Speaker recognition systems based on deep speaker embeddings have achieved significant performance in controlled conditions according to the results obtained for early NIST SRE (Speaker Recognition Evaluation) datasets. From the practical point of view, taking into account the increased interest in virtual assistants (such as Amazon Alexa, Google Home, AppleSiri, etc.), speaker verification on short utterances in uncontrolled noisy environment conditions is one of the most challenging and highly demanded tasks. This paper presents approaches aimed to achieve two goals: a) improve the quality of far-field speaker verification systems in the presence of environmental noise, reverberation and b) reduce the system qualitydegradation for short utterances. For these purposes, we considered deep neural network architectures based on TDNN (TimeDelay Neural Network) and ResNet (Residual Neural Network) blocks. We experimented with state-of-the-art embedding extractors and their training procedures. Obtained results confirm that ResNet architectures outperform the standard x-vector approach in terms of speaker verification quality for both long-duration and short-duration utterances. We also investigate the impact of speech activity detector, different scoring models, adaptation and score normalization techniques. The experimental results are presented for publicly available data and verification protocols for the VoxCeleb1, VoxCeleb2, and VOiCES datasets.

Machine learning has shown growing success in recent years. However, current machine learning systems are highly specialized, trained for particular problems or domains, and typically on a single narrow dataset. Human learning, on the other hand, is highly general and adaptable. Never-ending learning is a machine learning paradigm that aims to bridge this gap, with the goal of encouraging researchers to design machine learning systems that can learn to perform a wider variety of inter-related tasks in more complex environments. To date, there is no environment or testbed to facilitate the development and evaluation of never-ending learning systems. To this end, we propose the Jelly Bean World testbed. The Jelly Bean World allows experimentation over two-dimensional grid worlds which are filled with items and in which agents can navigate. This testbed provides environments that are sufficiently complex and where more generally intelligent algorithms ought to perform better than current state-of-the-art reinforcement learning approaches. It does so by producing non-stationary environments and facilitating experimentation with multi-task, multi-agent, multi-modal, and curriculum learning settings. We hope that this new freely-available software will prompt new research and interest in the development and evaluation of never-ending learning systems and more broadly, general intelligence systems.

Recent advances in the field of language modeling have improved state-of-the-art results on many Natural Language Processing tasks. Among them, the Machine Reading Comprehension task has made significant progress. However, most of the results are reported in English since labeled resources available in other languages, such as French, remain scarce. In the present work, we introduce the French Question Answering Dataset (FQuAD). FQuAD is French Native Reading Comprehension dataset that consists of 25,000+ questions on a set of Wikipedia articles. A baseline model is trained which achieves an F1 score of 88.0% and an exact match ratio of 77.9% on the test set. The dataset is made freely available at https://fquad.illuin.tech.

Variable selection problems generally present more than a single solution and, sometimes, it is worth to find as many solutions as possible. The use of Evolutionary Algorithms applied to this kind of problem proves to be one of the best methods to find optimal solutions. Moreover, there are variants designed to find all or almost all local optima, known as Niching Genetic Algorithms (NGA). There are several different NGA methods developed in order to achieve this task. The present work compares the behavior of eight different niching techniques, applied to a climatic database of four weather stations distributed in Tucuman, Argentina. The goal is to find different sets of input variables that have been used as the input variable by the estimation method. Final results were evaluated based on low estimation error and low dispersion error, as well as a high number of different results and low computational time. A second experiment was carried out to study the capability of the method to identify critical variables. The best results were obtained with Deterministic Crowding. In contrast, Steady State Worst Among Most Similar and Probabilistic Crowding showed good results but longer processing times and less ability to determine the critical factors.

Last Tuesday, 14th January 2020, was a big day for the Norwegian IT sector as the government's national strategy for artificial intelligence was presented at a breakfast meeting at MESH, central Oslo. Over 160 people from business, academia and the public sector participated in the launch, as well as many who followed the event online. The presented strategy claims to serve as a framework for both public and private sectors that aim to develop and use artificial intelligence, especially in areas where Norway already is greatly positioned and has strong foundations, such as in health, oil and gas, energy, and marine industry. Regarding the current digital development, we see many countries have high ambitions where one worth mentioning is the UK. Their AI strategy was initiated in 2017, which has by 2019 opened 16 New Centres for Doctoral Training in AI at universities across the country, industry funding for new Masters positions and numerous governmental funded scholar-ships.

