No abstract

Voice controlled virtual assistants (VAs) are now available in smartphones, cars, and standalone devices in homes. In most cases, the user needs to first "wake-up" the VA by saying a particular word/phrase every time he or she wants the VA to do something. Eliminating the need for saying the wake-up word for every interaction could improve the user experience. This would require the VA to have the capability to detect the speech that is being directed at it and respond accordingly. In other words, the challenge is to distinguish between system-directed and non-system-directed speech utterances. In this paper, we present a number of neural network architectures for tackling this classification problem based on using only acoustic features. These architectures are based on using convolutional, recurrent and feed-forward layers. In addition, we investigate the use of an attention mechanism applied to the output of the convolutional and the recurrent layers. It is shown that incorporating the proposed attention mechanism into the models always leads to significant improvement in classification accuracy. The best model achieved equal error rates of 16.25% and 15.62% on two distinct realistic datasets.

Generative Adversarial Networks (GANs) can successfully approximate a probability distribution and produce realistic samples. However, open questions such as sufficient convergence conditions and mode collapse still persist. In this paper, we build on existing work in the area by proposing a novel framework for training the generator against an ensemble of discriminator networks, which can be seen as a one-student/multiple-teachers setting. We formalize this problem within the full-information adversarial bandit framework, where we evaluate the capability of an algorithm to select mixtures of discriminators for providing the generator with feedback during learning. To this end, we propose a reward function which reflects the progress made by the generator and dynamically update the mixture weights allocated to each discriminator. We also draw connections between our algorithm and stochastic optimization methods and then show that existing approaches using multiple discriminators in literature can be recovered from our framework. We argue that less expressive discriminators are smoother and have a general coarse grained view of the modes map, which enforces the generator to cover a wide portion of the data distribution support. On the other hand, highly expressive discriminators ensure samples quality. Finally, experimental results show that our approach improves samples quality and diversity over existing baselines by effectively learning a curriculum. These results also support the claim that weaker discriminators have higher entropy improving modes coverage.

Spatial bias continues to be a major challenge in high-throughput screening technologies. Its successful detection and elimination are critical for identifying the most promising drug candidates. Here, we examine experimental small molecule assays from the popular ChemBank database and show that screening data are widely affected by both assay-specific and plate-specific spatial biases. Importantly, the bias affecting screening data can fit an additive or multiplicative model. We show that the use of appropriate statistical methods is essential for improving the quality of experimental screening data. The presented methodology can be recommended for the analysis of current and next-generation screening data.

Recent years have seen a steep rise in the number of skin cancer detection applications. While modern advances in deep learning made possible reaching new heights in terms of classification accuracy, no publicly available skin cancer detection software provide confidence estimates for these predictions. We present DUNEScan (Deep Uncertainty Estimation for Skin Cancer), a web server that performs an intuitive in-depth analysis of uncertainty in commonly used skin cancer classification models based on convolutional neural networks (CNNs). DUNEScan allows users to upload a skin lesion image, and quickly compares the mean and the variance estimates provided by a number of new and traditional CNN models. Moreover, our web server uses the Grad-CAM and UMAP algorithms to visualize the classification manifold for the user’s input, hence providing crucial information about its closeness to skin lesion images  from the popular ISIC database. DUNEScan is freely available at: https://www.dunescan.org.