AI Hits the Mainstream

For Robert Welborn, head of data science for the insurer and finance company USAA, 2015 was the year machine learning started to make commercial sense. Access to improved machine-learning tools, cheaper processing technology, and a sharp decline in the cost of storing data were key. When those developments were combined with USAA’s abundance of data, a technology studied for decades suddenly seemed practical.

Insurance, finance, manufacturing, oil and gas, auto manufacturing, health care: these may not be the industries that first spring to mind when you think of artificial intelligence. But as technology companies like Google and Baidu build labs and pioneer advances in the field, a broader group of industries are beginning to investigate how AI can work for them, too.

How will AI develop as it is commercialized, and how will the technology change these diverse industries? Those are the big questions of this Business Report.

Today the industry selling AI software and services remains a small one. Dave Schubmehl, research director at IDC, calculates that sales for all companies selling cognitive software platforms —excluding companies like Google and Facebook, which do research for their own use—added up to $1 billion last year. He predicts that by 2020 that number will exceed $10 billion. Other than a few large players like IBM and Palantir Technologies, AI remains a market of startups: 2,600 companies, by Bloomberg’s count.

That’s because despite rapid progress in the technologies collectively known as artificial intelligence—pattern recognition, natural language processing, image recognition, and hypothesis generation, among others—there still remains a long way to go.

USAA, just one early adopter, has been testing ways to use AI to fine-tune its detection of identity theft. Its system looks for patterns that don’t match a customer’s typical behavior and identifies those anomalies even on the first instance, Welborn says. Traditional systems wouldn’t catch a new pattern of crime until the second time it happened. “Our learning systems are really good at understanding things that look like fraud,” he says.

Another project being tested at USAA tries to improve customer service. It involves an AI technology built by Saffron, a division of Intel, using an approach designed to mimic the randomness of the connections made by the human brain. By combining 7,000 different factors, the technology can match broad patterns of customer behavior to that of specific members, and 88 percent of the time it can correctly predict things like how certain people might next contact USAA (Web? phone? e-mail?) and what products they will be looking for when they do. Without the AI, USAA’s systems were guessing right 50 percent of the time. That test is now being expanded.

General Electric is using AI to improve service on its highly engineered jet engines. By combining a form of AI called computer vision (originally developed to categorize movies and TV footage when GE owned NBC Universal) with CAD drawings and data from cameras and infrared detectors, GE has improved its detection of cracks and other problems in airplane engine blades.

The system eliminates errors common to traditional human reviews, such as a dip in detections on Fridays and Mondays, but also relies on human experts to confirm its alerts. The program then learns from that feedback, says Colin Parris, GE’s vice president of software research.

AI can be a driver of new products and services, too. Through its MyFitnessPal exercise- and calorie-tracking app and other products, sportswear maker Under Armour is connected to 160 million consumers. But rather than merely being limited to logging people’s exercise results, the company made a deal with IBM’s cognitive computing business, Watson, to combine its data about fitness and nutrition routines with information gleaned from research studies and other third-party data on sleep, activity, fitness, and nutrition. The goal: to tell people with a given goal how they can achieve it, making the company more relevant to those 160 million customers.

In what would come to be described as the world’s first computer game, Spanish inventor Leonardo Torres y Quevedo debuts “El Ajedrecista,” a machine that can automatically play chess based on a simple algorithm built into its mechanical design.

Neurophysiologist W. Grey Walter begins work on a series of robotic tortoises that responded to light cues. Their unpredictable movements, he later argued, “might be accepted as evidence of some degree of self-awareness.”

In a paper that helped establish a practical goal for future artificial intelligence research, Alan Turing proposes a game to answer, “Can machines think?” He predicted that by 2000 computers would be able to pass themselves off as human more than 30 percent of the time.

John McCarthy, Marvin Minsky, and Claude Shannon organize a summertime research meeting at Dartmouth that brings together the leading thinkers on information theory, artificial neural networks, and symbolic logic, christening the field “Artificial Intelligence.”

Oliver Selfridge presents a paper in England describing “Pandemonium,” a new model of a neural network based on lower-level “data demons” working in parallel with higher-level “cognitive demons” in order to perform pattern recognition and other tasks.

Frank Rosenblatt demonstrates the Mark-I Perceptron, an attempt to create an artificial neural network for image recognition that the New York Times optimistically calls the first step toward a computer that will “be able to walk, talk, see, write, reproduce itself and be conscious of its existence.”

Marvin Minsky publishes his foundational paper, “Steps Toward Artificial Intelligence.”

Joseph Weizenbaum demonstrates ELIZA, the world’s first chat program, which is able to converse using a series of preprogrammed phrases, sometimes to comic effect.

AI takes a hit when philosopher Hubert Dreyfus publishes “What Computers Can’t Do,” a manifesto that challenged the fanciful predictions of AI researchers, and  scientist James Lighthill publishes a pessimistic review of progress in AI research in the United Kingdom, which leads to local funding cuts.

A backgammon program developed by Hans Berliner defeats the reigning world champion in a match, the first time a computer is able to defeat a champion-level competitor in an intellectual game.

Douglas Lenat begins the Cyc project, an ambitious attempt to create a common-sense knowledge base that would eventually become self-educating. Nevertheless, it eventually falls short.

Ernst Dickmanns and his collaborators in Germany equip a Mercedes van with two analog video cameras, eight microprocessors, and other electronics to demonstrate completely autonomous driving at speeds of almost 60 miles per hour on an empty road.
After a series of research projects fall short of expectations, DARPA begins cuts to the budget for AI research, leading to an “AI Winter.”

IBM’s Deep Blue chess computer avenges its defeat to world champion Garry Kasparov the year before in a tense match that was the subject of a documentary film, The Man vs. The Machine.

Cynthia Breazeal designs a sociable humanoid robot named Kismet that is able to express emotion and recognize cues from interaction with humans.

DARPA sponsors its first “Grand Challenge,” an effort to spur innovation in driverless vehicles by pitting research teams against each other to design vehicles capable of independently traversing the Mojave Desert.

IBM Watson defeats Jeopardy! champions Ken Jennings and Brad Rutter in a televised two-game, three-night face-off that ended with the computer amassing more than three times the winnings of his human competitors.

A team led by Geoff Hinton wins the ImageNet Large Scale Visual Recognition Challenge, demonstrating  that deep-learning software could identify correctly (within five guesses) a thousand diverse types of objects, from mosquito nets to mosques, about 85 percent of the time.

Google acquires DeepMind Technologies, a small, London-based startup focused on deep learning, a relatively new field of artificial intelligence research that aims to achieve tasks like recognizing faces in video or words in human speech.

Google’s AlphaGo beats the European champion of the complex board game Go. Later this year the program will take on the top Go player in the world.

comments powered by Disqus