Today, artificial intelligence is a hot topic in almost every industry. In fact, you’d be hard-pressed to find any business that wouldn’t benefit from a bit of intuitive supercomputing.
And when you look at the projections, it’s poised to only grow even bigger. By 2020, for example, the artificial intelligence market is expected to surpass $40 billion. That number is estimated to grow to a whopping $100 billion just five years later.
In the modern world of AI supercomputing, there are three systems that are at the forefront of innovation and progress: Google, Watson, and Baidu.
We’ll take a closer look at each of these systems below, point out their claims to fame, and identify their drawbacks as well.
Google: A Deep Learning Machine
If deep learning was a party, Google would be the belle of the ball.
Beginning its first foray into this exciting and intuitive form of artificial intelligence in 2011 with the Google Brain project, Google has since made some incredible strides in the field.
In 2012, the tech giant announced that it had built a neural network which is designed to mimic the categorization and identification techniques of the human brain. From this system, Google Brain was able to successfully identify an image of a cat, a feat that (though seemingly underwhelming) was actually one of the first of its kind in the field.
From there, Google then acquired the deep learning startup Deep Mind and further expanded its deep learning efforts.
A few of the most notable applications that Google’s deep learning systems are being used for today are in image recognition, automated video summaries, language processing, and video suggestions on YouTube.
According to Google AI researcher Francois Chollet, deep learning also has its problems. “The most important problem for AI today is abstraction and reasoning,” Chollet says. “Current supervised perception and reinforcement learning algorithms require lots of data, are terrible at planning, and are only doing straightforward pattern recognition.”
IBM’s Watson: A Real Trivia Whiz
Originally constructed to outperform two of the biggest names in the TV trivia show Jeopardy!, IBM’s Watson is a question-answering computing system that specializes in natural language processing, information retrieval, knowledge representation, and a variety of other higher-level AI functions.
In 2011, Watson faced off against Brad Rutter, winner of the highest Jeopardy! prize pot of $3.25 million, and Ken Jennings who had 74 total winning appearances under his belt.
IBM’s Watson bested the two by over $50,000 over the course of two games and solidified its place in history as the first machine to ever beat a human on the trivia show.
Given the complexity of the Jeopardy! questions, many of which involve a complex question structure, an intricate interplay of words, and other variations in language, many AI systems have a hard time fully understanding what’s being asked.
It seems quite clear that Watson, then, is especially adept at natural language processing.
Watson is currently being used in the healthcare industry to aid in the diagnosis of lung cancer at the Memorial Sloan Kettering Cancer Center in New York City. In fact, Watson has proven to be more effective at identifying cancer than even human doctors.
Where other deep learning systems pull away from the rest, namely in video and picture identification and analysis, Watson is a bit late to the game. Just this year, IBM released a video analysis aspect to Watson while other systems like Google Deep Mind had this capability years ago.
Baidu’s Minwa: A Big Data Powerhouse
Much like its U.S. counterpart, China’s search engine leader Baidu has also developed a supercomputer that harnesses the power of deep learning and machine learning to analyze and learn from enormous swaths of data at an impressive rate.
But unlike its U.S. counterpart, it hasn’t garnered nearly as much media face time in the west. But as the Chinese company expands its marketing on a global scale, we can expect to hear more and more about this powerful name in the AI field.
One of Baidu’s primary advantages comes from its ability to collect a large amount of personal data from the enormous Chinese population. What’s more, the country’s closed-off internet means that Google has had little to no exposure with its citizens and, thus, has not been able to benefit from their data.
That puts Baidu at a distinct advantage over its largest western competitor.
The AI systems of Baidu are currently being applied to a number of fields. In healthcare, its machine and deep learning capabilities are being used to develop a chat function to help diagnose illnesses just like a human doctor.
Augmented reality (AR) is another industry that Baidu’s systems are also currently being applied to. In addition to having a hand in the company’s marketing of the system, the Baidu AI is also playing a role in using it to virtually reconstruct important historic monuments. Also similar to Google, Baidu plans on using their systems to mass-produce driverless cars by 2021.
Like Google though, Baidu may have their hands in too many different ventures to become the be-all end-all of any of them. That being said, they certainly have enough money to handle diversifying their investments.
Three Companies at the Head of The AI Race
There are a variety of startups and small businesses that are investing heavily in AI today but three companies in particular are leading the pack: Google, IBM, and Baidu. Each also comes with their own unique strengths, weaknesses, and applications.
And given their substantial investments of both time and money in the field, it’s hard to imagine the future of AI without them.