Much of the focus in the design of deep neural networks has been on improving accuracy, leading to more powerful yet highly complex network architectures that are difficult to deploy in practical scenarios, particularly on edge devices such as mobile and other consumer devices, given their high computational and memory requirements. As a result, there has been a recent interest in the design of quantitative metrics for evaluating deep neural networks that accounts for more than just model accuracy as the sole indicator of network performance. In this study, we continue the conversation towards universal metrics for evaluating the performance of deep neural networks for practical usage. In particular, we propose a new balanced metric called NetScore, which is designed specifically to provide a quantitative assessment of the balance between accuracy, computational complexity, and network architecture complexity of a deep neural network. In what is one of the largest comparative analysis between deep neural networks in literature, the NetScore metric, the top-1 accuracy metric, and the popular information density metric were compared across a diverse set of 50 different deep convolutional neural networks for image classification on the ImageNet Large Scale Visual Recognition Challenge (ILSVRC 2012) dataset. The evaluation results across these three metrics for this diverse set of networks are presented in this study to act as a reference guide for practitioners in the field. The proposed NetScore metric, along with the other tested metrics, are by no means perfect, but the hope is to push the conversation towards better universal metrics for evaluating deep neural networks for use in practical scenarios to help guide practitioners in model design.

In many forecasting applications, it is valuable to predict not only the value of a signal at a certain time point in the future, but also the values leading up to that point. This is especially true in clinical applications, where the future state of the patient can be less important than the patient's overall trajectory. This requires multi-step forecasting, a forecasting variant where one aims to predict multiple values in the future simultaneously. Standard methods to accomplish this can propagate error from prediction to prediction, reducing quality over the long term. In light of these challenges, we propose multi-output deep architectures for multi-step forecasting in which we explicitly model the distribution of future values of the signal over a prediction horizon. We apply these techniques to the challenging and clinically relevant task of blood glucose forecasting. Through a series of experiments on a real-world dataset consisting of 550K blood glucose measurements, we demonstrate the effectiveness of our proposed approaches in capturing the underlying signal dynamics. Compared to existing shallow and deep methods, we find that our proposed approaches improve performance individually and capture complementary information, leading to a large improvement over the baseline when combined (4.87 vs. 5.31 absolute percentage error (APE)). Overall, the results suggest the efficacy of our proposed approach in predicting blood glucose level and multi-step forecasting more generally.

Automated service classification plays a crucial role in service management such as service discovery, selection, and composition. In recent years, machine learning techniques have been used for service classification. However, they can only predict around 10 to 20 service categories due to the quality of feature engineering and the imbalance problem of service dataset. In this paper, we present a deep neural network ServeNet with a novel dataset splitting algorithm to deal with these issues. ServeNet can automatically abstract low-level representation to high-level features, and then predict service classification based on the service datasets produced by the proposed splitting algorithm. To demonstrate the effectiveness of our approach, we conducted a comprehensive experimental study on 10,000 real-world services in 50 categories. The result shows that ServeNet can achieve higher accuracy than other machine learning methods.

Recurrent neural networks have gained widespread use in modeling sequence data across various domains. While many successful recurrent architectures employ a notion of gating, the exact mechanism that enables such remarkable performance is not well understood. We develop a theory for signal propagation in recurrent networks after random initialization using a combination of mean field theory and random matrix theory. To simplify our discussion, we introduce a new RNN cell with a simple gating mechanism that we call the minimalRNN and compare it with vanilla RNNs. Our theory allows us to define a maximum timescale over which RNNs can remember an input. We show that this theory predicts trainability for both recurrent architectures. We show that gated recurrent networks feature a much broader, more robust, trainable region than vanilla RNNs, which corroborates recent experimental findings. Finally, we develop a closed-form critical initialization scheme that achieves dynamical isometry in both vanilla RNNs and minimalRNNs. We show that this results in significantly improvement in training dynamics. Finally, we demonstrate that the minimalRNN achieves comparable performance to its more complex counterparts, such as LSTMs or GRUs, on a language modeling task.

The representation of nonlinear sub-grid processes, especially clouds, has been a major source of uncertainty in climate models for decades. Cloud-resolving models better represent many of these processes and can now be run globally but only for short-term simulations of at most a few years because of computational limitations. Here we demonstrate that deep learning can be used to capture many advantages of cloud-resolving modeling at a fraction of the computational cost. We train a deep neural network to represent all atmospheric sub-grid processes in a climate model by learning from a multi-scale model in which convection is treated explicitly. The trained neural network then replaces the traditional sub-grid parameterizations in a global general circulation model in which it freely interacts with the resolved dynamics and the surface-flux scheme. The prognostic multi-year simulations are stable and closely reproduce not only the mean climate of the cloud-resolving simulation but also key aspects of variability, including precipitation extremes and the equatorial wave spectrum. Furthermore, the neural network approximately conserves energy despite not being explicitly instructed to. Finally, we show that the neural network parameterization generalizes to new surface forcing patterns but struggles to cope with temperatures far outside its training manifold. Our results show the feasibility of using deep learning for climate model parameterization. In a broader context, we anticipate that data-driven Earth System Model development could play a key role in reducing climate prediction uncertainty in the coming decade.

