Fei-Fei Li is a big deal in the world of AI. As the director of the Artificial Intelligence and Vision labs at Stanford University, she oversaw the creation of ImageNet, a vast database of images designed to accelerate the development of AI that can "see." And, well, it worked, helping to drive the creation of deep learning systems that can recognize objects, animals, people, and even entire scenes in photos--technology that has become commonplace on the world's biggest photo-sharing sites. Now, Fei-Fei will help run a brand new AI group inside Google, a move that reflects just how aggressively the world's biggest tech companies are remaking themselves around this breed of artificial intelligence. Intel Looks to a New Chip to Power the Coming Age of AI Giant Corporations Are Hoarding the World's AI Talent OpenAI Joins Microsoft on the Cloud's Next Big Front: Chips Facebook Manages to Squeeze an AI Into Its Mobile App Giant Corporations Are Hoarding the World's AI Talent Giant Corporations Are Hoarding the World's AI Talent Alongside a former Stanford researcher--Jia Li, who more recently ran research for the social networking service Snapchat--the China-born Fei-Fei will lead a team inside Google's cloud computing operation, building online services that any coder or company can use to build their own AI. This new Cloud Machine Learning Group is the latest example of AI not only re-shaping the technology that Google uses, but also changing how the company organizes and operates its business.

Recent advances in machine learning have led to innovative applications and services that use computational structures to reason about complex phenomenon. Over the past several years, the security and machine-learning communities have developed novel techniques for constructing adversarial samples--malicious inputs crafted to mislead (and therefore corrupt the integrity of) systems built on computationally learned models. The authors consider the underlying causes of adversarial samples and the future countermeasures that might mitigate them.

Cognitive computing is the simulation of human thought processes in a computerized model. Cognitive computing involves self-learning systems that use data mining, pattern recognition and natural language processing to mimic the way the human brain works. The goal of cognitive computing is to create automated IT systems that are capable of solving problems without requiring human assistance. Cognitive computing systems use machine learning algorithms. Such systems continually acquire knowledge from the data fed into them by mining data for information.

DeepMind, Google's artificial intelligence outfit, wants to streamline health care by using machine learning to provide medics with intelligent notifications. But not everyone is happy with the piles of data being shared with the company. The project will provide medics across a number of London hospitals with alerts about patients via an app called Streams. The app is meant to provide easy access to patient histories and test results for nurses and doctors. But its AI will also learn to track patterns in patients' blood test data and flag cases that show early signs of kidney injury to the appropriate doctors.

PredictionIO is an open source machine learning server, and is a recent addition to the Apache family.

The venture capital community is seeing literally hundreds, if not thousands, of new companies focused on applying artificial intelligence to resolving business issues and improving consumer experience on and offline. This explosion of interest, on both sides of the Atlantic, is driven by the well-known trends of big data, faster processing speeds, more bandwidth and increasing broadband access. The market numbers are staggering: a financial services firm issued a 300-page report in 2015 explaining why the AI market is projected to grow to $153 billion by 2020: that's $83 billion for robotics and $70 billion for AI-based analytics. The joke in the VC industry is that any start-up that claims to use AI will expect at least a 20 percent valuation premium. There will indeed be great value created but also lots of wasted experiments.

The Machine Learning Canvas is a template for developing new (or documenting existing) intelligent systems based on data and machine learning. It is a visual chart with elements describing the key aspects of such systems: the value proposition, the data to learn from (to create predictive models), the utilization of predictions (to create proposed value), requirements and measures of performance. It assists teams of data scientists, software engineers, product and business managers, in aligning their activities. This tutorial will help you get into the right mindset to go beyond the current hype around machine learning, beyond proofs of concept, and to clearly see how this technology can have an actual impact in your domain. I'll present the general structure of the Canvas, the different boxes it is composed of and the associated questions to answer. We'll see how to fill it in iteratively on a churn prevention example.

Enter a bookstore, while they still exist. Walk toward the philosophy section, toward shelves of fat books by Plato, Nietzsche, Spinoza. Perhaps you browse through their pages before putting them back in their place, respectfully but with a bit of a yawn. More appealing, perhaps: the books at the front of the store, the best sellers, the ones that portend crises (Rise of the Robots: Technology and the Threat of a Jobless Future); others advise on surviving one (Humans Need Not Apply: A Guide to Wealth and Work in the Age of Artificial Intelligence). The business best sellers are more gung-ho about the changes to come: Zero to One: Notes on Startups, or How to Build the Future.

In March 2016, AlphaGo, a computer program developed by Google's London-based DeepMind subsidiary, beat leading professional Go player Lee Se-dol in four games out of five. It was a result that surprised the Go and technology communities in equal measure, and one that has been heralded as a breakthrough in artificial intelligence. The rules of Go are simple, but the sheer number of possible moves available to the players means 2,500-year-old game is considered significantly harder than chess. IBM's Deep Blue computer beat chess champion Garry Kasparov in 1997 but, until 2015, Go programs had only managed to play as well as good amateurs. As significant as AlphaGo's level of competence is the way it was achieved. Rather than basing its decisions on explicit rules about the relative value of different moves, as chess computers do, AlphaGo "taught" itself how to play well, running millions of game simulations against versions of itself and gradually adjusting its algorithms to achieve better results.

Over the past few decades, artificial intelligence, or AI, has morphed from science fiction into an integral part of 21st century life. And what we've seen so far is just the beginning. Experts expect its use to skyrocket in coming years, and market researcher IDC forecasts that by 2020 spending on AI will rise nearly 500 percent to $47 billion from current levels. As Goldman Sachs (GS) noted in a recent report to clients, AI's potential appears boundless. IBM's (IBM) Jeopardy-playing supercomputer Watson may be the technology's best-known example.