Supervised training of neural networks requires large, diverse and well annotated data sets. In the medical field, this is often difficult to achieve due to constraints in time, expert knowledge and prevalence of an event. Artificial data augmentation can help to prevent overfitting and improve the detection of rare events as well as overall performance. However, most augmentation techniques use purely spatial transformations, which are not sufficient for video data with temporal correlations. In this paper, we present a novel methodology for workflow augmentation and demonstrate its benefit for event recognition in cataract surgery. The proposed approach increases the frequency of event alternation by creating artificial videos. The original video is split into event segments and a workflow graph is extracted from the original annotations. Finally, the segments are assembled into new videos based on the workflow graph. Compared to the original videos, the frequency of event alternation in the augmented cataract surgery videos increased by 26%. Further, a 3% higher classification accuracy and a 7.8% higher precision was achieved compared to a state-of-the-art approach. Our approach is particularly helpful to increase the occurrence of rare but important events and can be applied to a large variety of use cases.

Chest X-Ray (CXR) is one of the most common diagnostic techniques used in everyday clinical practice all around the world. We hereby present a work which intends to investigate and analyse the use of Deep Learning (DL) techniques to extract information from such images and allow to classify them, trying to keep our methodology as general as possible and possibly also usable in a real world scenario without much effort, in the future. To move in this direction, we trained several beta-Variational Autoencoder (beta-VAE) models on the CheXpert dataset, one of the largest publicly available collection of labeled CXR images; from these models, latent features have been extracted and used to train other Machine Learning models, able to classify the original images from the features extracted by the beta-VAE. Lastly, tree-based models have been combined together in ensemblings to improve the results without the necessity of further training or models engineering. Expecting some drop in pure performance with the respect to state of the art classification specific models, we obtained encouraging results, which show the viability of our approach and the usability of the high level features extracted by the autoencoders for classification tasks.

Ulcerative colitis (UC) is a chronic inflammatory bowel disease characterized by relapsing inflammation of the large intestine. The severity of UC is often represented by the Mayo Endoscopic Subscore (MES) which quantifies mucosal disease activity from endoscopy videos. In clinical trials, an endoscopy video is assigned an MES based upon the most severe disease activity observed in the video. For this reason, severe inflammation spread throughout the colon will receive the same MES as an otherwise healthy colon with severe inflammation restricted to a small, localized segment. Therefore, the extent of disease activity throughout the large intestine, and overall response to treatment, may not be completely captured by the MES. In this work, we aim to automatically estimate UC severity for each frame in an endoscopy video to provide a higher resolution assessment of disease activity throughout the colon. Because annotating severity at the frame-level is expensive, labor-intensive, and highly subjective, we propose a novel weakly supervised, ordinal classification method to estimate frame severity from video MES labels alone. Using clinical trial data, we first achieved 0.92 and 0.90 AUC for predicting mucosal healing and remission of UC, respectively. Then, for severity estimation, we demonstrate that our models achieve substantial Cohen's Kappa agreement with ground truth MES labels, comparable to the inter-rater agreement of expert clinicians. These findings indicate that our framework could serve as a foundation for novel clinical endpoints, based on a more localized scoring system, to better evaluate UC drug efficacy in clinical trials.

Deep neural networks (DNN) have been used successfully in many scientific problems for their high prediction accuracy, but their application to genetic studies remains challenging due to their poor interpretability. In this paper, we consider the problem of scalable, robust variable selection in DNN for the identification of putative causal genetic variants in genome sequencing studies. We identified a pronounced randomness in feature selection in DNN due to its stochastic nature, which may hinder interpretability and give rise to misleading results. We propose an interpretable neural network model, stabilized using ensembling, with controlled variable selection for genetic studies. The merit of the proposed method includes: (1) flexible modelling of the non-linear effect of genetic variants to improve statistical power; (2) multiple knockoffs in the input layer to rigorously control false discovery rate; (3) hierarchical layers to substantially reduce the number of weight parameters and activations to improve computational efficiency; (4) de-randomized feature selection to stabilize identified signals. We evaluated the proposed method in extensive simulation studies and applied it to the analysis of Alzheimer's disease genetics. We showed that the proposed method, when compared to conventional linear and nonlinear methods, can lead to substantially more discoveries. Introduction Recent advances in whole genome sequencing (WGS) technology have led the way to explore the contribution of common and rare variants in both coding and non-coding regions towards risk for complex traits. Large-scale genome sequencing studies, such as the Trans-Omics for Precision Medicine (TOPMed) study and the Alzheimer's Disease Sequencing Project (ADSP), have collected thousands of samples with directly sequenced whole genomes. Genetic variants or genes below a p-value threshold are deemed as associated variants. The marginal association tests are well-known for their simplicity and effectiveness, but they often identify proxy variants that are only correlated with the true causal variants, and the statistical power can be suboptimal. One obstacle for the widespread application of DNN to genetic data is their interpretability.

In the context of finite mixture models one considers the problem of classifying as many observations as possible in the classes of interest while controlling the classification error rate in these same classes. Similar to what is done in the framework of statistical test theory, different type I and type II-like classification error rates can be defined, along with their associated optimal rules, where optimality is defined as minimizing type II error rate while controlling type I error rate at some nominal level. It is first shown that finding an optimal classification rule boils down to searching an optimal region in the observation space where to apply the classical Maximum A Posteriori (MAP) rule. Depending on the misclassification rate to be controlled, the shape of the optimal region is provided, along with a heuristic to compute the optimal classification rule in practice. In particular, a multiclass FDR-like optimal rule is defined and compared to the thresholded MAP rules that is used in most applications. It is shown on both simulated and real datasets that the FDR-like optimal rule may be significantly less conservative than the thresholded MAP rule.

