Let's face it; training a machine learning model is time-consuming. Even with the advancement in computing prowess over the past few years, training machine learning models takes a lot of time. Even the most trivial models have more than a million parameters. On a bigger scale, these models have over a billion parameters(GPT-3 has over 175 billion parameters!), and training these models takes days, if not weeks. As a Data Scientist, we would want to keep an eye on the model's metrics to know if the model performs as per expectations.

As of October 2020, the COVID-19 pandemic has claimed over 1 million lives across the world and over 41 million people have been infected. Understanding the factors and policies that influence the spread of the virus can help governments make informed decisions in order to control infections and deaths until a vaccine becomes widely available. The data used for this project can be divided into four different parts, each represented as separate data frames/tables in the code: policy data, mobility data, demographic data, and COVID-19 time-series statistics. The policy data, extracted from the OxCGRT dataset, contains information about the policies implemented by the government in each country to control the spread of COVID-19. The policy data is available for each day after the start of the pandemic.

Research conducted by Gartner suggests that artificial intelligence or AI will create a business value of US $3.9 trillion by 2022. What's more, artificial intelligence is expected to be the most disruptive technology category for the next decade, due to advances in computing power, capacity, speed, and data diversity, along with the further evolution of deep neural networks (DNN). This growth is fueling a demand for talent in a number of related disciplines, including that of artificial intelligence engineering. But what is artificial intelligence engineering? Before answering that question, it's worth stepping back a little, to look at the evolution of artificial intelligence itself, and how it is enabling new ways of doing things that new require new skill sets to implement.

How do you build a robust end-to-end AI model to automatically detect defects in pipes in an aircraft? That was the big challenge for Airbus, which used MATLAB to quickly prototype and develop deep learning models to meet their needs. Working with the MathWorks Consulting Services team, Airbus adopted MATLAB to address the three main steps in the process. The first step was to have an integrated tool to build and train deep learning models from scratch for approaches such as semantic segmentation, as well as an easy and interactive environment for labeling videos. The positions of the ventilation holes and the wires on the pipe, found by the deep learning model in MATLAB, were used to measure distances and angles required by industry standard.

TensorFlow is a well established, open source machine learning and deep learning framework that can be used to create and run a wide range of different models, usually using powerful machines in the cloud. In addition, TensorFlow also supports running models on mobile devices through the TensorFlow.Mobile library, taking advantage of the hardware acceleration available on modern phones to run models incredibly fast on low powered mobile devices. In this post, we'll discuss how to build an image classifier to identify different fruits using the Azure Custom Vision service, and create a simple Android app to use this model to classify images. Creating ML models can be time consuming and require large data sets. To make it easier to create image classification models, Microsoft has created the Custom Vision Service, which uses a technique called transfer learning to allow you to train an image classifier using only a small number of images, instead of the thousands that traditionally would be required to train such a model.

It's no secret that deep learning is amazing. For the past eight years, the deep learning revolution has taken hold of computing and made technologies we thought might never be possible a daily reality. How can regular machine learning engineers and developers take those cutting edge algorithms from the R&D labs and put them into production? Deep Learning Design Patterns is here to help. It presents deep learning models in a unique-but-familiar new way: as extendable design patterns you can easily plug-and-play into your software projects.

The AutoSpeech challenge calls for automated machine learning (AutoML) solutions to automate the process of applying machine learning to speech processing tasks. These tasks, which cover a large variety of domains, will be shown to the automated system in a random order. Each time when the tasks are switched, the information of the new task will be hinted with its corresponding training set. Thus, every submitted solution should contain an adaptation routine which adapts the system to the new task. Compared to the first edition, the 2020 edition includes advances of 1) more speech tasks, 2) noisier data in each task, 3) a modified evaluation metric. This paper outlines the challenge and describe the competition protocol, datasets, evaluation metric, starting kit, and baseline systems.

This paper revisits spectral graph convolutional neural networks (graph-CNNs) given in Defferrard (2016) and develops the Laplace-Beltrami CNN (LB-CNN) by replacing the graph Laplacian with the LB operator. We then define spectral filters via the LB operator on a graph. We explore the feasibility of Chebyshev, Laguerre, and Hermite polynomials to approximate LB-based spectral filters and define an update of the LB operator for pooling in the LBCNN. We employ the brain image data from Alzheimer's Disease Neuroimaging Initiative (ADNI) and demonstrate the use of the proposed LB-CNN. Based on the cortical thickness of the ADNI dataset, we showed that the LB-CNN didn't improve classification accuracy compared to the spectral graph-CNN. The three polynomials had a similar computational cost and showed comparable classification accuracy in the LB-CNN or spectral graph-CNN. Our findings suggest that even though the shapes of the three polynomials are different, deep learning architecture allows us to learn spectral filters such that the classification performance is not dependent on the type of the polynomials or the operators (graph Laplacian and LB operator).

Current autoencoder-based disentangled representation learning methods achieve disentanglement by penalizing the (aggregate) posterior to encourage statistical independence of the latent factors. This approach introduces a trade-off between disentangled representation learning and reconstruction quality since the model does not have enough capacity to learn correlated latent variables that capture detail information present in most image data. To overcome this trade-off, we present a novel multi-stage modelling approach where the disentangled factors are first learned using a preexisting disentangled representation learning method (such as $\beta$-TCVAE); then, the low-quality reconstruction is improved with another deep generative model that is trained to model the missing correlated latent variables, adding detail information while maintaining conditioning on the previously learned disentangled factors. Taken together, our multi-stage modelling approach results in a single, coherent probabilistic model that is theoretically justified by the principal of D-separation and can be realized with a variety of model classes including likelihood-based models such as variational autoencoders, implicit models such as generative adversarial networks, and tractable models like normalizing flows or mixtures of Gaussians. We demonstrate that our multi-stage model has much higher reconstruction quality than current state-of-the-art methods with equivalent disentanglement performance across multiple standard benchmarks.