then propose a generic private federated learning framework \fedsync that accommodates a large family of state-of-the-art FL algorithms, which provably achieves {federated $f$-differential privacy}. Finally, we empirically demonstrate the trade-off between privacy guarantee and prediction performance for models trained by \fedsync in computer vision tasks.

Alzheimer's disease (AD) is a progressive neurodegenerative brain disorder characterized by memory loss and cognitive decline. Early detection and accurate prognosis of AD is an important research topic, and numerous machine learning methods have been proposed to solve this problem. However, traditional machine learning models are facing challenges in effectively integrating longitudinal neuroimaging data and biologically meaningful structure and knowledge to build accurate and interpretable prognostic predictors. To bridge this gap, we propose an interpretable graph neural network (GNN) model for AD prognostic prediction based on longitudinal neuroimaging data while embracing the valuable knowledge of structural brain connectivity. In our empirical study, we demonstrate that 1) the proposed model outperforms several competing models (i.e., DNN, SVM) in terms of prognostic prediction accuracy, and 2) our model can capture neuroanatomical contribution to the prognostic predictor and yield biologically meaningful interpretation to facilitate better mechanistic understanding of the Alzheimer's disease. Source code is available at https://github.com/JaesikKim/temporal-GNN.

Support vector machine (SVM) is a popular classification method for the analysis of a wide range of data including big biomedical data. Many SVM methods with feature selection have been developed under the frequentist regularization or Bayesian shrinkage frameworks. On the other hand, the value of incorporating a priori known biological knowledge, such as those from functional genomics and functional proteomics, into statistical analysis of -omic data has been recognized in recent years. Such biological information is often represented by graphs. We propose a novel method that assigns Laplace priors to the regression coefficients and incorporates the underlying graph information via a hyper-prior for the shrinkage parameters in the Laplace priors. This enables smoothing of shrinkage parameters for connected variables in the graph and conditional independence between shrinkage parameters for disconnected variables. Extensive simulations demonstrate that our proposed methods achieve the best performance compared to the other existing SVM methods in terms of prediction accuracy. The proposed method are also illustrated in analysis of genomic data from cancer studies, demonstrating its advantage in generating biologically meaningful results and identifying potentially important features.

Missing data are present in most real world problems and need careful handling to preserve the prediction accuracy and statistical consistency in the downstream analysis. As the gold standard of handling missing data, multiple imputation (MI) methods are proposed to account for the imputation uncertainty and provide proper statistical inference. In this work, we propose Multiple Imputation via Generative Adversarial Network (MI-GAN), a deep learning-based (in specific, a GAN-based) multiple imputation method, that can work under missing at random (MAR) mechanism with theoretical support. MI-GAN leverages recent progress in conditional generative adversarial neural works and shows strong performance matching existing state-of-the-art imputation methods on high-dimensional datasets, in terms of imputation error. In particular, MI-GAN significantly outperforms other imputation methods in the sense of statistical inference and computational speed.

Missing data are prevalent and present daunting challenges in real data analysis. While there is a growing body of literature on fairness in analysis of fully observed data, there has been little theoretical work on investigating fairness in analysis of incomplete data. In practice, a popular analytical approach for dealing with missing data is to use only the set of complete cases, i.e., observations with all features fully observed to train a prediction algorithm. However, depending on the missing data mechanism, the distribution of complete cases and the distribution of the complete data may be substantially different. When the goal is to develop a fair algorithm in the complete data domain where there are no missing values, an algorithm that is fair in the complete case domain may show disproportionate bias towards some marginalized groups in the complete data domain. To fill this significant gap, we study the problem of estimating fairness in the complete data domain for an arbitrary model evaluated merely using complete cases. We provide upper and lower bounds on the fairness estimation error and conduct numerical experiments to assess our theoretical results. Our work provides the first known theoretical results on fairness guarantee in analysis of incomplete data.

We present an effective deep multiview learning framework to identify population structure using multimodal imaging data. Our approach is based on canonical correlation analysis (CCA). We propose to use deep generalized CCA (DGCCA) to learn a shared latent representation of non-linearly mapped and maximally correlated components from multiple imaging modalities with reduced dimensionality. In our empirical study, this representation is shown to effectively capture more variance in original data than conventional generalized CCA (GCCA) which applies only linear transformation to the multi-view data. Furthermore, subsequent cluster analysis on the new feature set learned from DGCCA is able to identify a promising population structure in an Alzheimer's disease (AD) cohort. Genetic association analyses of the clustering results demonstrate that the shared representation learned from DGCCA yields a population structure with a stronger genetic basis than several competing feature learning methods.

