As machine learning methods see greater adoption and implementation in high stakes applications such as medical image diagnosis, the need for model inter-pretability and explanation has become more critical. Classical approaches that assess feature importance ( e.g., saliency maps) do not explain how and why a particular region of an image is relevant to the prediction. We propose a method that explains the outcome of a classification black-box by gradually exaggerating the semantic effect of a given class. Given a query input to a classifier, our method produces a progressive set of plausible variations of that query, which gradually change the posterior probability from its original class to its negation. These counter-factually generated samples preserve features unrelated to the classification decision, such that a user can employ our method as a "tuning knob" to traverse a data manifold while crossing the decision boundary. Our method is model agnostic and only requires the output value and gradient of the predictor with respect to its input. With the explosive adoption of deep learning for real-world applications, explanation and model interpretability have received substantial attention from the research community (Kim, 2015; Doshi-V elez & Kim, 2017; Molnar, 2019; Guidotti et al., 2019). Explaining an outcome of a model in high stake applications, such as medical diagnosis from radiology images, is of paramount importance to detect hidden biases in data (Cramer et al., 2018), evaluate the fairness of the model (Doshi-V elez & Kim, 2017), and build trust in the system (Glass et al., 2008). For example, consider evaluating a computer-aided diagnosis of Alzheimer's disease from medical images. The physician should be able to assess whether or not the model pays attention to age-related or disease-related variations in an image in order to trust the system. Given a query, our model provides an explanation that gradually exaggerates the semantic effect of one class, which is equivalent to traversing the decision boundary from side to another. Although not always clear, there are subtle differences between interpretability and explanation (Turner, 2016). While the former mainly focuses on building or approximating models that are locally or globally interpretable (Ribeiro et al., 2016), the latter aims at explaining a predictor a-posteriori.

Physics-based simulations are often used to model and understand complex physical systems and processes in domains like fluid dynamics. Such simulations, although used frequently, have many limitations which could arise either due to the inability to accurately model a physical process owing to incomplete knowledge about certain facets of the process or due to the underlying process being too complex to accurately encode into a simulation model. In such situations, it is often useful to rely on machine learning methods to fill in the gap by learning a model of the complex physical process directly from simulation data. However, as data generation through simulations is costly, we need to develop models, being cognizant of data paucity issues. In such scenarios it is often helpful if the rich physical knowledge of the application domain is incorporated in the architectural design of machine learning models. Further, we can also use information from physics-based simulations to guide the learning process using aggregate supervision to favorably constrain the learning process. In this paper, we propose PhyDNN, a deep learning model using physics-guided structural priors and physics-guided aggregate supervision for modeling the drag forces acting on each particle in a Computational Fluid Dynamics-Discrete Element Method(CFD-DEM). We conduct extensive experiments in the context of drag force prediction and showcase the usefulness of including physics knowledge in our deep learning formulation both in the design and through learning process. Our proposed PhyDNN model has been compared to several state-of-the-art models and achieves a significant performance improvement of 8.46% on average across all baseline models. The source code has been made available and the dataset used is detailed in [1, 2].

While there has been much recent progress using deep learning techniques to separate speech and music audio signals, these systems typically require large collections of isolated sources during the training process. When extending audio source separation algorithms to more general domains such as environmental monitoring, it may not be possible to obtain isolated signals for training. Here, we propose objective functions and network architectures that enable training a source separation system with weak labels. In this scenario, weak labels are defined in contrast with strong time-frequency (TF) labels such as those obtained from isolated sources, and refer either to frame-level weak labels where one only has access to the time periods when different sources are active in an audio mixture, or to clip-level weak labels that only indicate the presence or absence of sounds in an entire audio clip. We train a separator that estimates a TF mask for each type of sound event, using a sound event classifier as an assessor of the separator's performance to bridge the gap between the TF-level separation and the ground truth weak labels only available at the frame or clip level. Our objective function requires the classifier applied to a separated source to assign high probability to the class corresponding to that source and low probability to all other classes. The objective function also enforces that the separated sources sum up to the mixture. We benchmark the performance of our algorithm using synthetic mixtures of overlapping events created from a database of sounds recorded in urban environments. Compared to training a network using isolated sources, our model achieves somewhat lower but still significant SI-SDR improvement, even in scenarios with significant sound event overlap.

The approach makes use of the deep neural network to approximate solutions of PDEs through the compositional construction and employs least-squares functionals as loss functions to determine parameters of the deep neural network. There are various least-squares functionals for a partial differential equation. This paper focuses on the so-called first-order system least-squares (FOSLS) functional studied in [3], which is based on a first-order system of scalar second-order elliptic PDEs. Numerical results for second-order elliptic PDEs in one dimension are presented.

