Automatic heart sound abnormality detection can play a vital role in the early diagnosis of heart diseases, particularly in low-resource settings. The state-of-the-art algorithms for this task utilize a set of Finite Impulse Response (FIR) band-pass filters as a front-end followed by a Convolutional Neural Network (CNN) model. In this work, we propound a novel CNN architecture that integrates the front-end bandpass filters within the network using time-convolution (tConv) layers, which enables the FIR filter-bank parameters to become learnable. Different initialization strategies for the learnable filters, including random parameters and a set of predefined FIR filter-bank coefficients, are examined. Using the proposed tConv layers, we add constraints to the learnable FIR filters to ensure linear and zero phase responses. Experimental evaluations are performed on a balanced 4-fold cross-validation task prepared using the PhysioNet/CinC 2016 dataset. Results demonstrate that the proposed models yield superior performance compared to the state-of-the-art system, while the linear phase FIR filterbank method provides an absolute improvement of 9.54% over the baseline in terms of an overall accuracy metric.

We present a novel event embedding algorithm for crime data that can jointly capture time, location, and the complex free-text component of each event. The embedding is achieved by regularized Restricted Boltzmann Machines (RBMs), and we introduce a new way to regularize by imposing a $\ell_1$ penalty on the conditional distributions of the observed variables of RBMs. This choice of regularization performs feature selection and it also leads to efficient computation since the gradient can be computed in a closed form. The feature selection forces embedding to be based on the most important keywords, which captures the common modus operandi (M. O.) in crime series. Using numerical experiments on a large-scale crime dataset, we show that our regularized RBMs can achieve better event embedding and the selected features are highly interpretable from human understanding.

In this work, we propose a deep neural network architecture motivated by primal-dual splitting methods from convex optimization. We show theoretically that there exists a close relation between the derived architecture and residual networks, and further investigate this connection in numerical experiments. Moreover, we demonstrate how our approach can be used to unroll optimization algorithms for certain problems with hard constraints. Using the example of speech dequantization, we show that our method can outperform classical splitting methods when both are applied to the same task.

Recently, a variety of approaches has been enriching the field of Remote Sensing (RS) image processing and analysis. Unfortunately, existing methods remain limited faced to the rich spatio-spectral content of today's large datasets. It would seem intriguing to resort to Deep Learning (DL) based approaches at this stage with regards to their ability to offer accurate semantic interpretation of the data. However, the specificity introduced by the coexistence of spectral and spatial content in the RS datasets widens the scope of the challenges presented to adapt DL methods to these contexts. Therefore, the aim of this paper is firstly to explore the performance of DL architectures for the RS hyperspectral dataset classification and secondly to introduce a new three-dimensional DL approach that enables a joint spectral and spatial information process. A set of three-dimensional schemes is proposed and evaluated. Experimental results based on well knownhyperspectral datasets demonstrate that the proposed method is able to achieve a better classification rate than state of the art methods with lower computational costs.

In recent years machine learning methods that nearly interpolate the data have achieved remarkable success. In many settings achieving near-zero training error leads to excellent test results. In this work we show how the mathematical and conceptual simplicity of interpolation can be harnessed to construct a framework for very efficient, scalable and accurate kernel machines. Our main innovation is in constructing kernel machines that output solutions mathematically equivalent to those obtained using standard kernels, yet capable of fully utilizing the available computing power of a parallel computational resource, such as GPU. Such utilization is key to strong performance since much of the computational resource capability is wasted by the standard iterative methods. The computational resource and data adaptivity of our learned kernels is based on theoretical convergence bounds. The resulting algorithm, which we call EigenPro 2.0, is accurate, principled and very fast. For example, using a single GPU, training on ImageNet with $1.3\times 10^6$ data points and $1000$ labels takes under an hour, while smaller datasets, such as MNIST, take seconds. Moreover, as the parameters are chosen analytically, based on the theory, little tuning beyond selecting the kernel and kernel parameter is needed, further facilitating the practical use of these methods.

Learning representations of multimodal data is a fundamentally complex research problem due to the presence of multiple sources of information. To address the complexities of multimodal data, we argue that suitable representation learning models should: 1) factorize representations according to independent factors of variation in the data, capture important features for both 2) discriminative and 3) generative tasks, and 4) couple both modality-specific and multimodal information. To encapsulate all these properties, we propose the Multimodal Factorization Model (MFM) that factorizes multimodal representations into two sets of independent factors: multimodal discriminative factors and modality-specific generative factors. Multimodal discriminative factors are shared across all modalities and contain joint multimodal features required for discriminative tasks such as predicting sentiment. Modality-specific generative factors are unique for each modality and contain the information required for generating data. Our experimental results show that our model is able to learn meaningful multimodal representations and achieve state-of-the-art or competitive performance on five multimodal datasets. Our model also demonstrates flexible generative capabilities by conditioning on the independent factors. We further interpret our factorized representations to understand the interactions that influence multimodal learning.

Sequential learning studies the problem of learning tasks in a sequence with restricted access to only the data of the current task. In the setting with a fixed model capacity, the learning process should not be selfish and account for later tasks to be added and therefore aim at utilizing a minimum number of neurons, leaving enough capacity for future needs. We explore different regularization strategies and activation functions that could lead to less interference between the different tasks. We show that learning a sparse representation is more beneficial for sequential learning than encouraging parameter sparsity regardless of their corresponding neurons. We particularly propose a novel regularizer that encourages representation sparsity by means of neural inhibition. It results in few active neurons which in turn leaves more free neurons to be utilized by upcoming tasks. We combine our regularizer with state-of-the-art lifelong learning methods that penalize changes on important previously learned parts of the network. We show that increased sparsity translates in a performance improvement on the different tasks that are learned in a sequence.

Due to recent advances - compute, data, models - the role of learning in autonomous systems has expanded significantly, rendering new applications possible for the first time. While some of the most significant benefits are obtained in the perception modules of the software stack, other aspects continue to rely on known manual procedures based on prior knowledge on geometry, dynamics, kinematics etc. Nonetheless, learning gains relevance in these modules when data collection and curation become easier than manual rule design. Building on this coarse and broad survey of current research, the final sections aim to provide insights into future potentials and challenges as well as the necessity of structure in current practical applications.

Adversarial attacks against Neural Networks are a problem of considerable importance, for which effective defenses are not yet readily available. We make progress toward this problem by showing that non-negative weight constraints can be used to improve resistance in specific scenarios. In particular, we show that they can provide an effective defense for binary classification problems with asymmetric cost, such as malware or spam detection. We also show how non-negativity can be leveraged to reduce an attacker's ability to perform targeted misclassification attacks in other domains such as image processing.

Close human-robot cooperation is a key enabler for new developments in advanced manufacturing and assistive applications. Close cooperation require robots that can predict human actions and intent, and understand human non-verbal cues. Recent approaches based on neural networks have led to encouraging results in the human action prediction problem both in continuous and discrete spaces. Our approach extends the research in this direction. Our contributions are three-fold. First, we validate the use of gaze and body pose cues as a means of predicting human action through a feature selection method. Next, we address two shortcomings of existing literature: predicting multiple and variable-length action sequences. This is achieved by introducing an encoder-decoder recurrent neural network topology in the discrete action prediction problem. In addition, we theoretically demonstrate the importance of predicting multiple action sequences as a means of estimating the stochastic reward in a human robot cooperation scenario. Finally, we show the ability to effectively train the prediction model on a action prediction dataset, involving human motion data, and explore the influence of the model's parameters on its performance. Source code repository: https://github.com/pschydlo/ActionAnticipation