In the past few years, generative models like Generative Adversarial Networks (GANs) have dramatically advanced our ability to represent and parameterize high-dimensional, non-linear image manifolds. As a result, they have been widely adopted across a variety of applications, ranging from challenging inverse problems like image completion, to being used as a prior in problems such as anomaly detection and adversarial defense. A recurring theme in many of these applications is the notion of projecting an image observation onto the manifold that is inferred by the generator. In this context, Projected Gradient Descent (PGD) has been the most popular approach, which essentially searches for a latent representation with the goal of minimizing discrepancy between a generated image and the given observation. However, PGD is an extremely brittle optimization technique that fails to identify the right projection when the observation is corrupted, even by a small amount. Unfortunately, such corruptions are common in the real world, for example arbitrary images with unknown crops, rotations, missing pixels, or other kinds of distribution shifts requiring a more robust projection technique. In this paper we propose corruption-mimicking, a new strategy that utilizes a surrogate network to approximate the unknown corruption directly at test time, without the need for additional supervision or data augmentation. The proposed projection technique significantly improves the robustness of PGD under a wide variety of corruptions, thereby enabling a more effective use of GANs in real-world applications. More importantly, we show that our approach produces state-of-the-art performance in several GAN-based applications -- anomaly detection, domain adaptation, and adversarial defense, that rely on an accurate projection.

Systems incorporating Artificial Intelligence (AI) and machine learning (ML) techniques are increasingly used to guide decision-making in the healthcare sector. While AI-based systems provide powerful and promising results with regard to their classification and prediction accuracy (e.g., in differentiating between different disorders in human gait), most share a central limitation, namely their black-box character. Understanding which features classification models learn, whether they are meaningful and consequently whether their decisions are trustworthy is difficult and often impossible to comprehend. This severely hampers their applicability as decision-support systems in clinical practice. There is a strong need for AI-based systems to provide transparency and justification of predictions, which are necessary also for ethical and legal compliance. As a consequence, in recent years the field of explainable AI (XAI) has gained increasing importance. The primary aim of this article is to investigate whether XAI methods can enhance transparency, explainability and interpretability of predictions in automated clinical gait classification. We utilize a dataset comprising bilateral three-dimensional ground reaction force measurements from 132 patients with different lower-body gait disorders and 62 healthy controls. In our experiments, we included several gait classification tasks, employed a representative set of classification methods, and a well-established XAI method - Layer-wise Relevance Propagation - to explain decisions at the signal (input) level. The presented approach exemplifies how XAI can be used to understand and interpret state-of-the-art ML models trained for gait classification tasks, and shows that the features that are considered relevant for machine learning models can be attributed to meaningful and clinically relevant biomechanical gait characteristics.

Along with the advent of deep neural networks came various methods of exploitation, such as fooling the classifier or co n-taminating its training data. Another such attack is known a s model extraction, where provided API access to some black box neural network, the adversary extracts the underlying model. This is done by querying the model in such a way that the underlying neural network provides enough information to the adversary to be reconstructed. While several works have achieved impressive results with neural network extraction in the propositional domain, this problem has no t yet been considered over the relational domain, where data samples are no longer considered to be independent and identically distributed (iid). Graph Neural Networks (GNNs) ar e a popular deep learning framework to perform machine learning tasks over relational data. In this work, we formalize an instance of GNN extraction, present a solution with preliminary results, and discuss our assumptions and future direc - tions.

Neural architecture search (NAS) aims to discover network architectures with desired properties such as high accuracy or low latency. Recently, differentiable NAS (DNAS) has demonstrated promising results while maintaining a search cost orders of magnitude lower than reinforcement learning (RL) based NAS. However, DNAS models can only optimize differentiable loss functions in search, and they require an accurate differentiable approximation of non-differentiable criteria. In this work, we present UNAS, a unified framework for NAS, that encapsulates recent DNAS and RL-based approaches under one framework. Our framework brings the best of both worlds, and it enables us to search for architectures with both differentiable and non-differentiable criteria in one unified framework while maintaining a low search cost. Further, we introduce a new objective function for search based on the generalization gap that prevents the selection of architectures prone to overfitting. We present extensive experiments on the CIFAR-10, CIFAR-100 and ImageNet datasets and we perform search in two fundamentally different search spaces. We show that UNAS obtains the state-of-the-art average accuracy on all three datasets when compared to the architectures searched in the DARTS space. Moreover, we show that UNAS can find an efficient and accurate architecture in the ProxylessNAS search space, that outperforms existing MobileNetV2 based architectures.

Although Deep Neural Networks have seen great success in recent years through various changes in overall architectures and optimization strategies, their fundamental underlying design remains largely unchanged. Computational neuroscience on the other hand provides more biologically realistic models of neural processing mechanisms, but they are still high level abstractions of the actual experimentally observed behaviour. Here a model is proposed that bridges Neuroscience, Machine Learning and Evolutionary Algorithms to evolve individual soma and synaptic compartment models of neurons in a scalable manner. Instead of attempting to manually derive models for all the observed complexity and diversity in neural processing, we propose an Evolvable Neural Unit (ENU) that can approximate the function of each individual neuron and synapse. We demonstrate that this type of unit can be evolved to mimic Integrate-And-Fire neurons and synaptic Spike-Timing-Dependent Plasticity. Additionally, by constructing a new type of neural network where each synapse and neuron is such an evolvable neural unit, we show it is possible to evolve an agent capable of learning to solve a T-maze environment task. This network independently discovers spiking dynamics and reinforcement type learning rules, opening up a new path towards biologically inspired artificial intelligence.