Support vector machine (SVM) is a powerful classification tool for analysis of high dimensional data such as genomics. Regularized linear and nonlinear SVM methods with feature selection have been developed. On the other hand, there is a growing body of literature showing that incorporating prior biological knowledge such as functional genomics, which are typically represented by graphs, into the analysis of genomic data can improve feature selection and prediction. In practice, however, such biological knowledge can often be inaccurate or unavailable. To attack this problem, we propose a Bayesian modeling approach which enables us to learn the graph structure among features and perform feature selection simultaneously. Our approach employs a Gaussian graphical model for inferring the graphical information and exploits the inferred graph to guide feature selection for SVM. An efficient MCMC algorithm is developed and our numerical analysis demonstrates that the proposed method has advantages over existing methods in feature selection and prediction via simulations and an application to the analysis of glioblastoma patient data.

Datasets containing sensitive information are often sequentially analyzed by many algorithms and, accordingly, a fundamental question in differential privacy is concerned with how the overall privacy bound degrades under composition. To address this question, we introduce a family of analytical and sharp privacy bounds under composition using the Edgeworth expansion in the framework of the recently proposed $f$-differential privacy. In short, whereas the existing composition theorem, for example, relies on the central limit theorem, our new privacy bounds under composition gain improved tightness by leveraging the refined approximation accuracy of the Edgeworth expansion. Our approach is easy to implement and computationally efficient for any number of compositions. The superiority of these new bounds is confirmed by an asymptotic error analysis and an application to quantifying the overall privacy guarantees of noisy stochastic gradient descent used in training private deep neural networks.

Integrative analysis jointly analyzes multiple data sets to overcome curse of dimensionality. It can detect important but weak signals by jointly selecting features for all data sets, but unfortunately the sets of important features are not always the same for all data sets. Variations which allows heterogeneous sparsity structure-a subset of data sets can have a zero coefficient for a selected feature-have been proposed, but it compromises the effect of integrative analysis recalling the problem of losing weak important signals. We propose a new integrative analysis approach which not only aggregates weak important signals well in homogeneity setting but also substantially alleviates the problem of losing weak important signals in heterogeneity setting. Our approach exploits a priori known graphical structure of features by forcing joint selection of adjacent features, and integrating such information over multiple data sets can increase the power while taking into account the heterogeneity across data sets. We confirm the problem of existing approaches and demonstrate the superiority of our method through a simulation study and an application to gene expression data from ADNI.

BACKGROUND: Multiple co-inertia analysis (mCIA) is a multivariate analysis method that can assess relationships and trends in multiple datasets. Recently it has been used for integrative analysis of multiple high-dimensional -omics datasets. However, its estimated loading vectors are non-sparse, which presents challenges for identifying important features and interpreting analysis results. We propose two new mCIA methods: 1) a sparse mCIA method that produces sparse loading estimates and 2) a structured sparse mCIA method that further enables incorporation of structural information among variables such as those from functional genomics. RESULTS: Our extensive simulation studies demonstrate the superior performance of the sparse mCIA and structured sparse mCIA methods compared to the existing mCIA in terms of feature selection and estimation accuracy. Application to the integrative analysis of transcriptomics data and proteomics data from a cancer study identified biomarkers that are suggested in the literature related with cancer disease. CONCLUSION: Proposed sparse mCIA achieves simultaneous model estimation and feature selection and yields analysis results that are more interpretable than the existing mCIA. Furthermore, proposed structured sparse mCIA can effectively incorporate prior network information among genes, resulting in improved feature selection and enhanced interpretability.

Support vector machine (SVM) is a popular classification method for analysis of high dimensional data such as genomics data. Recently a number of linear SVM methods have been developed to achieve feature selection through either frequentist regularization or Bayesian shrinkage, but the linear assumption may not be plausible for many real applications. In addition, recent work has demonstrated that incorporating known biological knowledge, such as those from functional genomics, into the statistical analysis of genomic data offers great promise of improved predictive accuracy and feature selection. Such biological knowledge can often be represented by graphs. In this article, we propose a novel knowledge-guided nonlinear Bayesian SVM approach for analysis of high-dimensional data. Our model uses graph information that represents the relationship among the features to guide feature selection. To achieve knowledge-guided feature selection, we assign an Ising prior to the indicators representing inclusion/exclusion of the features in the model. An efficient MCMC algorithm is developed for posterior inference. The performance of our method is evaluated and compared with several penalized linear SVM and the standard kernel SVM method in terms of prediction and feature selection in extensive simulation studies. Also, analyses of genomic data from a cancer study show that our method yields a more accurate prediction model for patient survival and reveals biologically more meaningful results than the existing methods.

