Camera traps are a critical part of wildlife monitoring but analyzing those videos is a big bottleneck. Developing an automated procedure is of the utmost importance to conservation research. While camera traps have become powerful non-invasive tools in research and conservation efforts, they can't yet autonomously label the species they observe. It takes a lot of valuable time to determine whether there are any animals present (or just passing winds), and if there are, which ones. Then, the DrivenData community used cutting edge computer vision techniques to turn those labels into algorithms for automated species detection.

Today, Zensors, a Carnegie Mellon spinout and maker of cloud-based visual sensing technology is announcing the release of its latest deep learning technology, Car Pose Net. Previously, tracking rigid, three dimensional objects (like cars) using only single-view cameras was problematic. Car Pose Net fits 3D pose wireframes to cars, improving tracking results, especially in difficult conditions like snow or partial visual obstructions. This unlocks incredible potential for existing city and autonomous vehicle camera systems. Because the technology can be deployed using legacy camera hardware and Zensors edge or cloud compute platforms, more advanced, accurate, and real time traffic data can be unlocked.

The field of nuclear cardiology has evolved over the last several decades. The field has advanced from basic first-pass radionuclide ventriculography to gated myocardial perfusion imaging with single-photon emission computed tomography (SPECT) with solid state cadmium zinc telluride cameras. There has also been adjunctive use of computed tomography (CT) technology for attenuation correction with the benefit of utilizing the transmission CT image to obtain additional information on the presence or absence of coronary calcification for additional prognostic information. Additionally, there has also been development of gated myocardial perfusion acquisition with positron emission tomography (PET) and use of CT transmission imaging for attenuation correction as well. The utilization of PET cardiac imaging has led to the ability to assess myocardial blood flow, thus improving sensitivity and specificity of myocardial PET imaging.

The overall usage of emergency CT for patients with suspected urinary stones has doubled over the past decade. The growth in imaging volume has resulted in longer turnaround times, additional burden for radiologists, and longer hospital stays, according to the MGH researchers. They sought to investigate the accuracy of a cascading deep-learning system for detecting urinary stones on unenhanced images and also wanted to assess the effect of transfer learning to determine if the performance of pretrained models would be consistent across different types of scanners.

In this paper, a method for malfunctioning smart meter detection, based on Long Short-Term Memory (LSTM) and Temporal Phase Convolutional Neural Network (TPCNN), is proposed originally. This method is very useful for some developing countries where smart meters have not been popularized but in high demand. In addition, it is a new topic that people try to increase the service life span of smart meters to prevent unnecessary waste by detecting malfunctioning meters. We are the first people complete a combination of malfunctioning meters detection and prediction model based on deep learning methods. To the best our knowledge, our approach is the first method that achieves the malfunctioning meter detection of specific residential areas with their residents' data in practice. The procedure proposed creatively in this paper mainly consists of four components: data collecting and cleaning, prediction about electricity consumption based on LSTM, sliding window detection, and single user classification based on CNN. To make better classifying of malfunctioned user meters, we combine recurrence plots as image-input and combine them with sequence-input, which is the first work that applies one and two dimensions as two paths CNN's input for sequence data classification. Finally, many classical methods are compared with the method proposed in this paper. After comparison with classical methods, Elastic Net and Gradient Boosting Regression, the result shows that our method has higher accuracy. The average area under the Receiver Operating Characteristic (ROC) curve is 0.80 and the standard deviation is 0.04. The average area under the Precision-Recall Curve (PRC) is 0.84.

Deep models, while being extremely versatile and accurate, are vulnerable to adversarial attacks: slight perturbations that are imperceptible to humans can completely flip the prediction of deep models. Many attack and defense mechanisms have been proposed, although a satisfying solution still largely remains elusive. In this work, we give strong evidence that during training, deep models maximize the minimum margin in order to achieve high accuracy, but at the same time decrease the \emph{average} margin hence hurting robustness. Our empirical results highlight an intrinsic trade-off between accuracy and robustness for current deep model training. To further address this issue, we propose a new regularizer to explicitly promote average margin, and we verify through extensive experiments that it does lead to better robustness. Our regularized objective remains Fisher-consistent, hence asymptotically can still recover the Bayes optimal classifier.