None of those are proven cures for the coronavirus, but this kind of misinformation has been spreading online, in some places seemingly faster than the disease itself. Internet giants like Facebook, Google, Twitter and TikTok have all pledged to promote fact-based information on the epidemic. And the World Health Organization has pledged to partner with technology firms to push out authoritative data. It won't be easy, experts say. If you were online when the virus broke, you may have seen that … bat video.

Measuring similarity between two objects is the core operation in existing cluster analyses in grouping similar objects into clusters. Cluster analyses have been applied to a number of applications, including image segmentation, social network analysis, and computational biology. This paper introduces a new similarity measure called point-set kernel which computes the similarity between an object and a sample of objects generated from an unknown distribution. The proposed clustering procedure utilizes this new measure to characterize both the typical point of every cluster and the cluster grown from the typical point. We show that the new clustering procedure is both effective and efficient such that it can deal with large scale datasets. In contrast, existing clustering algorithms are either efficient or effective; and even efficient ones have difficulty dealing with large scale datasets without special hardware. We show that the proposed algorithm is more effective and runs orders of magnitude faster than the state-of-the-art density-peak clustering and scalable kernel k-means clustering when applying to datasets of millions of data points, on commonly used computing machines.

Graph convolutional networks (GCNs) are a widely used method for graph representation learning. We investigate the power of GCNs, as a function of their number of layers, to distinguish between different random graph models on the basis of the embeddings of their sample graphs. In particular, the graph models that we consider arise from graphons, which are the most general possible parameterizations of infinite exchangeable graph models and which are the central objects of study in the theory of dense graph limits. We exhibit an infinite class of graphons that are well-separated in terms of cut distance and are indistinguishable by a GCN with nonlinear activation functions coming from a certain broad class if its depth is at least logarithmic in the size of the sample graph. These results theoretically match empirical observations of several prior works. Finally, we show a converse result that for pairs of graphons satisfying a degree profile separation property, a very simple GCN architecture suffices for distinguishability. To prove our results, we exploit a connection to random walks on graphs.

Large-scale distributed training of deep neural networks results in models with worse generalization performance as a result of the increase in the effective mini-batch size. Previous approaches attempt to address this problem by varying the learning rate and batch size over epochs and layers, or ad hoc modifications of batch normalization. We propose Scalable and Practical Natural Gradient Descent (SP-NGD), a principled approach for training models that allows them to attain similar generalization performance to models trained with first-order optimization methods, but with accelerated convergence. Furthermore, SP-NGD scales to large mini-batch sizes with a negligible computational overhead as compared to first-order methods. We evaluated SP-NGD on a benchmark task where highly optimized first-order methods are available as references: training a ResNet-50 model for image classification on ImageNet. We demonstrate convergence to a top-1 validation accuracy of 75.4% in 5.5 minutes using a mini-batch size of 32,768 with 1,024 GPUs, as well as an accuracy of 74.9% with an extremely large mini-batch size of 131,072 in 873 steps of SP-NGD.

Distances are pervasive in machine learning. They serve as similarity measures, loss functions, and learning targets; it is said that a good distance measure solves a task. When defining distances, the triangle inequality has proven to be a useful constraint, both theoretically--to prove convergence and optimality guarantees--and empirically--as an inductive bias. Deep metric learning architectures that respect the triangle inequality rely, almost exclusively, on Euclidean distance in the latent space. Though effective, this fails to model two broad classes of subadditive distances, common in graphs and reinforcement learning: asymmetric metrics, and metrics that cannot be embedded into Euclidean space. To address these problems, we introduce novel architectures that are guaranteed to satisfy the triangle inequality. We prove our architectures universally approximate norm-induced metrics on $\mathbb{R}^n$, and present a similar result for modified Input Convex Neural Networks. We show that our architectures outperform existing metric approaches when modeling graph distances and have a better inductive bias than non-metric approaches when training data is limited in the multi-goal reinforcement learning setting.