MOTIVATION: Co-inertia analysis (CIA) is a multivariate statistical analysis method that can assess relationships and trends in two sets of data. Recently CIA has been used for an integrative analysis of multiple high-dimensional omics data. However, for classical CIA, all elements in the loading vectors are nonzero, presenting a challenge for the interpretation when analyzing omics data. For other multivariate statistical methods such as canonical correlation analysis (CCA), penalized least squares (PLS), various approaches have been proposed to produce sparse loading vectors via l1-penalization/constraint. We propose a novel CIA method that uses l1-penalization to induce sparsity in estimators of loading vectors. Our method simultaneously conducts model fitting and variable selection. Also, we propose another CIA method that incorporates structure/network information such as those from functional genomics, besides using sparsity penalty so that one can get biologically meaningful and interpretable results. RESULTS: Extensive simulations demonstrate that our proposed penalized CIA methods achieve the best or close to the best performance compared to the existing CIA method in terms of feature selection and recovery of true loading vectors. Also, we apply our methods to the integrative analysis of gene expression data and protein abundance data from the NCI-60 cancer cell lines. Our analysis of the NCI-60 cancer cell line data reveals meaningful variables for cancer diseases and biologically meaningful results that are consistent with previous studies. AVAILABILITY AND IMPLEMENTATION: Our algorithms are implemented as an R package which is freely available at: https://www.med.upenn.edu/long-lab/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

A key task in cancer genomics research is to identify cancer driver genes.As these genes initialise and progress cancer, understanding them is critical in designing effective cancer interventions.Although there are several methods developed to discover cancer drivers, most of them only identify coding drivers.However, non-coding RNAs can regulate driver mutations to develop cancer.Hence, novel methods are required to reveal both coding and non-coding cancer drivers.In this paper, we develop a novel framework named Controllability based Biological Network Analysis (CBNA) to uncover coding and non-coding cancer drivers (i.e.miRNA cancer drivers).CBNA integrates different genomic data types, including gene expression, gene network, mutation data, and contains a two-stage process: (1) Building a network for a condition (e.g.cancer condition) and (2) Identifying drivers.The application of CBNA to the BRCA dataset demonstrates that it is more effective than the existing methods in detecting coding cancer drivers.In addition, CBNA also predicts 17 miRNA drivers for breast cancer.Some of these predicted miRNA drivers have been validated by literature and the rest can be good candidates for wet-lab validation.We further use CBNA to detect subtype-specific cancer drivers and several predicted drivers have been confirmed to be related to breast cancer subtypes.Another application of CBNA is to discover epithelial-mesenchymal transition (EMT) drivers.Of the predicted EMT drivers, 7 coding and 6 miRNA drivers are in the known EMT gene lists. Author summaryCancer is a disease of cells in human body and it causes a high rate of deaths worldwide.There has been evidence that coding and non-coding RNAs are key players in the initialisation and progression of cancer.These coding and non-coding RNAs are considered as cancer drivers.To design better diagnostic and therapeutic plans for cancer patients, we need to know the roles of cancer drivers in cancer development as well as their regulatory

Support vector machine (SVM) is a popular classification method for the analysis of wide range of data including big data. Many SVM methods with feature selection have been developed under frequentist regularization or Bayesian shrinkage frameworks. On the other hand, the importance of incorporating a priori known biological knowledge, such as gene pathway information which stems from the gene regulatory network, into the statistical analysis of genomic data has been recognized in recent years. In this article, we propose a new Bayesian SVM approach that enables the feature selection to be guided by the knowledge on the graphical structure among predictors. The proposed method uses the spike-and-slab prior for feature selection, combined with the Ising prior that encourages group-wise selection of the predictors adjacent to each other on the known graph. Gibbs sampling algorithm is used for Bayesian inference. The performance of our method is evaluated and compared with existing SVM methods in terms of prediction and feature selection in extensive simulation settings. In addition, our method is illustrated in the analysis of genomic data from a cancer study, demonstrating its advantage in generating biologically meaningful results and identifying potentially important features.

