As machine learning systems become increasingly integrated into human-centered domains such as healthcare, ensuring fairness while maintaining high predictive performance is critical. Existing bias mitigation techniques often impose a trade-off between fairness and accuracy, inadvertently degrading performance for certain demographic groups. In high-stakes domains like clinical diagnosis, such trade-offs are ethically and practically unacceptable. In this study, we propose a fairness-without-harm approach by learning distinct representations for different demographic groups and selectively applying demographic experts consisting of group-specific representations and personalized classifiers through a no-harm constrained selection. We evaluate our approach on three real-world medical datasets—covering eye disease, skin cancer, and X-ray diagnosis—as well as two face datasets. Extensive empirical results demonstrate the effectiveness of our approach in achieving fairness without harm.

Post-surgery care involves ongoing collaboration between provider teams and patients, which starts from post-surgery hospitalization through home recovery after discharge. While prior HCI research has primarily examined patients' challenges at home, less is known about how provider teams coordinate discharge preparation and care handoffs, and how breakdowns in communication and care pathways may affect patient recovery. To investigate this gap, we conducted semi-structured interviews with 13 healthcare providers and 4 patients in the context of gastrointestinal (GI) surgery. We found coordination boundaries between in- and out-patient teams, coupled with complex organizational structures within teams, impeded the "invisible work" of preparing patients' home care plans and triaging patient information. For patients, these breakdowns resulted in inadequate preparation for home transition and fragmented self-collected data, both of which undermine timely clinical decision-making. Based on these findings, we outline design opportunities to formalize task ownership and handoffs, contextualize co-temporal signals, and align care plans with home resources.

Individual Treatment Effect (ITE) estimation is to evaluate the causal effects of treatment strategies on some important outcomes, which is a crucial problem in healthcare. Most existing ITE estimation methods are designed for centralized settings. However, in real-world clinical scenarios, the raw data are usually not shareable among hospitals due to the potential privacy and security risks, which makes the methods not applicable. In this work, we study the ITE estimation task in a federated setting, which allows us to harness the decentralized data from multiple hospitals. Due to the unavoidable confounding bias in the collected data, a model directly learned from it would be inaccurate. One well-known solution is Inverse Probability Treatment Weighting (IPTW), which uses the conditional probability of treatment given the covariates to re-weight each training example. Applying IPTW in a federated setting, however, is non-trivial. We found that even with a well-estimated conditional probability, the local model training step using each hospital's data alone would still suffer from confounding bias. To address this, we propose FED-IPTW, a novel algorithm to extend IPTW into a federated setting that enforces both global (over all the data) and local (within each hospital) decorrelation between covariates and treatments. We validated our approach on the task of comparing the treatment effects of mechanical ventilation on improving survival probability for patients with breadth difficulties in the Intensive Care Unit (ICU). We conducted experiments on both synthetic and real-world eICU datasets, and the results show that FED-IPTW outperforms state-of-the-art methods on all the metrics on factual prediction and ITE estimation tasks, paving the way for personalized treatment strategy design in mechanical ventilation usage.

Domain adaptation (DA) tackles the issue of distribution shift by learning a model from a source domain that generalizes to a target domain. However, most existing DA methods are designed for scenarios where the source and target domain data lie within the same feature space, which limits their applicability in real-world situations. Recently, heterogeneous DA (HeDA) methods have been introduced to address the challenges posed by heterogeneous feature space between source and target domains. Despite their successes, current HeDA techniques fall short when there is a mismatch in both feature and label spaces. To address this, this paper explores a new DA scenario called open-set HeDA (OSHeDA). In OSHeDA, the model must not only handle heterogeneity in feature space but also identify samples belonging to novel classes. To tackle this challenge, we first develop a novel theoretical framework that constructs learning bounds for prediction error on target domain. Guided by this framework, we propose a new DA method called Representation Learning for OSHeDA (RL-OSHeDA). This method is designed to simultaneously transfer knowledge between heterogeneous data sources and identify novel classes. Experiments across text, image, and clinical data demonstrate the effectiveness of our algorithm. Model implementation is available at https://github.com/pth1993/OSHeDA.

Despite recent advances in cancer treatments that prolong patients' lives, treatment-induced cardiotoxicity (i.e., the various heart damages caused by cancer treatments) emerges as one major side effect. The clinical decision-making process of cardiotoxicity is challenging, as early symptoms may happen in non-clinical settings and are too subtle to be noticed until life-threatening events occur at a later stage; clinicians already have a high workload focusing on the cancer treatment, no additional effort to spare on the cardiotoxicity side effect. Our project starts with a participatory design study with 11 clinicians to understand their decision-making practices and their feedback on an initial design of an AI-based decision-support system. Based on their feedback, we then propose a multimodal AI system, CardioAI, that can integrate wearables data and voice assistant data to model a patient's cardiotoxicity risk to support clinicians' decision-making. We conclude our paper with a small-scale heuristic evaluation with four experts and the discussion of future design considerations.

Parameter-Efficient fine-tuning (PEFT) has attracted significant attention due to the growth of pre-trained model sizes and the need to fine-tune (FT) them for superior downstream performance. Despite a surge in new PEFT methods, a systematic study to understand their performance and suitable application scenarios is lacking, leaving questions like "when to apply PEFT" and "which method to use" largely unanswered, especially in visual recognition. In this paper, we conduct a unifying empirical study of representative PEFT methods with Vision Transformers. We systematically tune their hyperparameters to fairly compare their accuracy on downstream tasks. Our study offers a practical user guide and unveils several new insights. First, if tuned carefully, different PEFT methods achieve similar accuracy in the low-shot benchmark VTAB-1K. This includes simple approaches like FT the bias terms that were reported inferior. Second, despite similar accuracy, we find that PEFT methods make different mistakes and high-confidence predictions, likely due to their different inductive biases. Such an inconsistency (or complementarity) opens up the opportunity for ensemble methods, and we make preliminary attempts at this. Third, going beyond the commonly used low-shot tasks, we find that PEFT is also useful in many-shot regimes, achieving comparable or better accuracy than full FT while using significantly fewer parameters. Lastly, we investigate PEFT’s ability to preserve a pre-trained model’s robustness to distribution shifts (e.g., CLIP). Perhaps not surprisingly, PEFT approaches outperform full FT alone. However, with weight-space ensembles, full FT can better balance target distribution and distribution shift performance, suggesting a future research direction for robust PEFT1.

