We investigate the performance of features that can capture nonlinear recurrence dynamics embedded in the speech signal for the task of Speech Emotion Recognition (SER). Reconstruction of the phase space of each speech frame and the computation of its respective Recurrence Plot (RP) reveals complex structures which can be measured by performing Recurrence Quantification Analysis (RQA). These measures are aggregated by using statistical functionals over segment and utterance periods. We report SER results for the proposed feature set on three databases using different classification methods. When fusing the proposed features with traditional feature sets, e.g., [1], we show an improvement in unweighted accuracy of up to 5.7% and 10.7% on Speaker-Dependent (SD) and Speaker-Independent (SI) SER tasks, respectively, over the baseline [1]. Following a segment-based approach we demonstrate state-of- the-art performance on IEMOCAP using a Bidirectional Recurrent Neural Network.

We study inverse problems consisting on determining medium properties using the responses to probing waves from the machine learning point of view. Based on the understanding of propagation of waves and their nonlinear interactions, we construct a deep convolutional neural network in which the parameters are used to classify and reconstruct the coefficients of nonlinear wave equations that model the medium properties. Furthermore, for given approximation accuracy, we obtain the depth and number of units of the network and their quantitative dependence on the complexity of the medium.

In image classification tasks, the ability of deep CNNs to deal with complex image data has proven to be unrivalled. However, they require large amounts of labeled training data to reach their full potential. In specialised domains such as healthcare, labeled data can be difficult and expensive to obtain. Active Learning aims to alleviate this problem, by reducing the amount of labelled data needed for a specific task while delivering satisfactory performance. We propose DEBAL, a new active learning strategy designed for deep neural networks. This method improves upon the current state-of-the-art deep Bayesian active learning method, which suffers from the mode collapse problem. We correct for this deficiency by making use of the expressive power and statistical properties of model ensembles. Our proposed method manages to capture superior data uncertainty, which translates into improved classification performance. We demonstrate empirically that our ensemble method yields faster convergence of CNNs trained on the MNIST and CIFAR-10 datasets.

A recent technical breakthrough in the domain of machine learning is the discovery and the multiple applications of Generative Adversarial Networks (GANs). Those generative models are computationally demanding, as a GAN is composed of two deep neural networks, and because it trains on large datasets. A GAN is generally trained on a single server. In this paper, we address the problem of distributing GANs so that they are able to train over datasets that are spread on multiple workers. MD-GAN is exposed as the first solution for this problem: we propose a novel learning procedure for GANs so that they fit this distributed setup. We then compare the performance of MD-GAN to an adapted version of Federated Learning to GANs, using the MNIST and CIFAR10 datasets. MD-GAN exhibits a reduction by a factor of two of the learning complexity on each worker node, while providing better performances than federated learning on both datasets. We finally discuss the practical implications of distributing GANs.

Recently deep neural networks have been successfully used for various classification tasks, especially for problems with massive perfectly labeled training data. However, it is often costly to have large-scale credible labels in real-world applications. One solution is to make supervised learning robust with imperfectly labeled input. In this paper, we develop a distribution correction approach that allows deep neural networks to avoid overfitting imperfect training data. Specifically, we treat the noisy input as samples from an incorrect distribution, which will be automatically corrected during our training process. We test our approach on several classification datasets with elaborately generated noisy labels. The results show significantly higher prediction and recovery accuracy with our approach compared to alternative methods.

Despite the huge success of deep learning, our understanding to how the non-convex neural networks are trained remains rather limited. Most of existing theoretical works only tackle neural networks with one hidden layer, and little is known for multi-layer neural networks. Recurrent neural networks (RNNs) are special multi-layer networks extensively used in natural language processing applications. They are particularly hard to analyze, comparing to feedforward networks, because the weight parameters are reused across the entire time horizon. We provide arguably the first theoretical understanding to the convergence speed of training RNNs (when activation functions are present). Specifically, when the weights are randomly initialized, and when the number of neurons is sufficiently large ---meaning polynomial in the training data size and the time horizon--- we show that gradient descent and stochastic gradient descent both minimize the regression loss in a linear convergence rate, that is, $\varepsilon \propto e^{-\Omega(T)}$.

The success of deep learning in recent years has raised concerns about adversarial examples, which allow attackers to force deep neural networks to output a specified target. Although a method by which to generate audio adversarial examples targeting a state-of-the-art speech recognition model has been proposed, this method cannot fool the model in the case of playing over the air, and thus, the threat was considered to be limited. In this paper, we propose a method to generate adversarial examples that can attack even when playing over the air in the physical world by simulating transformation caused by playback or recording and incorporating them in the generation process. Evaluation and a listening experiment demonstrated that audio adversarial examples generated by the proposed method may become a real threat.

Many people are interested in taking astonishing photos and sharing with others. Emerging hightech hardware and software facilitate ubiquitousness and functionality of digital photography. Because composition matters in photography, researchers have leveraged some common composition techniques to assess the aesthetic quality of photos computationally. However, composition techniques developed by professionals are far more diverse than well-documented techniques can cover. We leverage the vast underexplored innovations in photography for computational composition assistance. We propose a comprehensive framework, named CAPTAIN (Composition Assistance for Photo Taking), containing integrated deep-learned semantic detectors, sub-genre categorization, artistic pose clustering, personalized aesthetics-based image retrieval, and style set matching. The framework is backed by a large dataset crawled from a photo-sharing Website with mostly photography enthusiasts and professionals. The work proposes a sequence of steps that have not been explored in the past by researchers. The work addresses personal preferences for composition through presenting a ranked-list of photographs to the user based on user-specified weights in the similarity measure. The matching algorithm recognizes the best shot among a sequence of shots with respect to the user's preferred style set. We have conducted a number of experiments on the newly proposed components and reported findings. A user study demonstrates that the work is useful to those taking photos.

Executing natural language navigation instructions from raw observations requires solving language, perception, planning, and control problems. Consider instructing a quadcopter drone using natural language. Figure 1 shows an example instruction. Resolving the instruction requires identifying the blue fence, anvil and tree in the world, understanding the spatial constraints towards and on the right, planning a trajectory that satisfies these constraints, and continuously controlling the quadcopter to follow the trajectory. Existing work has addressed this problem mostly using manually-designed symbolic representations for language meaning and environment [1, 2, 3, 4, 5, 6].

Semantic Web knowledge representation standards, and in particular RDF and OWL, often come endowed with a formal semantics which is considered to be of fundamental importance for the field. Reasoning, i.e., the drawing of logical inferences from knowledge expressed in such standards, is traditionally based on logical deductive methods and algorithms which can be proven to be sound and complete and terminating, i.e. correct in a very strong sense. For various reasons, though, in particular, the scalability issues arising from the ever-increasing amounts of Semantic Web data available and the inability of deductive algorithms to deal with noise in the data, it has been argued that alternative means of reasoning should be investigated which bear high promise for high scalability and better robustness. From this perspective, deductive algorithms can be considered the gold standard regarding correctness against which alternative methods need to be tested. In this paper, we show that it is possible to train a Deep Learning system on RDF knowledge graphs, such that it is able to perform reasoning over new RDF knowledge graphs, with high precision and recall compared to the deductive gold standard.