Deep generative models are rapidly gaining traction in medical imaging. Nonetheless, most generative architectures struggle to capture the underlying probability distributions of volumetric data, exhibit convergence problems, and offer no robust indices of model uncertainty. By comparison, the autoregressive generative model PixelCNN can be extended to volumetric data with relative ease, it readily attempts to learn the true underlying probability distribution and it still admits a Bayesian reformulation that provides a principled framework for reasoning about model uncertainty. Our contributions in this paper are two fold: first, we extend PixelCNN to work with volumetric brain magnetic resonance imaging data. Second, we show that reformulating this model to approximate a deep Gaussian process yields a measure of uncertainty that improves the performance of semi-supervised learning, in particular classification performance in settings where the proportion of labelled data is low. We quantify this improvement across classification, regression, and semantic segmentation tasks, training and testing on clinical magnetic resonance brain imaging data comprising T1-weighted and diffusion-weighted sequences.

--Image reconstruction from undersampled k - space data has been playing an important role for fast MRI. Recently, deep learning has demonstrated tremendous success in various fields and has also shown potential to significantly speed up MRI reconstruction with reduced measurements. This article gives an overview of deep learning -based image reconstruction methods for MRI. Three types of deep learning -based approaches are reviewed, the data - driven, model - driven and integrated approaches. T he main structure of each network in three approaches is explained and the analysis of common parts of reviewed networks and differences in - between are highlighted. Based on the review, a number of signal processing issues are discussed for maximizing the potential of deep reconstruction for fast MRI. The discussion may facilitate further development of "optimal" network and performance analysis from a theoretical point of view. I. INTRODUCTION Since its inception in the early 70's, magnetic resonance imaging (MRI) has revolutionized radiology and medicine. However, MRI is known to be a slow imaging modality and many techniques have been devel oped to reconstruct the desired image from undersampled measured data to improve the imaging speed [1]. During the past decades, compressed sensing (CS) has become an important strategy for fast MR imaging based on the sparsity prior. However, the iterative solution procedure takes a relatively long time to achieve a high -quality reconstruction, and the selection of the regularization parameter is empirical.

Next-generation cosmic microwave background (CMB) experiments will have lower noise and therefore increased sensitivity, enabling improved constraints on fundamental physics parameters such as the sum of neutrino masses and the tensor-to-scalar ratio r. Achieving competitive constraints on these parameters requires high signal-to-noise extraction of the projected gravitational potential from the CMB maps. Standard methods for reconstructing the lensing potential employ the quadratic estimator (QE). However, the QE performs suboptimally at the low noise levels expected in upcoming experiments. Other methods, like maximum likelihood estimators (MLE), are under active development. In this work, we demonstrate reconstruction of the CMB lensing potential with deep convolutional neural networks (CNN) - ie, a ResUNet. The network is trained and tested on simulated data, and otherwise has no physical parametrization related to the physical processes of the CMB and gravitational lensing. We show that, over a wide range of angular scales, ResUNets recover the input gravitational potential with a higher signal-to-noise ratio than the QE method, reaching levels comparable to analytic approximations of MLE methods. We demonstrate that the network outputs quantifiably different lensing maps when given input CMB maps generated with different cosmologies. We also show we can use the reconstructed lensing map for cosmological parameter estimation. This application of CNN provides a few innovations at the intersection of cosmology and machine learning. First, while training and regressing on images, we predict a continuous-variable field rather than discrete classes. Second, we are able to establish uncertainty measures for the network output that are analogous to standard methods. We expect this approach to excel in capturing hard-to-model non-Gaussian astrophysical foreground and noise contributions.

We use deep neural networks to classify time series generated by discrete and continuous dynamical systems based on their chaotic behaviour. Our approach to circumvent the lack of precise models for some of the most challenging real-life applications is to train different neural networks on a data set from a dynamical system with a basic or low-dimensional phase space and then use these networks to classify time series of a dynamical system with more intricate or high-dimensional phase space. We illustrate this extrapolation approach using the logistic map, the sine-circle map, the Lorenz system, and the Kuramoto-Sivashinsky equation. We observe that the proposed convolutional neural network with large kernel size outperforms state-of-the-art neural networks for time series classification and is able to classify time series as chaotic or non-chaotic with high accuracy. Introduction Data and in particular time series are generated from numerous observations and experiments across different scientific fields such as atmospheric and oceanic sciences for climate predictions, nuclear fusion for control and safety, biology and medicine for diagnosis. Fourier transforms, radial basis functions approximation and standard numerical techniques have been extensively applied to perform short and long term predictions of chaotic time series [1, 2, 3, 4]. On the other hand, the spectacular success of machine learning and deep learning techniques to image classification [5, 6], which have recently surpassed human-level performance on the ImageNet data set [7], has inspired the development of neural network techniques for time series forecasting [8, 9] and classification [10]. Recently, deep learning approaches have been used to solve partial differential equations in high dimensions [11, 12, 13] and identify hidden physics models from experimental data [14, 15, 16, 17].