In this article, we propose a novel ECG classification framework for atrial fibrillation (AF) detection using spectro-temporal representation (i.e., time varying spectrum) and deep convolutional networks. In the first step we use a Bayesian spectro-temporal representation based on the estimation of time-varying coefficients of Fourier series using Kalman filter and smoother. Next, we derive an alternative model based on a stochastic oscillator differential equation to accelerate the estimation of the spectro-temporal representation in lengthy signals. Finally, after comparative evaluations of different convolutional architectures, we propose an efficient deep convolutional neural network to classify the 2D spectro-temporal ECG data. The ECG spectro-temporal data are classified into four different classes: AF, non-AF normal rhythm (Normal), non-AF abnormal rhythm (Other), and noisy segments (Noisy). The performance of the proposed methods is evaluated and scored with the PhysioNet/Computing in Cardiology (CinC) 2017 dataset. The experimental results show that the proposed method achieves the overall F1 score of 80.2%, which is in line with the state-of-the-art algorithms.

Inspired by the phenomenon of catastrophic forgetting, we investigate the learning dynamics of neural networks as they train on single classification tasks. Our goal is to understand whether a related phenomenon occurs when data does not undergo a clear distributional shift. We define a `forgetting event' to have occurred when an individual training example transitions from being classified correctly to incorrectly over the course of learning. Across several benchmark data sets, we find that: (i) certain examples are forgotten with high frequency, and some not at all; (ii) a data set's (un)forgettable examples generalize across neural architectures; and (iii) based on forgetting dynamics, a significant fraction of examples can be omitted from the training data set while still maintaining state-of-the-art generalization performance.

An unprecedented booming has been witnessed in the research area of artistic style transfer ever since Gatys et al. introduced the neural method. One of the remaining challenges is to balance a trade-off among three critical aspects---speed, flexibility, and quality: (i) the vanilla optimization-based algorithm produces impressive results for arbitrary styles, but is unsatisfyingly slow due to its iterative nature, (ii) the fast approximation methods based on feed-forward neural networks generate satisfactory artistic effects but bound to only a limited number of styles, and (iii) feature-matching methods like AdaIN achieve arbitrary style transfer in a real-time manner but at a cost of the compromised quality. We find it considerably difficult to balance the trade-off well merely using a single feed-forward step and ask, instead, whether there exists an algorithm that could adapt quickly to any style, while the adapted model maintains high efficiency and good image quality. Motivated by this idea, we propose a novel method, coined MetaStyle, which formulates the neural style transfer as a bilevel optimization problem and combines learning with only a few post-processing update steps to adapt to a fast approximation model with satisfying artistic effects, comparable to the optimization-based methods for an arbitrary style. The qualitative and quantitative analysis in the experiments demonstrates that the proposed approach achieves high-quality arbitrary artistic style transfer effectively, with a good trade-off among speed, flexibility, and quality.

Bayesian methods have been successfully applied to sparsify weights of neural networks and to remove structure units from the networks, e. g. neurons. We apply and further develop this approach for gated recurrent architectures. Specifically, in addition to sparsification of individual weights and neurons, we propose to sparsify preactivations of gates and information flow in LSTM. It makes some gates and information flow components constant, speeds up forward pass and improves compression. Moreover, the resulting structure of gate sparsity is interpretable and depends on the task. Code is available on github: https://github.com/tipt0p/SparseBayesianRNN

Heart disease remains the leading cause of death in the United States. Compared with risk assessment guidelines that require manual calculation of scores, machine learning-based prediction for disease outcomes such as mortality can be utilized to save time and improve prediction accuracy. This study built and evaluated various machine learning models to predict one-year mortality in patients diagnosed with acute myocardial infarction or post myocardial infarction syndrome in the MIMIC-III database. The results of the best performing shallow prediction models were compared to a deep feedforward neural network (Deep FNN) with back propagation. We included a cohort of 5436 admissions. Six datasets were developed and compared. The models applying Logistic Model Trees (LMT) and Simple Logistic algorithms to the combined dataset resulted in the highest prediction accuracy at 85.12% and the highest AUC at .901. In addition, other factors were observed to have an impact on outcomes as well.