High-throughput screening techniques, such as microscopy imaging of cellular responses to genetic and chemical perturbations, play a crucial role in drug discovery and biomedical research. However, robust perturbation screening for de novo cell lines remains challenging due to the significant morphological and biological heterogeneity across cell lines. To address this, we propose a novel framework that integrates external biological knowledge into existing pretraining strategies to enhance microscopy image profiling models. Our approach explicitly disentangles perturbation-specific and cell line-specific representations using external biological information. Specifically, we construct a knowledge graph leveraging protein interaction data from STRING and Hetionet databases to guide models toward perturbation-specific features during pretraining. Additionally, we incorporate transcriptomic features from single-cell foundation models to capture cell line-specific representations. By learning these disentangled features, our method improves the generalization of imaging models to de novo cell lines. We evaluate our framework on the RxRx database through one-shot fine-tuning on an RxRx1 cell line and few-shot fine-tuning on cell lines from the RxRx19a dataset. Experimental results demonstrate that our method enhances microscopy image profiling for de novo cell lines, highlighting its effectiveness in real-world phenotype-based drug discovery applications.

drug discovery. Most existing drug repurposing studies using real-world patient data often treat the entire population as homogeneous, ignoring the heterogeneity of treatment responses across patient subgroups. This approach may overlook promising drugs that benefit specific subgroups but lack notable treatment effects across the entire population, potentially limiting the number of repurposable candidates identified. To address this, we introduce STEDR, a novel drug repurposing framework that integrates subgroup analysis with treatment effect estimation. Our approach first identifies repurposing candidates by emulating multiple clinical trials on real-world patient data and then characterizes patient subgroups by learning subgroup-specific treatment effects. We deploy STEDR to Alzheimer's Disease (AD), a condition with few approved drugs and known heterogeneity in treatment responses. We emulate trials for over one thousand medications on a large-scale real-world database covering over 8 million patients, identifying 14 drug candidates with beneficial effects to AD in characterized subgroups. Experiments demonstrate STEDR's superior capability in identifying repurposing candidates compared to existing approaches. Additionally, our method can characterize clinically relevant patient subgroups associated with important AD-related risk factors, paving the way for precision drug repurposing.

Living environments play a vital role in the prevalence and progression of diseases, and understanding their impact on patient's health status becomes increasingly crucial for developing AI models. However, due to the lack of long-term and fine-grained spatial and temporal data in public and population health studies, most existing studies fail to incorporate environmental data, limiting the models' performance and real-world application. To address this shortage, we developed SatHealth, a novel dataset combining multimodal spatiotemporal data, including environmental data, satellite images, all-disease prevalences estimated from medical claims, and social determinants of health (SDoH) indicators. We conducted experiments under two use cases with SatHealth: regional public health modeling and personal disease risk prediction. Experimental results show that living environmental information can significantly improve AI models' performance and temporal-spatial generalizability on various tasks. Finally, we deploy a web-based application to provide an exploration tool for SatHealth and one-click access to both our data and regional environmental embedding to facilitate plug-and-play utilization. SatHealth is now published with data in Ohio, and we will keep updating SatHealth to cover the other parts of the US. With the web application and published code pipeline, our work provides valuable angles and resources to include environmental data in healthcare research and establishes a foundational framework for future research in environmental health informatics.

., the six-organ dysfunction assessment of SOFA in Figure 1) play a vital role in sepsis identification within clinicians' workflow, providing evidence-based risk assessments essential for sepsis diagnosis. However, artificial intelligence (AI) sepsis prediction models typically generate a single sepsis risk score without incorporating clinical calculators for assessing organ dysfunctions, making the models less convincing and transparent to clinicians. To bridge the gap, we propose to mimic clinicians' workflow with a novel framework SepsisCalc to integrate clinical calculators into the predictive model, yielding a clinically transparent and precise model for utilization in clinical settings. Practically, clinical calculators usually combine information from multiple component variables in Electronic Health Records (EHR), and might not be applicable when the variables are (partially) missing. We mitigate this issue by representing EHRs as temporal graphs and integrating a learning module to dynamically add the accurately estimated calculator to the graphs. Experimental results on real-world datasets show that the proposed model outperforms state-of-the-art methods on sepsis prediction tasks. Moreover, we developed a system to identify organ dysfunctions and potential sepsis risks, providing a human-AI interaction tool for deployment, which can help clinicians understand the prediction outputs and prepare timely interventions for the corresponding dysfunctions, paving the way for actionable clinical decision-making support for early intervention.

As machine learning (ML) algorithms are increasingly used in medical image analysis, concerns have emerged about their potential biases against certain social groups. Although many approaches have been proposed to ensure the fairness of ML models, most existing works focus only on medical image diagnosis tasks, such as image classification and segmentation, and overlooked prognosis scenarios, which involve predicting the likely outcome or progression of a medical condition over time. To address this gap, we introduce FairTTE, the first comprehensive framework for assessing fairness in time-to-event (TTE) prediction in medical imaging. FairTTE encompasses a diverse range of imaging modalities and TTE outcomes, integrating cutting-edge TTE prediction and fairness algorithms to enable systematic and fine-grained analysis of fairness in medical image prognosis. Leveraging causal analysis techniques, FairTTE uncovers and quantifies distinct sources of bias embedded within medical imaging datasets. Our large-scale evaluation reveals that bias is pervasive across different imaging modalities and that current fairness methods offer limited mitigation. We further demonstrate a strong association between underlying bias sources and model disparities, emphasizing the need for holistic approaches that target all forms of bias. Notably, we find that fairness becomes increasingly difficult to maintain under distribution shifts, underscoring the limitations of existing solutions and the pressing need for more robust, equitable prognostic models.

Semi-supervised learning (SSL) enhances model performance by leveraging abundant unlabeled data alongside limited labeled data. As vision foundation models (VFMs) become central to modern vision applications, this paper revisits SSL in the context of these powerful pre-trained models. We conduct a systematic study on tasks where frozen VFMs underperform and reveal several key insights when fine-tuning them. First, parameter-efficient fine-tuning (PEFT) using only labeled data often surpasses traditional SSL methods---even without access to unlabeled data. Second, pseudo-labels generated by PEFT models offer valuable supervisory signals for unlabeled data, and different PEFT techniques yield complementary pseudo-labels. These findings motivate a simple yet effective SSL baseline for the VFM era: \emph{ensemble pseudo-labeling across diverse PEFT methods and VFM backbones}. Extensive experiments validate the effectiveness of this approach, offering actionable insights into SSL with VFMs and paving the way for more scalable and robust semi-supervised learning in the foundation model era.

Sleep staging has become a critical task in diagnosing and treating sleep disorders to prevent sleep-related diseases. With growing large-scale sleep databases, significant progress has been made toward automatic sleep staging. However, previous studies face critical problems in sleep studies; the heterogeneity of subjects' physiological signals, the inability to extract meaningful information from unlabeled data to improve predictive performances, the difficulty in modeling correlations between sleep stages, and the lack of an effective mechanism to quantify predictive uncertainty. In this study, we propose a neural network-based sleep staging model, DREAM, to learn domain generalized representations from physiological signals and model sleep dynamics. DREAM learns sleep-related and subject-invariant representations from diverse subjects' sleep signals and models sleep dynamics by capturing interactions between sequential signal segments and between sleep stages. We conducted a comprehensive empirical study to demonstrate the superiority of DREAM, including sleep stage prediction experiments, a case study, the usage of unlabeled data, and uncertainty. Notably, the case study validates DREAM's ability to learn the generalized decision function for new subjects, especially in case there are differences between testing and training subjects. Uncertainty quantification shows that DREAM provides prediction uncertainty, making the model reliable and helping sleep experts in real-world applications.

