This weekly column will feature the latest topics and trends constantly emerging in a rapidly-growing technical world. Each week, we will focus on a specific technical topic, with topics covering a wide range of technical issues and the latest technical gadgets.
If there are any topics or products that you would like to see covered in this column, please email editor@etown.edu.
Tech companies are currently racing to be the elite power in artificial intelligence, or AI. Researchers in the field of AI are not in any sort of agreement as to how soon the world will have an AI that can beat a human in task completion and efficiency.
AI research is full of assumptions, and experts can only make educated guesses on the speed of advancements. Nevertheless, experts push forward in hopes of being the ones to reach the pinnacle achievement.
Tech giants such as Facebook, Google and Microsoft are in a heated competition to recruit top talent for their AI research programs.
This competition is so intense that IBM, an older giant in the industry, has committed to a $240 million partnership with researchers at the Massachusetts Institute of Technology (MIT) over the next 10 years.
According to Forbes, this project will bring together approximately 100 different academics. The project will focus on investigating and improving the four primary areas of AI research. The areas are hardware, new algorithms, social effects and business applications of AI.
The algorithm research dives into the ability of AI to move beyond the current deep learning of machine algorithms.
The hope is to get AI to quantum computing. This push into AI research is an attempt to stay relevant and begin to once more move in the direction of positive growth. With the scarcity of AI talent in the workforce, IBM is likely to lose this race for AI. The market to research in the field booms though, as companies are in bidding wars for talent.
Forbes reports that an entry level Ph.D. or someone with a lower degree and experience can make anywhere from $300,000-$500,000 a year.
Google made a huge announcement last week about one of its most successful AI projects. Go, an ancient Chinese board game, is the most complex two-player board game ever, and in 2015 Google’s AI program AlphaGo was tested on the game. AlphaGo was able to beat 60 grandmasters and the world champion.
The AI was trained by pitting it against human opponents. Google announced last week that the AlphaGo AI was just beaten by a new AI named AlphaGo Zero. AlphaGo Zero was taught the rules by human input and progressed over the course of three days to learn the game via matches against itself.
The AI uses probability calculations to analyze moves and solve problems. Researchers say that the victory of AlphaGo Zero over its predecessor is huge because while AlphaGo trained with humans, AlphaGo Zero had no human input after learning the rules. This self-teaching could enable AI to adapt to real-world problems and find ways to solve them.
A group of Japanese researchers challenged the ideas of whether an AI could learn like humans do. To test the levels of AI comprehension, the researchers developed Todai Robot, an AI developed to take the entrance exam of the University of Tokyo, the highest university in Japan.
The AI was able to pass 60 percent of entrance exams in Japan. It finished in the top one percent in the first math portion of the University of Tokyo entrance exam. Todai Robot finished in the top 20 percent of the written University of Tokyo entrance exam but did not score high enough to be accepted.
The flaws still lay in the lack of ability to read and comprehend. This technology can search and compute but cannot comprehend.
Artificial Intelligence has come a very long way over recent years. It is breaking barriers and pushing the boundaries of human knowledge through its processing and computational abilities.
This technology is fast on the rise and tech companies are racing to be successful.
