Dr. Quoc Viet Le is a research scientist at Google Brain known for his path-breaking work on deep neural networks (DNN). He is especially famous for his Ph.D work in image processing under Andrew Ng, one of the pioneers of the DNN revolution. Le's and Ng's work demonstrated how computers could be used to learn complicated features and patterns in a way similar to how the mammalian brain learns. This revolutionized the interest in DNNs, and got the current giants of the computer industry such as Google, Facebook and Microsoft in a race to incorporate AI techniques into their software. DNNs perform effectively in tasks such as image processing, handwriting recognition and game-playing, and are being explored for solutions to other problems such as self-driving cars, robotics, medical diagnosis and environmental and social problems.

Last week, Microsoft inadvertently revealed the difficulty of creating moral robots. Chatbot Tay, designed to speak like a teenage girl, turned into a Nazi-loving racist after less than 24 hours on Twitter. "Repeat after me, Hitler did nothing wrong," she said, after interacting with various trolls. "Bush did 9/11 and Hitler would have done a better job than the monkey we have got now." Of course, Tay wasn't designed to be explicitly moral.

An international coalition of media outlets publishes what it says is an extensive investigation into the offshore financial dealings of the rich and famous. The EU agency that's responsible for processing asylum seekers has less than a tenth of the 2,300 officers that it needs to do the job. He's calling for the Ohio governor, who has won only his home state, to drop out of the GOP presidential race. U.S. Sen. Lindsey Graham, speaking in Cairo during a congressional tour of the region, says that Congress will continue to play a primary role in foreign policy "regardless of what Mr. Trump says or does." The accident just south of Philadelphia kills two Amtrak workers and sends more than 30 passengers to the hospital.

Access to data is one thing. Moving from data to insight is another entirely. Jane Zavalishina, CEO of the Yandex Data Factory wants to make it easier for businesses to gain real insight from their data. Built on the real time personalization and predictive analytics technology of Russia's largest Internet business Yandex- Yandex Data Factory helps clients improve their business through the exploitation of their own data by machine learning. Disproportionate growth of figures suggesting that 90% of the world's data was generated over the course of the past two years has led Zavalishina to be of the opinion that "we should not really care about data anymore," rather, we should see it as a thing of the past.

At a glance, machine learning and statistics seem to be very similar, but many people fail to stress the importance of the difference between these two disciplines. Machine learning and statistics share the same goals--they both focus on data modeling--but their methods are affected by their cultural differences. In order to empower collaboration and knowledge creation, it's very important to understand the fundamental underlying differences that reflect in the cultural profile of these two disciplines. To gain a deeper understanding of these differences, we need to take a step back and look at their historical roots. In 1946, the first computer system, ENIAC, was developed with the vision of reforming numerical computation using a machine (instead of manual numerical computation using pencil and paper).

This tutorial introduces the CMA Evolution Strategy (ES), where CMA stands for Covariance Matrix Adaptation. The CMA-ES is a stochastic, or randomized, method for real-parameter (continuous domain) optimization of non-linear, non-convex functions. We try to motivate and derive the algorithm from intuitive concepts and from requirements of non-linear, non-convex search in continuous domain.

Many retailers today employ inventory management systems based on Re-Order Point Policies, most of which rely on the assumption that all decreases in product inventory levels result from product sales. Unfortunately, it usually happens that small but random quantities of the product get lost, stolen or broken without record as time passes, e.g., as a consequence of shoplifting. This is usual for retailers handling large varieties of inexpensive products, e.g., grocery stores. In turn, over time these discrepancies lead to stock freezing problems (see Ref. [1]), i.e., situations where the system believes the stock is above the reorder point but the actual stock is at zero, and so no replenishments or sales occur. Motivated by these issues, we model the interaction between sales, losses, replenishments and inventory levels as a Dynamic Bayesian Network (DBN), where the inventory levels are unobserved (i.e., hidden) variables we wish to estimate. We present an Expectation-Maximization (EM) algorithm to estimate the parameters of the sale and loss distributions, which relies on solving a one-dimensional dynamic program for the E-step and on solving two separate one-dimensional nonlinear programs for the M-step.

We study nonconvex finite-sum problems and analyze stochastic variance reduced gradient (SVRG) methods for them. SVRG and related methods have recently surged into prominence for convex optimization given their edge over stochastic gradient descent (SGD); but their theoretical analysis almost exclusively assumes convexity. In contrast, we prove non-asymptotic rates of convergence (to stationary points) of SVRG for nonconvex optimization, and show that it is provably faster than SGD and gradient descent. We also analyze a subclass of nonconvex problems on which SVRG attains linear convergence to the global optimum. We extend our analysis to mini-batch variants of SVRG, showing (theoretical) linear speedup due to mini-batching in parallel settings.

Topic models (e.g., pLSA, LDA, SLDA) have been widely used for segmenting imagery. These models are confined to crisp segmentation. Yet, there are many images in which some regions cannot be assigned a crisp label (e.g., transition regions between a foggy sky and the ground or between sand and water at a beach). In these cases, a visual word is best represented with partial memberships across multiple topics. To address this, we present a partial membership latent Dirichlet allocation (PM-LDA) model and associated parameter estimation algorithms. Experimental results on two natural image datasets and one SONAR image dataset show that PM-LDA can produce both crisp and soft semantic image segmentations; a capability existing methods do not have.

Automatic Offline Handwritten Signature Verification has been researched over the last few decades from several perspectives, using insights from graphology, computer vision, signal processing, among others. In spite of the advancements on the field, building classifiers that can separate between genuine signatures and skilled forgeries (forgeries made targeting a particular signature) is still hard. We propose approaching the problem from a feature learning perspective. Our hypothesis is that, in the absence of a good model of the data generation process, it is better to learn the features from data, instead of using hand-crafted features that have no resemblance to the signature generation process. To this end, we use Deep Convolutional Neural Networks to learn features in a writer-independent format, and use this model to obtain a feature representation on another set of users, where we train writer-dependent classifiers. We tested our method in two datasets: GPDS-960 and Brazilian PUC-PR. Our experimental results show that the features learned in a subset of the users are discriminative for the other users, including across different datasets, reaching close to the state-of-the-art in the GPDS dataset, and improving the state-of-the-art in the Brazilian PUC-PR dataset.