Integrative clustering is a clustering approach for multiple datasets, which provide different views of a common group of subjects. It enables analyzing multi-omics data jointly to, for example, identify the subtypes of diseases, cells, and so on, capturing the complex underlying biological processes more precisely. On the other hand, there has been a great deal of interest in incorporating the prior structural knowledge on the features into statistical analyses over the past decade. The knowledge on the gene regulatory network (pathways) can potentially be incorporated into many genomic studies. In this paper, we propose a novel integrative clustering method which can incorporate the prior graph knowledge. We first develop a generalized Bayesian factor analysis (GBFA) framework, a sparse Bayesian factor analysis which can take into account the graph information. Our GBFA framework employs the spike and slab lasso (SSL) prior to impose sparsity on the factor loadings and the Markov random field (MRF) prior to encourage smoothing over the adjacent factor loadings, which establishes a unified shrinkage adaptive to the loading size and the graph structure. Then, we use the framework to extend iCluster+, a factor analysis based integrative clustering approach. A novel variational EM algorithm is proposed to efficiently estimate the MAP estimator for the factor loadings. Extensive simulation studies and the application to the NCI60 cell line dataset demonstrate that the propose method is superior and delivers more biologically meaningful outcomes.

Ultra-high dimensional variable selection has become increasingly important in analysis of neuroimaging data. For example, in the Autism Brain Imaging Data Exchange (ABIDE) study, neuroscientists are interested in identifying important biomarkers for early detection of the autism spectrum disorder (ASD) using high resolution brain images that include hundreds of thousands voxels. However, most existing methods are not feasible for solving this problem due to their extensive computational costs. In this work, we propose a novel multiresolution variable selection procedure under a Bayesian probit regression framework. It recursively uses posterior samples for coarser-scale variable selection to guide the posterior inference on finer-scale variable selection, leading to very efficient Markov chain Monte Carlo (MCMC) algorithms. The proposed algorithms are computationally feasible for ultra-high dimensional data. Also, our model incorporates two levels of structural information into variable selection using Ising priors: the spatial dependence between voxels and the functional connectivity between anatomical brain regions. Applied to the resting state functional magnetic resonance imaging (R-fMRI) data in the ABIDE study, our methods identify voxel-level imaging biomarkers highly predictive of the ASD, which are biologically meaningful and interpretable. Extensive simulations also show that our methods achieve better performance in variable selection compared to existing methods.

BACKGROUND: Sparse principal component analysis (PCA) is a popular tool for dimensionality reduction, pattern recognition, and visualization of high dimensional data. It has been recognized that complex biological mechanisms occur through concerted relationships of multiple genes working in networks that are often represented by graphs. Recent work has shown that incorporating such biological information improves feature selection and prediction performance in regression analysis, but there has been limited work on extending this approach to PCA. In this article, we propose two new sparse PCA methods called Fused and Grouped sparse PCA that enable incorporation of prior biological information in variable selection. RESULTS: Our simulation studies suggest that, compared to existing sparse PCA methods, the proposed methods achieve higher sensitivity and specificity when the graph structure is correctly specified, and are fairly robust to misspecified graph structures. Application to a glioblastoma gene expression dataset identified pathways that are suggested in the literature to be related with glioblastoma. CONCLUSIONS: The proposed sparse PCA methods Fused and Grouped sparse PCA can effectively incorporate prior biological information in variable selection, leading to improved feature selection and more interpretable principal component loadings and potentially providing insights on molecular underpinnings of complex diseases.

Classification with high dimensional variables is a popular goal in many modern statistical studies. Fisher's linear discriminant analysis (LDA) is a common and effective tool for classifying entities into existing groups. It is well known that classification using Fisher's discriminant for high dimensional data is as bad as random guessing due to the many noise features that increases misclassification rate. Recently, it is being acknowledged that complex biological mechanisms occur through multiple features working together, though individually these features may contribute to noise accumulation in the data. In view of these, it is important to perform classification with discriminant vectors that use a subset of important variables, while also utilizing prior biological relationships among features. We tackle this problem in this article and propose methods that incorporate variable selection into the classification problem, for the identification of important biomarkers. Furthermore, we incorporate into the LDA problem prior information on the relationships among variables using undirected graphs in order to identify functionally meaningful biomarkers. We compare our methods to existing sparse LDA approaches via simulation studies and real data analysis.