Microsoft (NASDAQ:MSFT) has delivered a rapid transformation under CEO Satya Nadella. After his first few weeks as CEO, Nadella described a Mobile First, Cloud First vision on which strategic and organizational decisions would be evaluated. He has stuck to that vision ever since and the results are dramatic. No longer is "mobile" defined by a device as was the case previously. Mobile now means mobile or computing on whatever platform a consumer chooses, wherever she wants. Cloud doesn't really mean just cloud.

From Alexa to Amazon Go, we use deep learning extensively across all areas of Amazon, and we've tried a lot of deep learning engines along the way. One has emerged as the most scalable, efficient way to perform deep learning, and for these reasons, we have selected MXNet as our engine of choice at Amazon. MXNet is an open source, state of the art deep learning engine, which allows developers to build sophisticated, custom artificial intelligence systems. Training these systems is significantly faster in MXNet, due to its scale and performance. For example, for the popular image recognition network, Resnet, MXNet has 2X the throughput compared to other engines, letting you train equivalent models in half the time.

As a practical example illustrating the shift from predictive to adaptive intelligent things, take the piece of machinery that is used to put the hubcaps on cars in an assembly line. When connected sensors are installed into this machine, it is given the ability to become predictive. These sensors can give insights into when the tool might break down and alert the car company, ultimately optimizing uptime. By increasing uptime the business is saving money. However, when machine learning is embedded into the device, it can analyze a plethora of data to eventually give the company that same desired output of saving money.

Peruse the headlines, and you'll find hundreds of articles predicting the "next Silicon Valley." Pundits claim that the growing cost of living in the Bay Area is driving businesses away to more affordable regions that are rich in tech talent such as Austin, Phoenix, Boulder, and Miami. But try as they might, these "hubs" won't ever beat Northern California at its own game – the area will continue to dominate information technology because of its unique arbitrage of thought, culture and research. But that's okay, because striving to replicate the success of the Bay Area limits us by imposing an arbitrary constraint on our imagination. After all, does Silicon Valley represent the pinnacle of success in human innovation, or are there regions that have the potential to evolve into something even greater?

The transport ministry said Monday it will introduce safety standards for self-driving vehicles in Japan as early as this fall, including an alarm system that sounds 15 seconds after a driver takes his hands off the steering wheel while traveling on a highway. The introduction of the integrated standards is expected to spur the development of self-driving vehicles by Japanese automakers as well as information technology companies as they will make clear the technology necessary for such cars. The safety standards are in line with an agreement reached Friday by a U.N. working party tasked with creating a uniform system of regulations for vehicle design to facilitate international trade. Japanese automakers will be able to sell vehicles that pass domestic safety tests based on the new standards in the European market in the future as the same standards are expected to be introduced there. The Ministry of Land, Infrastructure, Transport and Tourism will revise relevant ministerial ordinances under the Road Traffic Act in line with the new regulations.

I mentioned in a blog last year that we are at the dawn of a new age of artificial intelligence (A.I.). And 2017 certainly is the beginning of a world that is rapidly embracing A.I. The halls at CES were filled with talking devices, many powered by the same presence, Alexa, Amazon's slowly evolving virtual assistant. There were several conversations about the impact the impending robot revolution would have on our lives, jobs and future occupations. IDC predicts that spending on A.I. will grow from $8 billion to $47 billion by 2020.

This example simulates a multi-label document classification problem. In the above process, rejection sampling is used to make sure that n is more than 2, and that the document length is never zero. Likewise, we reject classes which have already been chosen. The documents that are assigned to both classes are plotted surrounded by two colored circles. Note that PCA is used to perform an unsupervised dimensionality reduction, while CCA is used to perform a supervised one.

I tend to post related books with nearly every blog post. Here are some of the most popular books from all of the articles on this AI blog. If you know a great book I should feature, please get in touch.

The first WiNLP workshop will be co-located with ACL 2017 in Vancouver. The workshop aims at highlighting research done by women, providing a supportive venue for junior members, and offering opportunities for networking and career discussion.

The main focus here is on finding the point that minimizes the sum of the "distances" to n points in a d-dimensional space, called centroid or center, especially in the presence of outliers. We also discuss an interesting physics problem: finding the point of maximum or minimum light, sound, radioactivity, or heat intensity, in the presence of an energy field produced by n energy source points. Both problems are closely related and use the same algorithm to find solutions. The sum of "distances" between an arbitrary point (u, v) and a set S { (x(1), y(1)) ... (x(n), y(n)) } of n points is defined as follows: The function H has one parameter p called power, and when p 2, we are facing the traditional problem of finding the centroid of a cloud of points: in this case, the solution is the classic average of the n points. This solution is notoriously sensitive to outliers.