Our goal is to design architectures that retain the groundbreaking performance of Convolutional Neural Networks (CNNs) for landmark localization and at the same time are lightweight, compact and suitable for applications with limited computational resources. To this end, we make the following contributions: (a) we are the first to study the effect of neural network binarization on localization tasks, namely human pose estimation and face alignment. We exhaustively evaluate various design choices, identify performance bottlenecks, and more importantly propose multiple orthogonal ways to boost performance. (b) Based on our analysis, we propose a novel hierarchical, parallel and multi-scale residual architecture that yields large performance improvement over the standard bottleneck block while having the same number of parameters, thus bridging the gap between the original network and its binarized counterpart. (c) We perform a large number of ablation studies that shed light on the properties and the performance of the proposed block. (d) We present results for experiments on the most challenging datasets for human pose estimation and face alignment, reporting in many cases state-of-the-art performance. (e) We further provide additional results for the problem of facial part segmentation. Code can be downloaded from https://www.adrianbulat.com/binary-cnn-landmark

Spurred on, no doubt, by the current spate of books and articles about AIED (artificial intelligence in education), the IATEFL Learning Technologies SIG is organising an online event on the topic in November of this year. Currently, the most visible online references to AI in language learning are related to Glossika, basically a language learning system that uses spaced repetition, whose marketing department has realised that references to AI might help sell the product. They're not alone – see, for example, Knowble which I reviewed earlier this year . In the wider world of education, where AI has made greater inroads than in language teaching, every day brings more stuff: How artificial intelligence is changing teaching, 32 Ways AI is Improving Education, How artificial intelligence could help teachers do a better job, etc., etc. Common to all these publications is the claim that AI will radically change education. When it comes to language teaching, a similar claim has been made by Donald Clark (described by Anthony Seldon as an education guru but perhaps best-known to many in ELT for his demolition of Sugata Mitra).

The research comes from the University of North Carolina at Chapel Hill, and it demonstrates how an artificial-intelligence design can teach itself how to design new drug molecules from scratch. Such a system could accelerate the design of new drug candidates for use across pharmaceuticals and healthcare. The new device is named "Reinforcement Learning for Structural Evolution" (abbreviated to ReLeaSE). The artificial intelligence is in the form of an algorithm which has been configured to work with a computer program, based on two neural networks. The networks are described by the researchers as being akin to a teacher and a student.

Irvine Calif.-based Syntiant thinks it can use embedded flash memory to greatly reduce the amount of power needed to perform deep-learning computations. Austin, Tex.-based Mythic thinks it can use embedded flash memory to greatly reduce the amount of power needed to perform deep-learning computations. They both might be right. A growing crowd of companies is hoping to deliver chips that accelerate otherwise onerous deep learning applications, and to some degree they all have similarities because "these are solutions that are created by the shape of the problem," explains Mythic founder and CTO Dave Fick. When executed in a CPU, that problem is shaped like a traffic jam of data. A neural network is made up of connections and "weights" that denote how strong those connections are, and having to move those weights around so they can be represented digitally in the right place and time is the major energy expenditure in doing deep learning today.

Whether it's business, IT companies, financial services or even education, Artificial Intelligence (AI) is being integrated into various industries. AI's digital, dynamic nature also offers opportunities for student engagement that cannot be found in often outdated documents or in the fixed environment. In a synergistic fashion, AI has the potential to propel and accelerate the discovery of new learning frontiers and the creation of innovative technologies. Though yet to become a standard cult in organizations and schools, Artificial Intelligence in learning or training has been a "big thing" since AI's uptick in the 1940s (when the first seeds of AI were sown with programmable computers). In many ways, the 2 seem made for each other.

A small team of student AI (artificial intelligence) coders outperformed codes from Google's researchers, reveal an important benchmark. Students from Fast.ai, a non-profit group that creates learning resources and is dedicated to making deep learning "accessible to all", have created an AI algorithm that beats code from Google's researchers. Researchers from Stanford measured the algorithm using a benchmark called DAWNBench that uses a common image classification task to track the speed of a deep-learning algorithm per dollar of compute power. According to the benchmark, the researchers found that the algorithm built by Fast.ai's team had beaten Google's code. Fast.ai consists of part-time students who are eager to try out machine learning and convert it into a career in data science.

The rise of the machines is coming. By that, I don't mean that someone is about to break through a singularity and create a rapidly self-teaching AI that will enslave all of humanity. I mean, that's probably on the cards too, but it's not really what we're talking about today. We're talking about machine learning: Smart machines performing roles traditionally held by human beings. They're already used today in medicine, robotics, remote sensors, and even in ATMs.

Neural Networks are a class of models within the general machine learning literature. So for example, if you took a Coursera course on machine learning, neural networks will likely be covered. Neural networks are a specific set of algorithms that has revolutionized the field of machine learning. They are inspired by biological neural networks and the current so called deep neural networks have proven to work quite very well. Neural Networks are themselves general function approximations, that is why they can be applied to literally almost any machine learning problem where the problem is about learning a complex mapping from the input to the output space. After finishing the famous Andrew Ng's Machine Learning Coursera course, I started developing interest towards neural networks and deep learning.

A basic grounding in the principles and practices around artificial intelligence (AI), automation and cognitive systems is something which is likely to become increasingly valuable, regardless of your field of business, expertise or profession. Fortunately, today you don't have to take years out of your life studying at university to become familiar with this seemingly hugely complex technology. A growing number of online courses have sprung up in recent years covering everything from the basics to advanced implementation. Some are aimed at people who want to dive straight into coding their own artificial neural networks, and understandably assume a certain level of technical ability. Others are useful for those who want to learn how this technology can be applied by anyone, regardless of prior technical expertise, to solving real-word problems.

When we think about job displacement our minds often go to factory workers and calls centers, but as leading artificial Intelligence expert Dr. Vivienne Ming points out in a recent interview with the Financial Times, the middle class in professional services may be the most challenged in this computed future. Ming cites a recent competition at Columbia University between human lawyers and AI counterparts reviewing agreements with loopholes. The AI found 95 per cent of them in 22 seconds, it took the humans over an hour. As a lawyer reading this you may have two conflicting emotions: concern, but the other should be joy that your life is about to get easier; free of the day to day drudgery of reviewing agreements, allowing you to concentrate on what really matters. This is what technology does – an advantage to highly complementary skills.