Gaussian Processes (GPs) with an appropriate kernel are known to provide accurate predictions and uncertainty estimates even with very small amounts of labeled data. However, GPs are generally unable to learn a good representation that can encode intricate structures in high dimensional data. The representation power of GPs depends heavily on kernel functions used to quantify the similarity between data points. Traditional GP kernels are not very effective at capturing similarity between high dimensional data points, while methods that use deep neural networks to learn a kernel are not sample-efficient. To overcome these drawbacks, we propose deep probabilistic kernels which use a probabilistic neural network to map high-dimensional data to a probability distribution in a low dimensional subspace, and leverage the rich work on kernels between distributions to capture the similarity between these distributions. Experiments on a variety of datasets show that building a GP using this covariance kernel solves the conflicting problems of representation learning and sample efficiency. Our model can be extended beyond GPs to other small-data paradigms such as few-shot classification where we show competitive performance with state-of-the-art models on the mini-Imagenet dataset.

Recent work in signal propagation theory has shown that dropout limits the depth to which information can propagate through a neural network. In this paper, we investigate the effect of initialisation on training speed and generalisation for ReLU networks within this depth limit. We ask the following research question: given that critical initialisation is crucial for training at large depth, if dropout limits the depth at which networks are trainable, does initialising critically still matter? We conduct a large-scale controlled experiment, and perform a statistical analysis of over $12000$ trained networks. We find that (1) trainable networks show no statistically significant difference in performance over a wide range of non-critical initialisations; (2) for initialisations that show a statistically significant difference, the net effect on performance is small; (3) only extreme initialisations (very small or very large) perform worse than criticality. These findings also apply to standard ReLU networks of moderate depth as a special case of zero dropout. Our results therefore suggest that, in the shallow-to-moderate depth setting, critical initialisation provides zero performance gains when compared to off-critical initialisations and that searching for off-critical initialisations that might improve training speed or generalisation, is likely to be a fruitless endeavour.

Deep Learning has made several breakthroughs in recent years. Compared to traditional computation platforms, it has become more sophisticated and advanced than ever. Smart homes, intelligent personal assistant, etc. are some of the major breakthroughs in the present era. In this article, we list down 8 platforms which can be used to build mobile deep learning solutions. Facebook's open-source deep learning framework, Caffe2 is a lightweight, modular, and scalable framework which provides an easy way to experiment with deep learning models and algorithms. The framework comes with native Python and C APIs that work interchangeably and integrates with Android Studio, Microsoft Visual Studio, or XCode for mobile development.

If you're a lover of coffee, it will come as unpleasant news that the price of coffee could potentially spike in the near future. Climate change and deforestation are threatening some of the biggest coffee species in the world, but AI could potentially help keep coffee relatively affordable. The combined forces of deforestation and climate change are threatening the production of many species of coffee, including the common Arabica species, which can be found in many of the most prolific blends and brews. Coffee farmers around the globe are having to deal with rising temperatures and the problems that are associated with them, such as periods of drought. One recent study published in the journals Global Change Biology and Science Advances found that there were substantial risks to many wild coffee species, with around 60% of 124 different wild coffee species being vulnerable to extinction.

Predicting discomfort glare in open-plan offices is a challenging problem since most of available glare metrics are developed for cellular offices which are typically daylight dominated. The problem with open-plan offices is that they are mainly dependent on electric lighting rather than daylight even when they have a fully glazed facade. In addition, the contrast between bright windows and the buildings interior can be problematic and may cause discomfort glare to the building occupants. These problems can affect occupant productivity and wellbeing. Thus, it is important to develop a predictive model to avoid discomfort glare when designing open plan offices. To the best of our knowledge, we are the first to adopt Machine Learning (ML) models to predict discomfort glare. In order to develop new glare predictive models for these types of offices, Post-Occupancy Evaluation (POE) and High Dynamic Range (HDR) images were collected from 80 occupants (n=80) in four different open-plan offices. Consequently, various multi-region luminance values, luminance and glare indices were calculated and used as input features to train ML models. The accuracy of the ML model was compared to the accuracy of 24 indices which were also evaluated using a Receiver Operating Characteristic (ROC) analysis to identify the best cutoff values (thresholds) for each index for open-plan configurations. Results showed that the ML glare model could predict glare in open-plan offices with an accuracy of 83.8% (0.80 true positive rate and 0.86 true negative rate) which outperformed the accuracy of the previously developed glare metrics.