Given radiology images, automatic radiology report generation aims to produce informative text that reports diseases. It can benefit current clinical practice in diagnostic radiology. Existing methods typically rely on large-scale medical datasets annotated by clinicians to train desirable models. However, for novel diseases, sufficient training data are typically not available. We propose a prompt-based deep learning framework, i.e., PromptLLM, to align, autoencode, and prompt the (large) language model to generate reports for novel diseases accurately and efficiently. Our method includes three major steps: 1) aligning visual images and textual reports to learn general knowledge across modalities from diseases where labeled data are sufficient, 2) autoencoding the LLM using unlabeled data of novel diseases to learn the specific knowledge and writing styles of the novel disease, and 3) prompting the LLM with learned knowledge and writing styles to report the novel diseases contained in the radiology images. Through the above three steps, with limited labels on novel diseases, we show that PromptLLM can rapidly learn the corresponding knowledge for accurate novel disease reporting. The experiments on COVID-19 and diverse thorax diseases show that our approach, utilizing 1% of the training data, achieves desirable performance compared to previous methods. It shows that our approach allows us to relax the reliance on labeled data that is common to existing methods. It could have a real-world impact on data analysis during the early stages of novel diseases.

Precise estimation of treatment effects is crucial for accurately evaluating the intervention. While deep learning models have exhibited promising performance in learning counterfactual representations for treatment effect estimation (TEE), a major limitation in most of these models is that they often overlook the diversity of treatment effects across potential subgroups that have varying treatment effects and characteristics, treating the entire population as a homogeneous group. This limitation restricts the ability to precisely estimate treatment effects and provide targeted treatment recommendations. In this paper, we propose a novel treatment effect estimation model, named SubgroupTE, which incorporates subgroup identification in TEE. SubgroupTE identifies heterogeneous subgroups with different responses and more precisely estimates treatment effects by considering subgroup-specific treatment effects in the estimation process. In addition, we introduce an expectation-maximization (EM)-based training process that iteratively optimizes estimation and subgrouping networks to improve both estimation and subgroup identification. Comprehensive experiments on the synthetic and semi-synthetic datasets demonstrate the outstanding performance of SubgroupTE compared to the existing works for treatment effect estimation and subgrouping models. Additionally, a real-world study demonstrates the capabilities of SubgroupTE in enhancing targeted treatment recommendations for patients with opioid use disorder (OUD) by incorporating subgroup identification with treatment effect estimation.

Scoring sleep stages is essential for evaluating the status of sleep continuity and comprehending its structure. Despite previous attempts, automating sleep scoring remains challenging. First, most existing works did not fuse local and global temporal information. Second, the correlation for special waves in different signals is rarely used in sleep staging modeling. Third, the logic of scoring rules based on adjacent epochs is not considered in developing sleep staging models. This paper introduces a multimodal hybrid-embedding fusion network (MHFNet), which aims to tackle these challenges in automating sleep stage scoring. MHFNet comprises multi-stream Xception blocks to extract wave characteristics, a hybrid time-embedding module to combine local and global temporal information, a dual-path gate transformer to fuse and enhance attention features, and a refined output header to reconstruct sleep scoring. We perform experiments using three publicly available datasets (SleepEDF-ST, SleepEDF-SC, and SHHS). Experimental results indicate the superiority of MHFNet over baseline approaches in cross-validation. Moreover, at the individual level, MHFNet yielded an average $R^{2}$ score improvement of 9$\%$ in the testing dataset compared to state-of-the-art models, paving the way for its applications in real-world sleep medicine.

Treatment effect estimation (TEE) is the task of determining the impact of various treatments on patient outcomes. Current TEE methods fall short due to reliance on limited labeled data and challenges posed by sparse and high-dimensional observational patient data. To address the challenges, we introduce a novel pre-training and fine-tuning framework, KG-TREAT, which synergizes large-scale observational patient data with biomedical knowledge graphs (KGs) to enhance TEE. Unlike previous approaches, KG-TREAT constructs dual-focus KGs and integrates a deep bi-level attention synergy method for in-depth information fusion, enabling distinct encoding of treatment-covariate and outcome-covariate relationships. KG-TREAT also incorporates two pre-training tasks to ensure a thorough grounding and contextualization of patient data and KGs. Evaluation on four downstream TEE tasks shows KG-TREAT’s superiority over existing methods, with an average improvement of 7% in Area under the ROC Curve (AUC) and 9% in Influence Function-based Precision of Estimating Heterogeneous Effects (IF-PEHE). The effectiveness of our estimated treatment effects is further affirmed by alignment with established randomized clinical trial findings.

Today's AI systems for medical decision support often succeed on benchmark datasets in research papers but fail in real-world deployment. This work focuses on the decision making of sepsis, an acute life-threatening systematic infection that requires an early diagnosis with high uncertainty from the clinician. Our aim is to explore the design requirements for AI systems that can support clinical experts in making better decisions for the early diagnosis of sepsis. The study begins with a formative study investigating why clinical experts abandon an existing AI-powered Sepsis predictive module in their electrical health record (EHR) system. We argue that a human-centered AI system needs to support human experts in the intermediate stages of a medical decision-making process (e.g., generating hypotheses or gathering data), instead of focusing only on the final decision. Therefore, we build SepsisLab based on a state-of-the-art AI algorithm and extend it to predict the future projection of sepsis development, visualize the prediction uncertainty, and propose actionable suggestions (i.e., which additional laboratory tests can be collected) to reduce such uncertainty. Through heuristic evaluation with six clinicians using our prototype system, we demonstrate that SepsisLab enables a promising human-AI collaboration paradigm for the future of AI-assisted sepsis diagnosis and other high-stakes medical decision making.

Sepsis is the leading cause of in-hospital mortality in the USA. Early sepsis onset prediction and diagnosis could significantly improve the survival of sepsis patients. Existing predictive models are usually trained on high-quality data with few missing information, while missing values widely exist in real-world clinical scenarios (especially in the first hours of admissions to the hospital), which causes a significant decrease in accuracy and an increase in uncertainty for the predictive models. The common method to handle missing values is imputation, which replaces the unavailable variables with estimates from the observed data. The uncertainty of imputation results can be propagated to the sepsis prediction outputs, which have not been studied in existing works on either sepsis prediction or uncertainty quantification. In this study, we first define such propagated uncertainty as the variance of prediction output and then introduce uncertainty propagation methods to quantify the propagated uncertainty. Moreover, for the potential high-risk patients with low confidence due to limited observations, we propose a robust active sensing algorithm to increase confidence by actively recommending clinicians to observe the most informative variables. We validate the proposed models in both publicly available data (i.e., MIMIC-III and AmsterdamUMCdb) and proprietary data in The Ohio State University Wexner Medical Center (OSUWMC). The experimental results show that the propagated uncertainty is dominant at the beginning of admissions to hospitals and the proposed algorithm outperforms state-of-the-art active sensing methods. Finally, we implement a SepsisLab system for early sepsis prediction and active sensing based on our pre-trained models. Clinicians and potential sepsis patients can benefit from the system in early prediction and diagnosis of sepsis.