--The great success of deep learning poses urgent challenges for understanding its working mechanism and rationality. The depth, structure, and massive size of the data are recognized to be three key ingredients for deep learning. In this paper, we aim at rigorous verification of the importance of massive data in embodying the out-performance of deep learning. T o approximate and learn spatially sparse and smooth functions, we establish a novel sampling theorem in learning theory to show the necessity of massive data. We then prove that implementing the classical empirical risk minimization on some deep nets facilitates in realization of the optimal learning rates derived in the sampling theorem. This perhaps explains why deep learning performs so well in the era of big data. With the rapid development of data mining and knowledge discovery, data of massive size are collected in various disciplines [50], including medical diagnosis, financial market analysis, computer vision, natural language processing, time series forecasting, and search engines. These massive data bring additional opportunities to discover subtle data features which cannot be reflected by data of small size while creating a crucial challenge on machine learning to develop learning schemes to realize benefits by exploring the use of massive data. Although numerous learning schemes such as distributed learning [26], localized learning [32] and sub-sampling [14] have been proposed to handle massive data, all these schemes focused on the tractability rather than the benefit of massiveness. Therefore, it remains open to explore the benefits brought from massive data and to develop feasible learning strategies for realizing these benefits. Deep learning [18], characterized by training deep neural networks (deep nets for short) to extract data features by using rich computational resources such as computational power of modern graphical processor units (GPUs) and custom processors, has made remarkable success in computer vision [23], speech recognition [24] and game theory [40], practically showing its power in tackling massive data. C.K. Chui is also associated with the Department of Statistics, Stanford University, CA 94305, USA. Shao-Bo Lin is with the Center of Intelligent Decision-making and Machine Learning, School of Management, Xi'an Jiaotong University, Xi'an, China.

In this paper we propose a new model-based unsupervised learning method, called VarNet, for the solution of partial differential equations (PDEs) using deep neural networks (NNs). Particularly, we propose a novel loss function that relies on the variational (integral) form of PDEs as apposed to their differential form which is commonly used in the literature. Our loss function is discretization-free, highly parallelizable, and more effective in capturing the solution of PDEs since it employs lower-order derivatives and trains over measure non-zero regions of space-time. Given this loss function, we also propose an approach to optimally select the space-time samples, used to train the NN, that is based on the feedback provided from the PDE residual. The models obtained using VarNet are smooth and do not require interpolation. They are also easily differentiable and can directly be used for control and optimization of PDEs. Finally, VarNet can straight-forwardly incorporate parametric PDE models making it a natural tool for model order reduction (MOR) of PDEs. We demonstrate the performance of our method through extensive numerical experiments for the advection-diffusion PDE as an important case-study.

Abstract-- VLSI mask optimization is one of the most critical stages in manufacturability aware design, which is costly due to the complicated mask optimization and lithography simulation. Recent researches have shown prominent advantages of machine learning techniques dealing with complicated and big data problems, which bring potential of dedicated machine learning solution for DFM problems and facilitate the VLSI design cycle. In this paper, we focus on a heterogeneous OPC framework that assists mask layout optimization. Preliminary results show the efficiency and effectiveness of proposed frameworks that have the potential to be alternatives to existing EDA solutions. I Introduction VLSI mask optimization is one of the most critical stages in manufacturability aware design, which is costly due to the complicated mask optimization and lithography simulation. Recent studies have shown prominent advantages of machine learning techniques dealing with complicated and big data problems, which bring the potential of dedicated machine learning solution for DFM problems and facilitate the VLSI design cycle [1, 2].

Breast CT provides image volumes with isotropic resolution in high contrast, enabling detection of calcification (down to a few hundred microns in size) and subtle density differences. Since breast is sensitive to x-ray radiation, dose reduction of breast CT is an important topic, and for this purpose low-dose few-view scanning is a main approach. In this article, we propose a Deep Efficient End-to-end Reconstruction (DEER) network for low-dose few-view breast CT. The major merits of our network include high dose efficiency, excellent image quality, and low model complexity. By the design, the proposed network can learn the reconstruction process in terms of as less as O(N) parameters, where N is the size of an image to be reconstructed, which represents orders of magnitude improvements relative to the state-of-the-art deep-learning based reconstruction methods that map projection data to tomographic images directly. As a result, our method does not require expensive GPUs to train and run. Also, validated on a cone-beam breast CT dataset prepared by Koning Corporation on a commercial scanner, our method demonstrates competitive performance over the state-of-the-art reconstruction networks in terms of image quality.

Applying machine learning models to meteorological data brings many opportunities to the Geosciences field, such as predicting future weather conditions more accurately. In recent years, modeling meteorological data with deep neural networks has become a relevant area of investigation. These works apply either recurrent neural networks (RNNs) or some hybrid approach mixing RNNs and convolutional neural networks (CNNs). In this work, we propose STConvS2S (short for Spatiotemporal Convolutional Sequence to Sequence Network), a new deep learning architecture built for learning both spatial and temporal data dependencies in weather data, using fully convolutional layers. Computational experiments using observations of air temperature and rainfall show that our architecture captures spatiotemporal context and outperforms baseline models and the state-of-art architecture for weather forecasting task.