Normative modeling has recently been introduced as a promising approach for modeling variation of neuroimaging measures across individuals in order to derive biomarkers of psychiatric disorders. Current implementations rely on Gaussian process regression, which provides coherent estimates of uncertainty needed for the method but also suffers from drawbacks including poor scaling to large datasets and a reliance on fixed parametric kernels. In this paper, we propose a deep normative modeling framework based on neural processes (NPs) to solve these problems. To achieve this, we define a stochastic process formulation for mixed-effect models and show how NPs can be adopted for spatially structured mixed-effect modeling of neuroimaging data. This enables us to learn optimal feature representations and covariance structure for the random-effect and noise via global latent variables. In this scheme, predictive uncertainty can be approximated by sampling from the distribution of these global latent variables. On a publicly available clinical fMRI dataset, we compare the novelty detection performance of multivariate normative models estimated by the proposed NP approach to a baseline multi-task Gaussian process regression approach and show substantial improvements for certain diagnostic problems.

Cervical spondylosis (CS) is a common chronic disease that affects up to two-thirds of the population and poses a serious burden on individuals and society. The early identification has significant value in improving cure rate and reducing costs. However, the pathology is complex, and the mild symptoms increase the difficulty of the diagnosis, especially in the early stage. Besides, the time-consuming and costliness of hospital medical service reduces the attention to the CS identification. Thus, a convenient, low-cost intelligent CS identification method is imperious demanded. In this paper, we present an intelligent method based on the deep learning to identify CS, using the surface electromyography (sEMG) signal. Faced with the complex, high dimensionality and weak usability of the sEMG signal, we proposed and developed a multi-channel EasiCSDeep algorithm based on the convolutional neural network, which consists of the feature extraction, spatial relationship representation and classification algorithm. To the best of our knowledge, this EasiCSDeep is the first effort to employ the deep learning and the sEMG data to identify CS. Compared with previous state-of-the-art algorithm, our algorithm achieves a significant improvement.

Recently, graph neural networks have been adopted in a wide variety of applications ranging from relational representations to modeling irregular data domains such as point clouds and social graphs. However, the space of graph neural network architectures remains highly fragmented impeding the development of optimized implementations similar to what is available for convolutional neural networks. In this work, we present BiGraphNet, a graph neural network architecture that generalizes many popular graph neural network models and enables new efficient operations similar to those supported by ConvNets. By explicitly separating the input and output nodes, BiGraphNet: (i) generalizes the graph convolution to support new efficient operations such as coarsened graph convolutions (similar to strided convolution in convnets), multiple input graphs convolution and graph expansions (unpooling) which can be used to implement various graph architectures such as graph autoencoders, and graph residual nets; and (ii) accelerates and scales the computations and memory requirements in hierarchical networks by performing computations only at specified output nodes.

In this work, we propose an advanced AI based grading system for OCT images. The proposed system is a very deep fully convolutional attentive classification network trained with end to end advanced transfer learning with online random augmentation. It uses quasi random augmentation that outputs confidence values for diseases prevalence during inference. Its a fully automated retinal OCT analysis AI system capable of pathological lesions understanding without any offline preprocessing/postprocessing step or manual feature extraction. We present a state of the art performance on the publicly available Mendeley OCT dataset.

In this paper, we use a well-known Deep Learning technique called Long Short Term Memory (LSTM) recurrent neural networks to find sessions that are prone to code failure in applications that rely on telemetry data for system health monitoring. We also use LSTM networks to extract telemetry patterns that lead to a specific code failure. For code failure prediction, we treat the telemetry events, sequence of telemetry events and the outcome of each sequence as words, sentence and sentiment in the context of sentiment analysis, respectively. Our proposed method is able to process a large set of data and can automatically handle edge cases in code failure prediction. We take advantage of Bayesian optimization technique to find the optimal hyper parameters as well as the type of LSTM cells that leads to the best prediction performance. We then introduce the Contributors and Blockers concepts. In this paper, contributors are the set of events that cause a code failure, while blockers are the set of events that each of them individually prevents a code failure from happening, even in presence of one or multiple contributor(s). Once the proposed LSTM model is trained, we use a greedy approach to find the contributors and blockers. To develop and test our proposed method, we use synthetic (simulated) data in the first step. The synthetic data is generated using a number of rules for code failures, as well as a number of rules for preventing a code failure from happening. The trained LSTM model shows over 99% accuracy for detecting code failures in the synthetic data. The results from the proposed method outperform the classical learning models such as Decision Tree and Random Forest. Using the proposed greedy method, we are able to find the contributors and blockers in the synthetic data in more than 90% of the cases, with a performance better than sequential rule and pattern mining algorithms.