Multilayered artificial neural networks are becoming a pervasive tool in a host of application fields. At the heart of this deep learning revolution are familiar concepts from applied and computational mathematics; notably, in calculus, approximation theory, optimization and linear algebra. This article provides a very brief introduction to the basic ideas that underlie deep learning from an applied mathematics perspective. Our target audience includes postgraduate and final year undergraduate students in mathematics who are keen to learn about the area. The article may also be useful for instructors in mathematics who wish to enliven their classes with references to the application of deep learning techniques. We focus on three fundamental questions: what is a deep neural network? how is a network trained? what is the stochastic gradient method? We illustrate the ideas with a short MATLAB code that sets up and trains a network. We also show the use of state-of-the art software on a large scale image classification problem. We finish with references to the current literature.

We formulate a supervised learning problem, referred to as continuous ranking, where a continuous real-valued label Y is assigned to an observable r.v. X taking its values in a feature space $\mathcal{X}$ and the goal is to order all possible observations x in $\mathcal{X}$ by means of a scoring function $s:\mathcal{X}\rightarrow \mathbb{R}$ so that s(X) and Y tend to increase or decrease together with highest probability. This problem generalizes bi/multi-partite ranking to a certain extent and the task of finding optimal scoring functions s(x) can be naturally cast as optimization of a dedicated functional criterion, called the IROC curve here, or as maximization of the Kendall ${\tau}$ related to the pair (s(X), Y ). From the theoretical side, we describe the optimal elements of this problem and provide statistical guarantees for empirical Kendall ${\tau}$ maximization under appropriate conditions for the class of scoring function candidates. We also propose a recursive statistical learning algorithm tailored to empirical IROC curve optimization and producing a piecewise constant scoring function that is fully described by an oriented binary tree. Preliminary numerical experiments highlight the difference in nature between regression and continuous ranking and provide strong empirical evidence of the performance of empirical optimizers of the criteria proposed.

This is the Udacity's Self-Driving Car Engineer Nanodegree Program final project for the 1st Term. To write a software pipeline to identify vehicles in a video from a front-facing camera on a car. In my implementation, I used a Deep Learning approach to image recognition. Specifically, I leveraged the extraordinary power of Convolutional Neural Networks (CNNs) to recognize images. However, the task at hand is not just to detect a vehicle's presence, but rather to point to its location. It turns out CNNs are suitable for these type of problems as well.

This is part two of my deep learning solution for lane detection, which covers the actual models I created in finding my final approach to the problem, as well as some potential improvements. Be sure to read Part One for the limitations of my previous approaches as well as the preliminary data used prior to the changes I made below. The code and data mentioned here and in the earlier post can be found in my Github repo. With a decent dataset created, I was ready to make my first model for using deep learning to detect lane lines. You may be asking, "Wait, I thought you were trying to get rid of perspective transformation?"

Artificial intelligence (AI) software developed by Alibaba Group has performed better than humans in a global reading comprehension test, the first time that machines have outperformed people. The AI research arm of China's biggest online commerce company developed a machine-learning model that scored higher on the Stanford Question Answering Dataset, a large-scale reading comprehension test with more than 100,000 questions, according to a release by the company. On January 11, Alibaba's machine-learning models scored 82.44 on the test, compared with 82.304 by humans. While computers have beaten humans at complex games like chess, where raw computing power and an infallible memory have given bots an advantage, languages are generally seen as harder for machines to master. AlphaGo's China showdown: why it's time to embrace artificial intelligence The win has broader implications for how companies deploy machine learning to replace customer service jobs that have so far relied on armies of call-centre employees to handle inquiries.

A Mission teacher wants to create a series of tutorial videos to teach people about machine learning, that is, if he can get his idea off the ground and running. Jeremy Ellis is a teacher at Mission Secondary School and he thinks machine learning is a valuable tool for almost everyone. But what is machine learning? Some websites define it as "a field of computer science that gives computers the ability to learn without being explicitly programmed," but Ellis has a simpler example. "Machine learning is what Google uses to figure out what ads to give you when your Google searching. So, it kind of knows a little bit of your history and extrapolates to something you might be interested in ."

The course is broad and pretty decent introductory course, but there is a number of presentation and course design flaws. First, while I'm not sure whether it is solely a Coursera's typical marketing approach to prevent users from refusing the course just because of the minimum amount of time required, or authors' unintended misestimations, but the actual time needed to complete the course is a way more than listed at the course home page, especially assignments. Often the time needed only to run an assignment training with no coding exceeds the given estimate. To get the value from the course one should be prepared to allocate much more time (2x-3x in total). Second, the course is too broad to be called an introductory one but too shallow in terms of math/practical/reasoning details to be named a deep one.

Think back to those pesky reading comprehension tests you used to take in school. Trust me, they are still on the SAT and still challenging. From the rapidly developing artificial intelligence perspective, we just got poned. Not one, but already two, artificial intelligence natural language processing algorithms outscored the average person's score on a rigorous reading comprehension test designed by Stanford researchers. The first to claim the honor of "better than people" was Alibaba's artificial intelligence.

Zhang Yong, head of Chinese tech giant Alibaba, introduces the company's artificial intelligence ET Brain at a conference in December 2017.Credit: Li Xin/Xinhua via ZUMA A mountainous district in western Beijing known for its temples and mushroom production is tipped to become China's hub for industries based on artificial intelligence (AI). Last week, the Chinese government announced that it will spend 13.8 billion yuan (US$2.1 billion) on an AI industrial park -- the first major investment in its plan to become a world leader in the field by 2030. But scientists there wonder whether the proposed 55-hectare AI park, in the Mentougou district 30 kilometres away from the city centre, will be able to attract enough researchers. The government wants it to house 400 companies that will make an estimated 50 billion yuan a year developing products and services in cloud computing, big data, biorecognition and deep learning. "I don't see any top talent willing to go to work and live there," says a scientist working at an AI start-up in Beijing, who asked to remain anonymous because the government is sensitive to criticism.