We investigate learning of the differential geometric structure of a data manifold embedded in a high-dimensional Euclidean space. We first analyze kernel-based algorithms and show that under the usual regularizations, non-probabilistic methods cannot recover the differential geometric structure, but instead find mostly linear manifolds or spaces equipped with teleports. To properly learn the differential geometric structure, non-probabilistic methods must apply regularizations that enforce large gradients, which go against common wisdom. We repeat the analysis for probabilistic methods and find that under reasonable priors, the geometric structure can be recovered. Fully exploiting the recovered structure, however, requires the development of stochastic extensions to classic Riemannian geometry. We take early steps in that regard. Finally, we partly extend the analysis to modern models based on neural networks, thereby highlighting geometric and probabilistic shortcomings of current deep generative models.

IntraVascular UltraSound (IVUS) is one of the most effective imaging modalities that provides assistance to experts in order to diagnose and treat cardiovascular diseases. We address a central problem in IVUS image analysis with Fully Convolutional Network (FCN): automatically delineate the lumen and media-adventitia borders in IVUS images, which is crucial to shorten the diagnosis process or benefits a faster and more accurate 3D reconstruction of the artery. Particularly, we propose an FCN architecture, called IVUS-Net, followed by a post-processing contour extraction step, in order to automatically segments the interior (lumen) and exterior (media-adventitia) regions of the human arteries. We evaluated our IVUS-Net on the test set of a standard publicly available dataset containing 326 IVUS B-mode images with two measurements, namely Jaccard Measure (JM) and Hausdorff Distances (HD). The evaluation result shows that IVUS-Net outperforms the state-of-the-art lumen and media segmentation methods by 4% to 20% in terms of HD distance. IVUS-Net performs well on images in the test set that contain a significant amount of major artifacts such as bifurcations, shadows, and side branches that are not common in the training set. Furthermore, using a modern GPU, IVUS-Net segments each IVUS frame only in 0.15 seconds. The proposed work, to the best of our knowledge, is the first deep learning based method for segmentation of both the lumen and the media vessel walls in 20 MHz IVUS B-mode images that achieves the best results without any manual intervention. Code is available at https://github.com/Kulbear/ivus-segmentation-icsm2018

With the rising popularity of machine learning and the ever increasing demand for computational power, there is a growing need for hardware optimized implementations of neural networks and other machine learning models. As the technology evolves, it is also plausible that machine learning or artificial intelligence will soon become consumer electronic products and military equipment, in the form of well-trained models. Unfortunately, the modern fabless business model of manufacturing hardware, while economic, leads to deficiencies in security through the supply chain. In this paper, we illuminate these security issues by introducing hardware Trojan attacks on neural networks, expanding the current taxonomy of neural network security to incorporate attacks of this nature. To aid in this, we develop a novel framework for inserting malicious hardware Trojans in the implementation of a neural network classifier. We evaluate the capabilities of the adversary in this setting by implementing the attack algorithm on convolutional neural networks while controlling a variety of parameters available to the adversary. Our experimental results show that the proposed algorithm could effectively classify a selected input trigger as a specified class on the MNIST dataset by injecting hardware Trojans into $0.03\%$, on average, of neurons in the 5th hidden layer of arbitrary 7-layer convolutional neural networks, while undetectable under the test data. Finally, we discuss the potential defenses to protect neural networks against hardware Trojan attacks.

Multigrid modeling algorithms are a technique used to accelerate relaxation models running on a hierarchy of similar graphlike structures. We introduce and demonstrate a new method for training neural networks which uses multilevel methods. Using an objective function derived from a graph-distance metric, we perform orthogonally-constrained optimization to find optimal prolongation and restriction maps between graphs. We compare and contrast several methods for performing this numerical optimization, and additionally present some new theoretical results on upper bounds of this type of objective function. Once calculated, these optimal maps between graphs form the core of Multiscale Artificial Neural Network (MsANN) training, a new procedure we present which simultaneously trains a hierarchy of neural network models of varying spatial resolution. Parameter information is passed between members of this hierarchy according to standard coarsening and refinement schedules from the multiscale modelling literature. In our machine learning experiments, these models are able to learn faster than default training, achieving a comparable level of error in an order of magnitude fewer training examples.

We introduce autoregressive implicit quantile networks (AIQN), a fundamentally different approach to generative modeling than those commonly used, that implicitly captures the distribution using quantile regression. AIQN is able to achieve superior perceptual quality and improvements in evaluation metrics, without incurring a loss of sample diversity. The method can be applied to many existing models and architectures. In this work we extend the PixelCNN model with AIQN and demonstrate results on CIFAR-10 and ImageNet using Inception score, FID, non-cherry-picked samples, and inpainting results. We consistently observe that AIQN yields a highly stable algorithm that improves perceptual quality while maintaining a highly diverse distribution.