Artificial Intelligence, one of the ripest fields in technology right now, may be on the verge of a major breakthrough in how software systems are designed. Researchers in the field of Artificial Intelligence have found out that Artificial Intelligence software can, in some cases, best humans in designing new AI software.
MIT took a look at the work of some of the leaders in the field of AI, including the Google Brain Team, OpenAi, UC Berkeley, DeepMind, and MIT itself.
Google’s Brain Team, dedicated to “making machines intelligent”, designed a particular AI-based software that could design a machine-learning system to take a test used to benchmark software that processes language. Not only did it do this on its own, but it even outdid the software designed by humans.
Much has been said about how the advent of computers has resulted in catapulting unemployment rates through the roof. However, if advancements in the field of Artificial Intelligence keep happening at the same pace at which they are happening currently, we might soon be seeing computers take over an even larger chunk of the job industry.
Truly intelligent machines are still long ways down the road and there is a very limited number of cases where AI can be used to design AI software on itself. Work, such as the one done by Google’s Brain Team, is just a glimpse of what lies ahead. Even if it is applicable, there are a lot of kinks and quirks that need to be worked out before such systems can be used.
One of the chief drawbacks is the need for a huge amount of processing power it needs currently. On the other hand, a lot of technology-pioneers, such as the self-made billionaire and entrepreneur Elon Musk, are also genuinely concerned about making sure that artificial intelligence is kept in check.
The advantages, however, are many. It lessens the amount of mundane work that needs to be done by humans and also eliminates the need for manually training the AI systems using huge data sets.
Image Credits — WorldVersus