Continuous representations have been widely adopted in recommender systems where a large number of entities are represented using embedding vectors. As the cardinality of the entities increases, the embedding components can easily contain millions of parameters and become the bottleneck in both storage and inference due to large memory consumption. This work focuses on post-training 4-bit quantization on the continuous embeddings. We propose row-wise uniform quantization with greedy search and codebook-based quantization that consistently outperforms state-of-the-art quantization approaches on reducing accuracy degradation. We deploy our uniform quantization technique on a production model in Facebook and demonstrate that it can reduce the model size to only 13.89% of the single-precision version while the model quality stays neutral.

Generative adversarial networks (GANs) are deep neural networks that allow us to sample from an arbitrary probability distribution without explicitly estimating the distribution. There is a generator that takes a latent vector as input and transforms it into a valid sample from the distribution. There is also a discriminator that is trained to discriminate such fake samples from true samples of the distribution; at the same time, the generator is trained to generate fakes that the discriminator cannot tell apart from the true samples. Instead of learning a global generator, a recent approach involves training multiple generators each responsible from one part of the distribution. In this work, we review such approaches and propose the hierarchical mixture of generators, inspired from the hierarchical mixture of experts model, that learns a tree structure implementing a hierarchical clustering with soft splits in the decision nodes and local generators in the leaves. Since the generators are combined softly, the whole model is continuous and can be trained using gradient-based optimization, just like the original GAN model. Our experiments on five image data sets, namely, MNIST, FashionMNIST, UTZap50K, Oxford Flowers, and CelebA, show that our proposed model generates samples of high quality and diversity in terms of popular GAN evaluation metrics. The learned hierarchical structure also leads to knowledge extraction.

Training of deep models for classification tasks is hindered by local minima problems and vanishing gradients, while unsupervised layer-wise pretraining does not exploit information from class labels. Here, we propose a new regularization technique, called diversifying regularization (DR), which applies a penalty on hidden units at any layer if they obtain similar features for different types of data. For generative models, DR is defined as divergence over the variational posteriori distributions and included in the maximum likelihood estimation as a prior. Thus, DR includes class label information for greedy pretraining of deep belief networks which result in a better weight initialization for fine-tuning methods. On the other hand, for discriminative training of deep neural networks, DR is defined as a distance over the features and included in the learning objective. With our experimental tests, we show that DR can help the backpropagation to cope with vanishing gradient problems and to provide faster convergence and smaller generalization errors.

Radio-frequency dosimetry is an important process in human safety and for compliance of related products. Recently, computational human models generated from medical images have often been used for such assessment, especially to consider the inter-variability of subjects. However, the common procedure to develop personalized models is time consuming because it involves excessive segmentation of several components that represent different biological tissues, which limits the inter-variability assessment of radiation safety based on personalized dosimetry. Deep learning methods have been shown to be a powerful approach for pattern recognition and signal analysis. Convolutional neural networks with deep architecture are proven robust for feature extraction and image mapping in several biomedical applications. In this study, we develop a learning-based approach for fast and accurate estimation of the dielectric properties and density of tissues directly from magnetic resonance images in a single shot. The smooth distribution of the dielectric properties in head models, which is realized using a process without tissue segmentation, improves the smoothness of the specific absorption rate (SAR) distribution compared with that in the commonly used procedure. The estimated SAR distributions, as well as that averaged over 10-g of tissue in a cubic shape, are found to be highly consistent with those computed using the conventional methods that employ segmentation.

Where dealing with temporal sequences it is fair to assume that the same kind of deformations that motivated the development of the Dynamic Time Warp algorithm could be relevant also in the calculation of the dot product ("convolution") in a 1-D convolution layer. In this work a method is proposed for aligning the convolution filter and the input where they are locally out of phase utilising an algorithm similar to the Dynamic Time Warp. The proposed method enables embedding a non-parametric warping of temporal sequences for increasing similarity directly in deep networks and can expand on the generalisation capabilities and the capacity of standard 1-D convolution layer where local sequential deformations are present in the input. Experimental results demonstrate the proposed method exceeds or matches the standard 1-D convolution layer in terms of the maximum accuracy achieved on a number of time series classification tasks. In addition the impact of different hyperparameters settings is investigated given different datasets and the results support the conclusions of previous work done in relation to the choice of DTW parameter values. The proposed layer can be freely integrated with other typical layers to compose deep artificial neural networks of an arbitrary architecture that are trained using standard stochastic gradient descent.

Neural Machine Translation (NMT) is a deep learning model that prov ides a robust method for machine translation using recurrent neural ne tworks (RNNs). Originally proposed in [1], NMT relies primarily on an encoder-decoder ar chi-tecture that provides increased fluency over phrase-based sys tems. This was implemented successfully in [2] for fast, accurate use on very large datasets. However, it has been suggested that there is significant redundan cy in the current method of neural network parametrization [3], presenting t he opportunity for significant speedup. Tensor Train (TT) decomposition [4] is a method by which large tenso rs can be approximated by the product of a'train' of smaller matrices (see Section 2.2). 1 TTdecomposition has been proposed as a method of speeding up an d reducing the memory usage of machine translation systems with dense weight matrices by reducing the number of parameters required to describe the sy stem [3].