Back

 Industry News Details

 
The 6 most exciting AI advances of 2016 Posted on : Dec 16 - 2016

Innovations in driverless vehicle technology, voice recognition, and deep learning show how much machines can do to improve our lives. Here are the best examples from 2016.

 In 2016, major automakers like Tesla and Ford announced timelines for releasing fully-autonomous vehicles. DeepMind's AlphaGo, Google's AI system, beat the world champ Lee Sedol at one of the most complex board games in history. And other major advancements in AI have had big implications in healthcare, with some systems proving more effective in diagnosing cancer than human doctors.

Want to learn what other cool things AI did in 2016?

1. AlphaGo beats world champion at the game Go

In January 2016, Google's DeepMind achieved a major victory in deep learning—AlphaGo, a division of the company, mastered the ancient Chinese game Go. According to Demis Hassabis, head of DeepMind, this milestone came about ten years earlier than experts thought it would. At the annual Association for the Advancement of Artificial Intelligence conference (AAAI) in February, Hassabis called Go the "most complex professional game man has ever devised," because of the huge number of potential moves that can be made. It's a game that many experts say relies on human intuition. AlphaGo was able to master the game by learning on millions of training samples from real games. In March, AlphaGo faced off with Lee Sedol, the world champion, and beat Sedol in four out of five games.

2. Tesla's Autopilot brings man with blood clot to hospital

Many news stories about Tesla's Autopilot—a semi-autonomous driving feature that includes speed-adjusting, lane-switching, and automatic braking—have focused on the fatality that occurred when Joshua Brown's Tesla Model S, engaged in this mode, crashed into an 18-wheeler crossing its path. But Elon Musk, Tesla's CEO, continues to stress the point that cars are safer with Autopilot than without it. And stats back this up: According to a 2015 report by the US National Safety Council, the estimated annual mileage death rate is 1.3 deaths per 100 million vehicle miles traveled, while Autopilot-users have driven more than 130 million miles with only one verified fatality.

But, here's one concrete way Autopilot did something great: It helped bring one man who was suffering a blood clot to the hospital. Joshua Neally began experiencing constriction in his chest while driving home from work, down the highway in Springfield, MO. His car, engaged in Autopilot, helped him navigate almost all the way to the hospital, and Neally credits the autonomous feature with saving his life.

3. Swarm AI predicts the Kentucky Derby

In May, the AI platform UNU made big waves after it was able to successfully predict the superfecta—top four horses, in exact order—at the Kentucky Derby. The "swarm" is a real-time online tool that brings people together to make a group decision. Louis Rosenberg, CEO of Unanimous A.I., which created the platform, had been asked to create a "swarm" that could predict the Kentucky Derby—a race that he and others thought would be virtually impossible to predict.

 The swarm predicted the exact superfecta, which none of the official Kentucky Derby experts had done, beating 540-1 odds.

4. Microsoft's AI can now understand speech better than humans

Speech recognition came a long way in 2016, with virtual assistants like Echo becoming hugely popular. In October, research from Microsoft showed that, in comparing the ability of AI with humans in a conversational speech task, "For the first time...automatic recognition performance [is] on par with human performance on this task." Its system used convolutional and recurrent neural networks, trained on 2,000 hours of data, to achieve this victory.

5. AI predicts US election

While the outcome of the US presidential came as a huge surprise to many, even insiders, one AI system saw it coming: An Indian startup in Mumbai called MogIA predicted a Trump win. The company analyzed social media sentiment through 20 million social media data points. In this way, it may have tapped into true voter preferences in a way that traditional polling did not. (TechRepublic held a "swarm AI" to predict the election, which gave a narrow edge to Clinton. While inaccurate, given her lead in the popular vote—over 2.8 million, as of this writing—it may not have been that far off, after all.)

While some AI experts are cautious about giving MogIA's victory too much weight, it was, nonetheless, able to predict an event that most humans were not.

6. AI helps diagnose cancer

"There have been major AI advances in healthcare," said Roman Yampolskiy, director of the Cybersecurity Lab at the University of Louisville. For instance: IBM Watson may be able to spot health issues that your doctor can't. Speaking at the Business Insider's Ignition Conference, David Kenny, general manager of IBM Watson, told a story about how the cognitive machine detected leukemia in a woman in Japan that had been previously missed. "Statistically, we're seeing that about one third of the time, Watson is proposing an additional diagnosis," Kenny said.

And one AI program at the Houston Methodist Research Institute in Texas reviewed millions of mammograms—30 times faster than humans— and was able to achieve 99% accuracy in diagnosing cancer. "My mom had breast cancer," said Fabio Cardenas, president of Sundown AI, "so it's encouraging to see AI to improve differential diagnoses and patient care." Source