Fine-tuning is arguably the most straightforward way to tailor a pre-trained model (e.g., a foundation model) to downstream applications, but it also comes with the risk of losing valuable knowledge the model had learned in pre-training. For example, fine-tuning a pre-trained classifier capable of recognizing a large number of classes to master a subset of classes at hand is shown to drastically degrade the model's accuracy in the other classes it had previously learned. As such, it is hard to further use the fine-tuned model when it encounters classes beyond the fine-tuning data. In this paper, we systematically dissect the issue, aiming to answer the fundamental question, "What has been damaged in the fine-tuned model?" To our surprise, we find that the fine-tuned model neither forgets the relationship among the other classes nor degrades the features to recognize these classes. Instead, the fine-tuned model often produces more discriminative features for these other classes, even if they were missing during fine-tuning! What really hurts the accuracy is the discrepant logit scales between the fine-tuning classes and the other classes, implying that a simple post-processing calibration would bring back the pre-trained model's capability and at the same time unveil the feature improvement over all classes. We conduct an extensive empirical study to demonstrate the robustness of our findings and provide preliminary explanations underlying them, suggesting new directions for future theoretical analysis.

Although recent advances in machine learning have shown its success to learn from independent and identically distributed (IID) data, it is vulnerable to out-of-distribution (OOD) data in an open world. Domain generalization (DG) deals with such an issue and it aims to learn a model from multiple source domains that can be generalized to unseen target domains. Existing studies on DG have largely focused on stationary settings with homogeneous source domains. However, in many applications, domains may evolve along a specific direction (e.g., time, space). Without accounting for such non-stationary patterns, models trained with existing methods may fail to generalize on OOD data. In this paper, we study domain generalization in non-stationary environment. We first examine the impact of environmental non-stationarity on model performance and establish the theoretical upper bounds for the model error at target domains. Then, we propose a novel algorithm based on adaptive invariant representation learning, which leverages the non-stationary pattern to train a model that attains good performance on target domains. Experiments on both synthetic and real data validate the proposed algorithm.

Deep learning models have demonstrated promising results in estimating treatment effects (TEE). However, most of them overlook the variations in treatment outcomes among subgroups with distinct characteristics. This limitation hinders their ability to provide accurate estimations and treatment recommendations for specific subgroups. In this study, we introduce a novel neural network-based framework, named SubgroupTE, which incorporates subgroup identification and treatment effect estimation. SubgroupTE identifies diverse subgroups and simultaneously estimates treatment effects for each subgroup, improving the treatment effect estimation by considering the heterogeneity of treatment responses. Comparative experiments on synthetic data show that SubgroupTE outperforms existing models in treatment effect estimation. Furthermore, experiments on a real-world dataset related to opioid use disorder (OUD) demonstrate the potential of our approach to enhance personalized treatment recommendations for OUD patients.

Rare diseases, which are severely underrepresented in basic and clinical research, can particularly benefit from machine learning techniques. However, current learning-based approaches usually focus on either mono-modal image data or matched multi-modal data, whereas the diagnosis of rare diseases necessitates the aggregation of unstructured and unmatched multi-modal image data due to their rare and diverse nature. In this study, we therefore propose diagnosis-guided multi-to-mono modal generation networks (TMM-Nets) along with training and testing procedures. TMM-Nets can transfer data from multiple sources to a single modality for diagnostic data structurization. To demonstrate their potential in the context of rare diseases, TMM-Nets were deployed to diagnose the lupus retinopathy (LR-SLE), leveraging unmatched regular and ultra-wide-field fundus images for transfer learning. The TMM-Nets encoded the transfer learning from diabetic retinopathy to LR-SLE based on the similarity of the fundus lesions. In addition, a lesion-aware multi-scale attention mechanism was developed for clinical alerts, enabling TMM-Nets not only to inform patient care, but also to provide insights consistent with those of clinicians. An adversarial strategy was also developed to refine multi- to mono-modal image generation based on diagnostic results and the data distribution to enhance the data augmentation performance. Compared to the baseline model, the TMM-Nets showed 35.19% and 33.56% F1 score improvements on the test and external validation sets, respectively. In addition, the TMM-Nets can be used to develop diagnostic models for other rare diseases.

Sleep staging is a key challenge in diagnosing and treating sleep-related diseases due to its labor-intensive, time-consuming, costly, and error-prone. With the availability of large-scale sleep signal data, many deep learning methods are proposed for automatic sleep staging. However, these existing methods face several challenges including the heterogeneity of patients’ underlying health conditions and the difficulty modeling complex interactions between sleep stages. In this paper, we propose a neural network architecture named DREAM to tackle these issues for automatic sleep staging. DREAM consists of (i) a feature representation network that generates robust representations for sleep signals via the variational auto-encoder framework and contrastive learning and (ii) a sleep stage classification network that explicitly models the interactions between sleep stages in the sequential context at both feature representation and label classification levels via Transformer and conditional random field architectures. Our experimental results indicate that DREAM significantly outperforms existing methods for automatic sleep staging on three sleep signal datasets.

Despite intense efforts in basic and clinical research, an individualized ventilation strategy for critically ill patients remains a major challenge. Recently, dynamic treatment regime (DTR) with reinforcement learning (RL) on electronic health records (EHR) has attracted interest from both the healthcare industry and machine learning research community. However, most learned DTR policies might be biased due to the existence of confounders. Although some treatment actions non-survivors received may be helpful, if confounders cause the mortality, the training of RL models guided by long-term outcomes (e.g., 90-day mortality) would punish those treatment actions causing the learned DTR policies to be suboptimal. In this study, we develop a new deconfounding actor-critic network (DAC) to learn optimal DTR policies for patients. To alleviate confounding issues, we incorporate a patient resampling module and a confounding balance module into our actor-critic framework. To avoid punishing the effective treatment actions non-survivors received, we design a short-term reward to capture patients' immediate health state changes. Combining short-term with long-term rewards could further improve the model performance. Moreover, we introduce a policy adaptation method to successfully transfer the learned model to new-source small-scale datasets. The experimental results on one semi-synthetic and two different real-world datasets show the proposed model outperforms the state-of-the-art models. The proposed model provides individualized treatment decisions for mechanical ventilation that could improve patient outcomes.

