Google and MIT Are Building an AI Capable of Using Human Senses to Learn
There is a reason why everyone's a little iffy when it comes to talking about artificial intelligence (AI). It runs the risk of privacy and security that people enjoy and it entails the possibility that someday, AIs can hold power over us because of their superior knowledge. However, this does not seem to deter Google and the Massachusetts Institute of Technology (MIT) as they entered a partnership geared in developing an AI.
The research was published in a paper that can be publicly accessed. According to the paper's abstract, it seems that Google and MIT's ambition is to create a system that can learn and solve problems across multiple dimensions. Furthermore, both have taken measures in outlining what the system will be capable of, including being able to interact and communicate with humans.
The main purpose of the experiments is to allow the AI to use algorithms in such a way that is reminiscent of the human senses. The researchers have thus created complicated codes to provide the system with the ability to use sight, sound, and text to communicate. In doing so, Google and MIT are hoping to make their system capable of interacting with the environment around them more effectively.
"We believe that this treads a path towards interesting future work on more deep learning architectures, especially since our model shows transfer learning from tasks with a large amount of available data to ones where the data is limited," the researchers said in the paper.
So far, technological advancements have only allowed AI to function through one human sense at a time. However, Google and MIT hope to radically change that by targeting three in their studies and experiments. If both are successful in their respective roles in the research, the world may be introduced to a system that can learn all about humans and how to interact with them in the world they know.