Breast cancer is the second most common cancer among women in the United States; as of January 2021, there are more than 3.8 million women with a history of breast cancer in the United States. Doctors often use ultrasound, mammograms, MRI, or biopsy to find or diagnose breast cancer. In a new study, researchers from NYU and NYU Abu Dhabi (NYUAD) report that they have developed a novel artificial intelligence (AI) system that achieves radiologist-level accuracy in identifying breast cancer in ultrasound images. Their findings are published in the journal Nature Communications, in a paper titled, "Artificial intelligence system reduces false-positive findings in the interpretation of breast ultrasound exams," and was led by Farah Shamout, PhD, NYUAD assistant professor emerging scholar of computer engineering and colleagues. "Though consistently shown to detect mammographically occult cancers, breast ultrasound has been noted to have high false-positive rates, the researchers wrote. "In this work, we present an AI system that achieves radiologist-level accuracy in identifying breast cancer in ultrasound images." "The AI system was developed and evaluated using the NYU Breast Ultrasound Dataset41 consisting of 5,442,907 images within 288,767 breast exams (including both screening and diagnostic exams) collected from 143,203 patients examined between 2012 and 2019 at NYU Langone Health in New York," noted the researchers. The primary goal of the AI system is to reduce the frequency of false-positive findings. It can detect cancer by assigning a probability for malignancy and highlight parts of ultrasound images that are associated with its predictions. When the researchers conducted a reader study to compare its diagnostic accuracy with board-certified breast radiologists, the system achieved higher accuracy than the ten radiologists on average. However, a hybrid model that aggregated the predictions of the AI system and radiologists achieved the best results in accurately detecting cancer in patients. "Our findings highlight the potential of AI to improve the accuracy, consistency, and efficiency of breast ultrasound diagnosis," explained Shamout. "Importantly, AI is not a replacement for the expertise of clinicians.

The novel severe acute respiratory syndrome coronavirus type-2 (SARS-CoV-2) caused a global pandemic that has taken more than 4.5 million lives and severely affected the global economy. To curb the spread of the virus, an accurate, cost-effective, and quick testing for large populations is exceedingly important in order to identify, isolate, and treat infected people. Current testing methods commonly use PCR (Polymerase Chain Reaction) based equipment that have limitations on throughput, cost-effectiveness, and simplicity of procedure which creates a compelling need for developing additional coronavirus disease-2019 (COVID-19) testing mechanisms, that are highly sensitive, rapid, trustworthy, and convenient to use by the public. We propose a COVID-19 testing method using artificial intelligence (AI) techniques on MALDI-ToF (matrix-assisted laser desorption/ionization time-of-flight) data extracted from 152 human gargle samples (60 COVID-19 positive tests and 92 COVID-19 negative tests). Our AI-based approach leverages explainable-AI (X-AI) methods to explain the decision rules behind the predictive algorithm both on a local (per-sample) and global (all-samples) basis to make the AI model more trustworthy. Finally, we evaluated our proposed method using a 70%-30% train-test-split strategy and achieved a training accuracy of 86.79% and a testing accuracy of 91.30%.

High-dimensional classification and feature selection tasks are ubiquitous with the recent advancement in data acquisition technology. In several application areas such as biology, genomics and proteomics, the data are often functional in their nature and exhibit a degree of roughness and non-stationarity. These structures pose additional challenges to commonly used methods that rely mainly on a two-stage approach performing variable selection and classification separately. We propose in this work a novel Gaussian process discriminant analysis (GPDA) that combines these steps in a unified framework. Our model is a two-layer non-stationary Gaussian process coupled with an Ising prior to identify differentially-distributed locations. Scalable inference is achieved via developing a variational scheme that exploits advances in the use of sparse inverse covariance matrices. We demonstrate the performance of our methodology on simulated datasets and two proteomics datasets: breast cancer and SARS-CoV-2. Our approach distinguishes itself by offering explainability as well as uncertainty quantification in addition to low computational cost, which are crucial to increase trust and social acceptance of data-driven tools.

Objectives: This paper develops two algorithms to achieve federated generalized linear mixed effect models (GLMM), and compares the developed model's outcomes with each other, as well as that from the standard R package (`lme4'). Methods: The log-likelihood function of GLMM is approximated by two numerical methods (Laplace approximation and Gaussian Hermite approximation), which supports federated decomposition of GLMM to bring computation to data. Results: Our developed method can handle GLMM to accommodate hierarchical data with multiple non-independent levels of observations in a federated setting. The experiment results demonstrate comparable (Laplace) and superior (Gaussian-Hermite) performances with simulated and real-world data. Conclusion: We developed and compared federated GLMMs with different approximations, which can support researchers in analyzing biomedical data to accommodate mixed effects and address non-independence due to hierarchical structures (i.e., institutes, region, country, etc.).

Most current anomaly detection methods suffer from the curse of dimensionality when dealing with high-dimensional data. We propose an anomaly detection algorithm that can scale to high-dimensional data using concepts from the theory of large deviations. The proposed Large Deviations Anomaly Detection (LAD) algorithm is shown to outperform state of art anomaly detection methods on a variety of large and high-dimensional benchmark data sets. Exploiting the ability of the algorithm to scale to high-dimensional data, we propose an online anomaly detection method to identify anomalies in a collection of multivariate time series. We demonstrate the applicability of the online algorithm in identifying counties in the United States with anomalous trends in terms of COVID-19 related cases and deaths. Several of the identified anomalous counties correlate with counties with documented poor response to the COVID pandemic.