Age-related macular degeneration (AMD) is the leading cause of irreversible blindness in developed countries. Identifying patients at high risk of progression to late AMD, the sight-threatening stage, is critical for clinical actions, including medical interventions and timely monitoring. Recently, deep-learning-based models have been developed and achieved superior performance for late AMD prediction. However, most existing methods are limited to the color fundus photography (CFP) from the last ophthalmic visit and do not include the longitudinal CFP history and AMD progression during the previous years' visits. Patients in different AMD subphenotypes might have various speeds of progression in different stages of AMD disease. Capturing the progression information during the previous years' visits might be useful for the prediction of AMD progression. In this work, we propose a C ontrastive-A ttention-based T ime-aware L ong S hort-T erm M emory network (CAT-LSTM ) to predict AMD progression. First, we adopt a convolutional neural network (CNN) model with a contrastive attention module (CA) to extract abnormal features from CFPs. Then we utilize a time-aware LSTM (T-LSTM) to model the patients' history and consider the AMD progression information. The combination of disease progression, genotype information, demographics, and CFP features are sent to T-LSTM. Moreover, we leverage an auto-encoder to represent temporal CFP sequences as fixed-size vectors and adopt k-means to cluster them into subphenotypes. We evaluate the proposed model based on real-world datasets, and the results show that the proposed model could achieve 0.925 on area under the receiver operating characteristic (AUROC) for 5-year late-AMD prediction and outperforms the state-of-the-art methods by more than 3%, which demonstrates the effectiveness of the proposed CAT-LSTM. After analyzing patient representation learned by an auto-encoder, we identify 3 novel subphenotypes of AMD patients with different characteristics and progression rates to late AMD, paving the way for improved personalization of AMD management. The code of CAT-LSTM can be found at https://github.com/yinchangchang/CAT-LSTM .

molecular design is a key challenge in drug discovery due to the complexity of chemical space. With the availability of molecular datasets and advances in machine learning, many deep generative models are proposed for generating novel molecules with desired properties. However, most of the existing models focus only on molecular distribution learning and target-based molecular design, thereby hindering their potentials in real-world applications. In drug discovery, phenotypic molecular design has advantages over target-based molecular design, especially in first-in-class drug discovery. In this work, we propose the first deep graph generative model (FAME) targeting phenotypic molecular design, in particular gene expression-based molecular design. FAME leverages a conditional variational autoencoder framework to learn the conditional distribution generating molecules from gene expression profiles. However, this distribution is difficult to learn due to the complexity of the molecular space and the noisy phenomenon in gene expression data. To tackle these issues, a gene expression denoising (GED) model that employs contrastive objective function is first proposed to reduce noise from gene expression data. FAME is then designed to treat molecules as the sequences of fragments and learn to generate these fragments in autoregressive manner. By leveraging this fragment-based generation strategy and the denoised gene expression profiles, FAME can generate novel molecules with a high validity rate and desired biological activity. The experimental results show that FAME outperforms existing methods including both SMILES-based and graph-based deep generative models for phenotypic molecular design. Furthermore, the effective mechanism for reducing noise in gene expression data proposed in our study can be applied to omics data modeling in general for facilitating phenotypic drug discovery.

Complication risk profiling is a key challenge in the healthcare domain due to the complex interaction between heterogeneous entities (e.g., visit, disease, medication) in clinical data. With the availability of real-world clinical data such as electronic health records and insurance claims, many deep learning methods are proposed for complication risk profiling. However, these existing methods face two open challenges. First, data heterogeneity relates to those methods leveraging clinical data from a single view only while the data can be considered from multiple views (e.g., sequence of clinical visits, set of clinical features). Second, generalized prediction relates to most of those methods focusing on single-task learning, whereas each complication onset is predicted independently, leading to suboptimal models. We propose a multi-view multi-task network (MuViTaNet) for predicting the onset of multiple complications to tackle these issues. In particular, MuViTaNet complements patient representation by using a multi-view encoder to effectively extract information by considering clinical data as both sequences of clinical visits and sets of clinical features. In addition, it leverages additional information from both related labeled and unlabeled datasets to generate more generalized representations by using a new multi-task learning scheme for making more accurate predictions. The experimental results show that MuViTaNet outperforms existing methods for profiling the development of cardiac complications in breast cancer survivors. Furthermore, thanks to its multi-view multi-task architecture, MuViTaNet also provides an effective mechanism for interpreting its predictions in multiple perspectives, thereby helping clinicians discover the underlying mechanism triggering the onset and for making better clinical treatments in real-world scenarios.

Disease progression modeling (DPM) involves using mathematical frameworks to quantitatively measure the severity of how certain disease progresses. DPM is useful in many ways such as predicting health state, categorizing disease stages, and assessing patients’ disease trajectory, etc. Recently, with the wider availability of electronic health records (EHR) and the broad application of data-driven machine learning methods, DPM has attracted much attention yet remains two major challenges: (i) Due to the existence of irregularity, heterogeneity, and long-term dependency in EHRs, most existing DPM methods might not be able to provide comprehensive patient representations. (ii) Lots of records in EHRs might be irrelevant to the target disease. Most existing models learn to automatically focus on the relevant information instead of explicitly capture the target-relevant events, which might make the learned model suboptimal. To address these two issues, we propose Temporal Clustering with External Memory Network (TC-EMNet) for DPM that groups patients with similar trajectories to form disease clusters/stages. TC-EMNet uses a variational autoencoder (VAE) to capture internal complexity from the input data and utilizes an external memory work to capture long-term distance information, both of which are helpful for producing comprehensive patient health states. Last but not least, the k-means algorithm is adopted to cluster the extracted comprehensive patient representation to capture disease progression. Experiments on two real-world datasets show that our model demonstrates competitive clustering performance against state-of-the-art methods and is able to identify clinically meaningful clusters. The visualization of the patient representations shows that the proposed model can generate better patient health states than the baselines.

Ultrasound is a widely used technology for diagnosing developmental dysplasia of the hip (DDH) because it does not use radiation. Due to its low cost and convenience, 2-D ultrasound is still the most common examination in DDH diagnosis. In clinical usage, the complexity of both ultrasound image standardization and measurement leads to a high error rate for sonographers. The automatic segmentation results of key structures in the hip joint can be used to develop a standard plane detection method that helps sonographers decrease the error rate. However, current automatic segmentation methods still face challenges in robustness and accuracy. Thus, we propose a neonatal hip bone segmentation network (NHBS-Net) for the first time for the segmentation of seven key structures. We design three improvements, an enhanced dual attention module, a two-class feature fusion module, and a coordinate convolution output head, to help segment different structures. Compared with current state-of-the-art networks, NHBS-Net gains outstanding performance accuracy and generalizability, as shown in the experiments. Additionally, image standardization is a common need in ultrasonography. The ability of segmentation-based standard plane detection is tested on a 50-image standard dataset. The experiments show that our method can help healthcare workers decrease their error rate from 6%-10% to 2%. In addition, the segmentation performance in another ultrasound dataset (fetal heart) demonstrates the ability of our network.

Estimating the individual treatment effect (ITE) from observational data is meaningful and practical in healthcare. Existing work mainly relies on the strong ignorability assumption that no hidden confounders exist, which may lead to bias in estimating causal effects. Some studies considering the hidden confounders are designed for static environment and not easily adaptable to a dynamic setting. In fact, most observational data (e.g., electronic medical records) is naturally dynamic and consists of sequential information. In this paper, we propose Deep Sequential Weighting (DSW) for estimating ITE with time-varying confounders. Specifically, DSW infers the hidden confounders by incorporating the current treatment assignments and historical information using a deep recurrent weighting neural network. The learned representations of hidden confounders combined with current observed data are leveraged for potential outcome and treatment predictions. We compute the time-varying inverse probabilities of treatment for re-weighting the population. We conduct comprehensive comparison experiments on fully-synthetic, semi-synthetic and real-world datasets to evaluate the performance of our model and baselines. Results demonstrate that our model can generate unbiased and accurate treatment effect by conditioning both time-varying observed and hidden confounders, paving the way for personalized medicine.