Non-negative tensor factorization models enable predictive analysis on count data. Among them, Bayesian Poisson-Gamma models are able to derive full posterior distributions of latent factors and are less sensitive to sparse count data. However, current inference methods for these Bayesian models adopt restricted update rules for the posterior parameters. They also fail to share the update information to better cope with the data sparsity. Moreover, these models are not endowed with a component that handles the imbalance in count data values. In this paper, we propose a novel variational auto-encoder framework called VAE-BPTF which addresses the above issues. It uses multi-layer perceptron networks to encode and share complex update information. The encoded information is then reweighted per data instance to penalize common data values before aggregated to compute the posterior parameters for the latent factors. Under synthetic data evaluation, VAE-BPTF tended to recover the right number of latent factors and posterior parameter values. It also outperformed current models in both reconstruction errors and latent factor (semantic) coherence across five real-world datasets. Furthermore, the latent factors inferred by VAE-BPTF are perceived to be meaningful and coherent under a qualitative analysis.

Spectrum management and resource allocation (RA) problems are challenging and critical in a vast number of research areas such as wireless communications and computer networks. The traditional approaches for solving such problems usually consume time and memory, especially for large size problems. Recently different machine learning approaches have been considered as potential promising techniques for combinatorial optimization problems, especially the generative model of the deep neural networks. In this work, we propose a resource allocation deep autoencoder network, as one of the promising generative models, for enabling spectrum sharing in underlay device-to-device (D2D) communication by solving linear sum assignment problems (LSAPs). Specifically, we investigate the performance of three different architectures for the conditional variational autoencoders (CVAE). The three proposed architecture are the convolutional neural network (CVAE-CNN) autoencoder, the feed-forward neural network (CVAE-FNN) autoencoder, and the hybrid (H-CVAE) autoencoder. The simulation results show that the proposed approach could be used as a replacement of the conventional RA techniques, such as the Hungarian algorithm, due to its ability to find solutions of LASPs of different sizes with high accuracy and very fast execution time. Moreover, the simulation results reveal that the accuracy of the proposed hybrid autoencoder architecture outperforms the other proposed architectures and the state-of-the-art DNN techniques.

ACCRA - Artificial intelligence took center stage as African female technology experts met at Women in Tech Week in Ghana to promote women's involvement in the field. When Lily Edinam Botsyoe was studying computer science at a university in Ghana, students wrote programming codes on a whiteboard because there were not enough computers. This made it difficult to apply the coding skills they were learning, she says, and the problem continues today. "We have students coming out of schools having the theoretical background -- which is very important because you can't actually appreciate something practical if you don't have the theory. But, the industry-ready skills is lacking because they didn't have the hands-on experience," Botsyoe said.

Artificial intelligence (AI) is one of the key drivers of digital health. Digital health and AI applications in medicine and biology are emerging worldwide, not only in resource-rich but also resource-limited regions. AI predates to the mid-20th century, but the current wave of AI builds in part on machine learning (ML), big data, and algorithms that can learn from massive amounts of online user data from patients or healthy persons. There are lessons to be learned from AI applications in different medical specialties and across developed and resource-limited contexts. A case in point is congenital heart defects (CHDs) that continue to plague sub-Saharan Africa, which calls for innovative approaches to improve risk prediction and performance of the available diagnostics.

The first decade of the 21st century introduced us to sweeping mobile and social revolutions largely driven by names like Jobs, Zuckerberg and Bezos. In the second decade that's now closing, things got a little more… complicated. During those years, a new collection of faces have joined the earlier tech titans to continue moving us into the future. A person wears a Guy Fawkes mask, which today is a trademark and symbol for the online hacktivist group Anonymous. More a decentralized collective than a personality, Anonymous was the name claimed by the loose affiliation of hackers who brought "hacktivism" into the mainstream. During the first half of the decade, Anonymous launched attacks against targets like ISIS, the governments of the US and Tunisia, and corporations such as Sony and PayPal.