A group of professors and researchers at the Technical University of Berlin, the University of Vienna, and ETH Zurich have recently been working on understanding deep neural networks (computer systems that are modelled after the human brain) in "a mathematically sound way", as Dr. Phillip Petersen refers to it. Although the official paper for this research, "Optimal Approximation with Sparse Deep Neural Networks," will not be published until next week, Professor Gitta Kutyniok graciously presented a preview of their work for the Center's Math & Data Seminar group this past Thursday. Neural networks, or artificial brains, represent functions in mathematics. For these researchers, the main goal is to uncover how well a deep neural network with sparse connectivity can approximate a function. Dr. Petersen likens the network to a tree -- deep neural networks are composed of multiple layers and are connected by edges. Those layers are made of nodes, or neurons where computation occurs, and are sparsely connected if they have few non-zero weights or edges.

What are the main differences between NLP for Chinese vs NLP for English? Linguistically speaking, Chinese is an isolating language different from English. There are no spaces between words in Chinese written texts, and Chinese grammatical relations are indicated by word order. These factors have multiplied the difficulty of Chinese disambiguation at lexical, syntactic and semantic levels, since modern linguistic concepts and principles are more suitable for English than for Chinese. Currently, most mainstream NLP methods are language-independent.

This is another installment of Mighty AI's "Conversations in Machine Learning" blog series. Each week, our content human, Cassie, shares a summary of a recent conversation we had with a machine learning team and potential customer--what they're building, how they're handling training data today, etc. Read more about the series here. This week I'm highlighting a call we had with the makers of a consumer mobile application that is killing it in terms of downloads, usage, and even business model. You've quite likely heard of the app, but due to my vow to not be creepy or shady, I won't name it. So first, do you remember a time when people were not snapping selfies, photo-documenting their brunch, or capturing touching moments with a freaking iPad that blocks everyone else's view?

"According to merriam-webster.com, the simple definition of artificial intelligence: 'an area of computer science that deals with giving machines the ability to seem like they have human intelligence; the power of a machine to copy intelligent behavior.'" Ready or not, artificial intelligence is here; and it's here to stay. Businesses and organizations that have taken an early lead in the adoption and use of artificial intelligence (be it natural language processing, machine learning, deep learning, or cognitive computing) are simply scratching the surface of its potential to not only improve sales, service, marketing and operations, but to discover and deliver new digital business models. In the 2017 Economist Intelligence Unit report, Artificial Intelligence in the Real World, 75% of more than 200 business executives surveyed said AI will be actively implemented in their companies within the next three years. And while many are wary of its potential to reduce human employment, 27% say introducing artificial intelligence to business will improve decision making; 26% believe it will improve customer service; 29% say it will improve operating efficiency; and 17% said it will increase sales revenue.

Move over, managers, there's a new boss in the office: artificial intelligence. The same technology that enables a navigation app to find the most efficient route to your destination or lets an online store recommend products based on past purchases is on the verge of transforming the office--promising to remake how we look for job candidates, get the most out of workers and keep our best workers on the job. These applications aim to analyze a vast amount of data and search for patterns--broadening managers' options and helping them systematize processes that are often driven simply by instinct. And just like shopping sites, the AIs are designed to learn from experience to get an ever-better idea of what managers want. A company can provide a job description, and AI will collect and crunch data from a variety of sources to find people with the right talents, with experience to match--candidates who might never have thought of applying to the company, and whom the company might never have thought of seeking out.

Few things have the transformative potential of artificial intelligence (AI). It has the power to completely change how we live, work and get around. If electricity and the internet changed humanity forever in the 20th century, then AI will in the 21st. It's predicted that AI will bring a massive shift in how people perceive and interact with technology, with machines performing a greater and greater number of tasks and, in many cases, doing a better job of it than humans. By its simplest definition, AI is an umbrella term for technologies that are inspired by biological systems that give computers human-like abilities related to seeing, reasoning, hearing, and learning.

In the consumer report, we are number one once again and just like the Q7, in the consumer report it also occupies the first position as the best luxury SUV. And I think this power of the brand makes it possible for us to grow significantly. There are couple of models which have not even be launched yet in this market, models which we already know here, for instance the S4, the A5, and the entirely new A5 Sportback. They are now being launched in the United States. All new models for this market, and I assume that this year once again we are going to experience very solid growth in the United States. And the question so whether we spend more money for this? I can tell you we even spend less money in form of sales discounts because of the powerful brand and the relatively young product portfolio. So you would take the second part?

Expanding on their Jetson TX1 and TK1 products for embedded computing, NVIDIA announced last week their Jetson TX2 platform--a hardware and software platform the size of a credit card designed to deliver AI computing at the edge. NVIDIA touts Jetson TX2 as delivering "unprecedented deep learning capabilities," and based on the form factor, they may be right as it paves the way for a number of cutting-edge uses--from highly intelligent factory robots and commercial drones, to cameras with AI for smart cities. NVIDIA has been running on all cylinders lately with datacenter machine learning, and I think this release, if it performs as promised, will solidify their place at the top of the machine learning class in certain classes of devices. NVIDIA announced the TX2 at an event I attended last week in San Francisco with many tier 1 vendors and startups with some interesting use cases. Jetson, by design, isn't targeted at every embedded device, it's for those non-mobile devices who need strong deep neural network performance at a given power draw. The TX2 is a significant step up from its predecessor.

A common applied statistics task involves building regression models to characterize non-linear relationships between variables. It is possible to fit such models by assuming a particular non-linear functional form, such as a sinusoidal, exponential, or polynomial function, to describe one variable's response to the variation in another. Unless this relationship is obvious from the outset, however, it involves possibly extensive model selection procedures to ensure the most appropriate model is retained. Alternatively, a non-parametric approach can be adopted by defining a set of knots across the variable space and use a spline or kernel regression to describe arbitrary non-linear relationships. However, knot layout procedures are somewhat ad hoc and can also involve variable selection. A third alternative is to adopt a Bayesian non-parametric strategy, and directly model the unknown underlying function.

Crime poses a particularly interesting data challenge -- it is both geospatial and temporal, and may be affected by many different types of variables -- weather, city infrastructure, population demographics, public events, and government policy. We will talk about how to merge, visualize, and model this type of complex data, using PostGIS, spatial mapping, time-series analyses, and machine learning, and ask what most predicts crime -- and what we can do to prevent it in the future. She is an award-winning researcher and instructor. Prior to joining Dataiku, she completed her Ph.D. in Cognitive Neuroscience at Massachusetts Institute of Technology and worked as a Postdoctoral Fellow at Harvard. Jorie currently resides in New York where she helps clients research, analyze, builds and deploy scalable data products.