Sepsis is a heterogeneous clinical syndrome that is the leading cause of mortality in hospital intensive care units (ICUs). Identification of sepsis subphenotypes may allow for more precise treatments and lead to more targeted clinical interventions. Recently, sepsis subtyping on electronic health records (EHRs) has attracted interest from healthcare researchers. However, most sepsis subtyping studies ignore the temporality of EHR data and suffer from missing values. In this paper, we propose a new sepsis subtyping framework to address the two issues. Our subtyping framework consists of a novel Time-Aware Multi-modal auto-Encoder (TAME) model which introduces time-aware attention mechanism and incorporates multi-modal inputs (e.g., demographics, diagnoses, medications, lab tests and vital signs) to impute missing values, a dynamic time wrapping (DTW) method to measure patients' temporal similarity based on the imputed EHR data, and a weighted k-means algorithm to cluster patients. Comprehensive experiments on real-world datasets show TAME outperforms the baselines on imputation accuracy. After analyzing TAME-imputed EHR data, we identify four novel subphenotypes of sepsis patients, paving the way for improved personalization of sepsis management.

MOTIVATION: Graph embedding learning that aims to automatically learn low-dimensional node representations, has drawn increasing attention in recent years. To date, most recent graph embedding methods are evaluated on social and information networks and are not comprehensively studied on biomedical networks under systematic experiments and analyses. On the other hand, for a variety of biomedical network analysis tasks, traditional techniques such as matrix factorization (which can be seen as a type of graph embedding methods) have shown promising results, and hence there is a need to systematically evaluate the more recent graph embedding methods (e.g. random walk-based and neural network-based) in terms of their usability and potential to further the state-of-the-art. RESULTS: We select 11 representative graph embedding methods and conduct a systematic comparison on 3 important biomedical link prediction tasks: drug-disease association (DDA) prediction, drug-drug interaction (DDI) prediction, protein-protein interaction (PPI) prediction; and 2 node classification tasks: medical term semantic type classification, protein function prediction. Our experimental results demonstrate that the recent graph embedding methods achieve promising results and deserve more attention in the future biomedical graph analysis. Compared with three state-of-the-art methods for DDAs, DDIs and protein function predictions, the recent graph embedding methods achieve competitive performance without using any biological features and the learned embeddings can be treated as complementary representations for the biological features. By summarizing the experimental results, we provide general guidelines for properly selecting graph embedding methods and setting their hyper-parameters for different biomedical tasks. AVAILABILITY AND IMPLEMENTATION: As part of our contributions in the paper, we develop an easy-to-use Python package with detailed instructions, BioNEV, available at: https://github.com/xiangyue9607/BioNEV, including all source code and datasets, to facilitate studying various graph embedding methods on biomedical tasks. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Drug repurposing, referring to the discovery of new indications for an existing drug, has become an important aspect of drug development. This study seeks to extract hidden disease-drug associations from existing literature. Specifically, we propose a systematic framework assuming that a drug has a new use for a certain disease if the clinical disease symptoms and clinical drug effects complement each other. Via literature mining, we generate two vectors to describe clinical disease symptoms and clinical drug effects. Then we use a logistic regression model to predict how effectively the drug counteracts the disease based on the generated disease/drug vectors. Comprehensive experiments show the effectiveness of proposed framework. According to our results, there are many hidden disease-drug associations which can be the candidates for drug repurposing.

Due to their promising performance in clinical risk prediction with Electronic Health Records (EHRs), deep learning methods have attracted significant interest from healthcare researchers. However, there are 4 challenges: (i) Data insufficiency. Many methods require large amounts of training data to achieve satisfactory results. (ii) Interpretability. Results from many methods are hard to explain to clinicians (e.g., why the models make particular predictions and which events cause clinical outcomes). (iii) Domain knowledge integration. No existing method dynamically exploits complicated medical knowledge (e.g., relations such as cause and is-caused-by between clinical events). (iv) Time interval information. Most existing methods only consider the relative order of visits from EHRs, but ignore the irregular time intervals between neighboring visits. In the study, we propose a new model, Domain Knowledge Guided Recurrent Neural Networks (DG-RNN), by directly introducing domain knowledge from the medical knowledge graph into an RNN architecture, as well as taking the irregular time intervals into account. Experimental results on heart failure risk prediction tasks show that our model not only outperforms state-of-the-art deep-learning based risk prediction models, but also associates individual medical events with heart failure onset, thus paving the way for interpretable accurate clinical risk predictions.

In contrast to the one-size-fits-all approach to medicine, precision medicine will allow targeted prescriptions based on the specific profile of the patient thereby avoiding adverse reactions and ineffective but expensive treatments. Longitudinal observational data such as Electronic Health Records (EHRs) have become an emerging data source for personalized medicine. In this paper, we propose a unified computational framework, called PerDREP, to predict the unique response patterns of each individual patient from EHR data. PerDREP models individual responses of each patient to the drug exposure by introducing a linear system to account for patients' heterogeneity, and incorporates a patient similarity graph as a network regularization. We formulate PerDREP as a convex optimization problem and develop an iterative gradient descent method to solve it. In the experiments, we identify the effect of drugs on Glycated hemoglobin test results. The experimental results provide evidence that the proposed method is not only more accurate than state-of-the-art methods, but is also able to automatically cluster patients into multiple coherent groups, thus paving the way for personalized medicine.

This paper presents a cloud-based system for health care applications. Our system has advanced features for preserving privacy which are essential for health care applications that deal with confidential data. We describe some of the bioinformatics applications which our system is designed for. Performance is significantly enhanced by caching, and enhanced clients for performing part of the computations are a key component of our system. Cloud, due to its pay-as-you-go pricing and API based deployment model, has become widely used for delivering and maintaining infrastructure technology for businesses. However, there are significant challenges with using the cloud for applications with strict privacy and compliance requirements; health care applications fall in this domain. This paper describes an architecture and solutions for handling these types of applications.

The identification of drug-target interactions (DTIs) is a key task in drug discovery, where drugs are chemical compounds and targets are proteins. Traditional DTI prediction methods are either time consuming (simulation-based methods) or heavily dependent on domain expertise (similarity-based and feature-based methods). In this work, we propose an end-to-end neural network model that predicts DTIs directly from low level representations. In addition to making predictions, our model provides biological interpretation using two-way attention mechanism. Instead of using simplified settings where a dataset is evaluated as a whole, we designed an evaluation dataset from BindingDB following more realistic settings where predictions of unobserved examples (proteins and drugs) have to be made. We experimentally compared our model with matrix factorization, similarity-based methods, and a previous deep learning approach. Overall, the results show that our model outperforms other approaches without requiring domain knowledge and feature engineering. In a case study, we illustrated the ability of our approach to provide biological insights to interpret the predictions.

BACKGROUND: Adverse drug reactions (ADRs) are unintended and harmful reactions caused by normal uses of drugs. Predicting and preventing ADRs in the early stage of the drug development pipeline can help to enhance drug safety and reduce financial costs. METHODS: In this paper, we developed machine learning models including a deep learning framework which can simultaneously predict ADRs and identify the molecular substructures associated with those ADRs without defining the substructures a-priori. RESULTS: We evaluated the performance of our model with ten different state-of-the-art fingerprint models and found that neural fingerprints from the deep learning model outperformed all other methods in predicting ADRs. Via feature analysis on drug structures, we identified important molecular substructures that are associated with specific ADRs and assessed their associations via statistical analysis. CONCLUSIONS: The deep learning model with feature analysis, substructure identification, and statistical assessment provides a promising solution for identifying risky components within molecular structures and can potentially help to improve drug safety evaluation.

Adverse drug-drug interactions (DDIs) remain a leading cause of morbidity and mortality around the world. Identifying potential DDIs during the drug design process is critical in guiding targeted clinical drug safety testing. Although detection of adverse DDIs is conducted during Phase IV clinical trials, there are still a large number of new DDIs founded by accidents after the drugs were put on market. With the arrival of big data era, more and more pharmaceutical research and development data are becoming available, which provides an invaluable resource for digging insights that can potentially be leveraged in early prediction of DDIs. Many computational approaches have been proposed in recent years for DDI prediction. However, most of them focused on binary prediction (with or without DDI), despite the fact that each DDI is associated with a different type. Predicting the actual DDI type will help us better understand the DDI mechanism and identify proper ways to prevent it. In this paper, we formulate the DDI type prediction problem as a multitask dyadic regression problem, where the prediction of each specific DDI type is treated as a task. Compared with conventional matrix completion approaches which can only impute the missing entries in the DDI matrix, our approach can directly regress those dyadic relationships (DDIs) and thus can be extend to new drugs more easily. We developed an effective proximal gradient method to solve the problem. Evaluation on real world datasets is presented to demonstrate the effectiveness of the proposed approach.

Adverse drug reaction (ADR) is a major burden for patients and healthcare industry. It usually causes preventable hospitalizations and deaths, while associated with a huge amount of cost. Traditional preclinical in vitro safety profiling and clinical safety trials are restricted in terms of small scale, long duration, huge financial costs and limited statistical signifi- cance. The availability of large amounts of drug and ADR data potentially allows ADR predictions during the drugs’ early preclinical stage with data analytics methods to inform more targeted clinical safety tests. Despite their initial success, existing methods have trade-offs among interpretability, predictive power and efficiency. This urges us to explore methods that could have all these strengths and provide practical solutions for real world ADR predictions. We cast the ADR-drug relation structure into a three-layer hierarchical Bayesian model. We interpret each ADR as a symbolic word and apply latent Dirichlet allocation (LDA) to learn topics that may represent certain biochemical mechanism that relates ADRs with drug structures. Based on LDA, we designed an equivalent regularization term to incorporate the hierarchical ADR domain knowledge. Finally, we developed a mixed input model leveraging a fast collapsed Gibbs sampling method that the complexity of each iteration of Gibbs sampling proportional only to the number of positive ADRs. Experiments on real world data show our models achieved higher prediction accuracy and shorter running time than the state-of-the-art alternatives.

Biological interpretation and understanding of machine learning based predictive models are highly desirable in healthcare analytics. Predicting Adverse Drug Reactions (ADRs) is extremely important for safe and precision medicine. There are various machine learning based approaches to predict adverse reactions for drugs. These models, though effective, lack biological interpretation and are treated as black-boxes. We propose DrugPathSeeker, a novel interactive user interface that integrates the machine learning model, database query API, statistical analysis, and visualization for exploring and understanding of the association between drugs and ADRs. The proposed UI can take a query drug, and provide a visual interface designed to support exploration of the predictions from the machine learning model for further understanding and interpretation. DrugPathSeeker uses a machine learning model, Small Molecular Risk Profiler, to make ADR predictions for a given drug. The visualization uses Sankey type flow diagrams for highlighting the relation between the drugs and ADRs. The main goal of Drug-PathSeeker is to mine the gene-pathways from public databases and analyze them in a visual manner to generate a biological hypothesis. DrugPathSeeker's effectiveness is demonstrated with two use cases: mechanisms of action for carbamazepine-induced dystonia, and fluorometholone-induced diabetes mellitus.

The proliferation of Electronic Health Records (EHRs) challenges data miners to discover potential and previously unknown patterns from a large collection of medical data. One of the tasks that we address in this paper is to reveal previously unknown effects of drugs on laboratory test results. We propose a method that leverages drug information to find a meaningful list of drugs that have an effect on the laboratory result. We formulate the problem as a convex non smooth function and develop a proximal gradient method to optimize it. The model has been evaluated on two important use cases: lowering low-density lipoproteins and glycated hemoglobin test results. The experimental results provide evidence that the proposed method is more accurate than the state-of-the-art method, rediscover drugs that are known to lower the levels of laboratory test results, and most importantly, discover additional potential drugs that may also lower these levels.

Computational Drug Discovery, which uses computational techniques to facilitate and improve the drug discovery process, has aroused considerable interests in recent years. Drug Repositioning (DR) and Drug-Drug Interaction (DDI) prediction are two key problems in drug discovery and many computational techniques have been proposed for them in the last decade. Although these two problems have mostly been researched separately in the past, both DR and DDI can be formulated as the problem of detecting positive interactions between data entities (DR is between drug and disease, and DDI is between pairwise drugs). The challenge in both problems is that we can only observe a very small portion of positive interactions. In this paper, we propose a novel framework called Dyadic Positive-Unlabeled learning (DyPU) to solve the problem of detecting positive interactions. DyPU forces positive data pairs to rank higher than the average score of unlabeled data pairs. Moreover, we also derive the dual formulation of the proposed method with the rectifier scoring function and we show that the associated non-trivial proximal operator admits a closed form solution. Extensive experiments are conducted on real drug data sets and the results show that our method achieves superior performance comparing with the state-of-the-art.

We study the problem of Polyadic Prediction, where the input consists of an ordered tuple of objects, and the goal is to predict a measurement associated with them. Many tasks can be naturally framed as Polyadic Prediction problems. In drug discovery, for instance, it is important to estimate the treatment effect of a drug on various tissue-specific diseases, as it is expressed over the available genes. Thus, we essentially predict the expression value measurements for several (drug, gene, tissue) triads. To tackle Polyadic Prediction problems, we propose a general framework, called Polyadic Regression, predicting measurements associated with multiple objects. Our framework is inductive, in the sense of enabling predictions for new objects, unseen during training. Our model is expressive, exploring high-order, polyadic interactions in an efficient manner. An alternating Proximal Gradient Descent procedure is proposed to fit our model. We perform an extensive evaluation using real-world chemogenomics data, where we illustrate the superior performance of Polyadic Regression over the prior art. Our method achieves an increase of 0.06 and 0.1 in Spearman correlation between the predicted and the actual measurement vectors, for predicting missing polyadic data and predicting polyadic data for new drugs, respectively.

In the last decade, advances in high-throughput technologies, growth of clinical data warehouses, and rapid accumulation of biomedical knowledge provided unprecedented opportunities and challenges to researchers in biomedical informatics. One distinct solution, to efficiently conduct big data analytics for biomedical problems, is the application of matrix computation and factorization methods such as non-negative matrix factorization, joint matrix factorization, tensor factorization. Compared to probabilistic and information theoretic approaches, matrix-based methods are fast, easy to understand and implement. In this tutorial, we provide a review of recent advances in algorithms and methods using matrix and their potential applications in biomedical informatics. We survey various related articles from data mining venues as well as from biomedical informatics venues to share with the audience key problems and trends in matrix computation research, with different novel applications such as drug repositioning, personalized medicine, and electronic phenotyping.

The recent years have witnessed a surge of interests in data analytics with patient Electronic Health Records (EHR). Data-driven healthcare, which aims at effective utilization of big medical data, representing the collective learning in treating hundreds of millions of patients, to provide the best and most personalized care, is believed to be one of the most promising directions for transforming healthcare. EHR is one of the major carriers for make this data-driven healthcare revolution successful. There are many challenges on working directly with EHR, such as temporality, sparsity, noisiness, bias, etc. Thus effective feature extraction, or phenotyping from patient EHRs is a key step before any further applications. In this paper, we propose a deep learning approach for phenotyping from patient EHRs. We first represent the EHRs for every patient as a temporal matrix with time on one dimension and event on the other dimension. Then we build a four-layer convolutional neural network model for extracting phenotypes and perform prediction. The first layer is composed of those EHR matrices. The second layer is a one-side convolution layer that can extract phenotypes from the first layer. The third layer is a max pooling layer introducing sparsity on the detected phenotypes, so that only those significant phenotypes will remain. The fourth layer is a fully connected softmax prediction layer. In order to incorporate the temporal smoothness of the patient EHR, we also investigated three different temporal fusion mechanisms in the model: early fusion, late fusion and slow fusion. Finally the proposed model is validated on a real world EHR data warehouse under the specific scenario of predictive modeling of chronic diseases.

Logistic regression is one core predictive modeling technique that has been used extensively in health and biomedical problems. Recently a lot of research has been focusing on enforcing sparsity on the learned model to enhance its effectiveness and interpretability, which results in sparse logistic regression model. However, no matter the original or sparse logistic regression, they require the inputs to be in vector form. This limits the applicability of logistic regression in the problems when the data cannot be naturally represented vectors (e.g., functional magnetic resonance imaging and electroencephalography signals). To handle the cases when the data are in the form of multi-dimensional arrays, we propose MulSLR: Multilinear Sparse Logistic Regression. MulSLR can be viewed as a high order extension of sparse logistic regression. Instead of solving one classification vector as in conventional logistic regression, we solve for K classification vectors in MulSLR (K is the number of modes in the data). We propose a block proximal descent approach to solve the problem and prove its convergence. The convergence rate of the proposed algorithm is also analyzed. Finally we validate the efficiency and effectiveness of MulSLR on predicting the onset risk of patients with Alzheimer's disease and heart failure.

OBJECTIVE: Common chronic diseases such as hypertension are costly and difficult to manage. Our ultimate goal is to use data from electronic health records to predict the risk and timing of deterioration in hypertension control. Towards this goal, this work predicts the transition points at which hypertension is brought into, as well as pushed out of, control. METHOD: In a cohort of 1294 patients with hypertension enrolled in a chronic disease management program at the Vanderbilt University Medical Center, patients are modeled as an array of features derived from the clinical domain over time, which are distilled into a core set using an information gain criteria regarding their predictive performance. A model for transition point prediction was then computed using a random forest classifier. RESULTS: The most predictive features for transitions in hypertension control status included hypertension assessment patterns, comorbid diagnoses, procedures and medication history. The final random forest model achieved a c-statistic of 0.836 (95% CI 0.830 to 0.842) and an accuracy of 0.773 (95% CI 0.766 to 0.780). CONCLUSIONS: This study achieved accurate prediction of transition points of hypertension control status, an important first step in the long-term goal of developing personalized hypertension management plans.

BACKGROUND: In many biomedical applications, there is a need for developing classification models based on noisy annotations. Recently, various methods addressed this scenario by relaying on unreliable annotations obtained from multiple sources. RESULTS: We proposed a probabilistic classification algorithm based on labels obtained by multiple noisy annotators. The new algorithm is capable of eliminating annotations provided by novice labellers and of providing a more accurate estimate of the ground truth by consensus labelling according to higher quality annotations. The approach is evaluated on text classification and prediction of protein disorder. Our study suggests that the higher levels of accuracy, effectiveness and performance can be achieved by the new method as compared to alternatives. CONCLUSIONS: The proposed method is applicable for meta-learning from multiple existing classification models and noisy annotations obtained by humans. It is particularly beneficial when many annotations are obtained by novice labellers. In addition, the proposed method can provide further characterization of each annotator that can help in developing more accurate classifiers by identifying the most competent annotators for each data instance.

Learning from noisy labels obtained from multiple annotators and without access to any true labels is an increasingly important problem in bioinformatics and biomedicine. In our method, this challenge is addressed by iteratively filtering low-quality annotators and estimating the consensus labels based only on the remaining experts that provide higher-quality annotations. Experiments on biomedical text classification and CASP9 protein disorder prediction tasks provide evidence that the proposed algorithm is more accurate than the majority voting and previously developed multi-annotator approaches. The benefit of using the new method is particularly large when low-quality annotators dominate. Moreover, the new algorithm also suggests the most relevant annotators for each instance, thus paving the way for understanding the behaviors of each annotator and building more reliable predictive models for bioinformatics applications.

We propose a novel semi-supervised learning algorithm, called IMCC, designed for co-training classifiers on single-view datasets. Our method runs the co-training algorithm for a predefined number of times, each time using a different random split of features. Thus, a set of diverse co-training classifiers is created. Each of these classifiers then labels each of the examples for which we want to determine the class label. In this way, each example for classification is assigned multiple labels. We then treat this as a problem of learning from inconsistent and unreliable annotators in a multi-annotator problem setting and estimate the single hidden true label for each example. In experimental results obtained on 25 benchmark datasets of various properties IMCC outperformed five considered alternative methods for co-training on single-view datasets, and resulted in a statistical tie with a Naive Bayes classifier trained using a much larger set of labeled examples.

Studies of intrinsically disordered proteins that lack a stable tertiary structure but still have important biological functions critically rely on computational methods that predict this property based on sequence information. Although a number of fairly successful models for prediction of protein disorder were developed over the last decade, the quality of their predictions is limited by available cases of confirmed disorders. To more reliably estimate protein disorder from protein sequences, an iterative algorithm is proposed that integrates predictions of multiple disorder models without relying on any protein sequences with confirmed disorder annotation. The iterative method alternately provides the maximum a posterior (MAP) estimation of disorder prediction and the maximum-likelihood (ML) estimation of quality of multiple disorder predictors. Experiments on data used at the Critical Assessment of Techniques for Protein Structure Prediction (CASP7 and CASP8) have shown the effectiveness of the proposed algorithm.