CHINA TOPIX

11/02/2024 03:30:19 am

Make CT Your Homepage

New Aritifical Intelligence Can Now Learn Like Humans Do

A team of scientists has developed an algorithm that captures our learning abilities, enabling computers to recognize and draw simple visual concepts that are mostly indistinguishable from those created by humans.

(Photo : Pixabay) A team of scientists has developed an algorithm that captures our learning abilities, enabling computers to recognize and draw simple visual concepts that are mostly indistinguishable from those created by humans.

A team of researchers have developed a machine and taught it to learn skills like humans do. According to researchers from New York University and MIT, they developed a special algorithm that can capture human learning abilities that enables the computers to recognize and even draw out visual concepts and ideas that are hard to distinguish that are created by humans.

Like Us on Facebook

This breakthrough work is considered to be a significant advancement in the field of artificial intelligence where this new computer can shorten the time it takes to learn new concepts that also broadens their scope when it come to creative tasks.

According to lead author of the study, data science fellow at NYU, Brenden Lake, these results reveal that reverse engineering can help develop better algorithms where this can narrow the gap when it comes to more promising methods for other machine learning tasks.

For example, when humans learn a new concept, like a new musical instrument, a new dance move or how to operate a new kitchen tool, humans need only a few examples to be able to grasp the concept and understand its composition where humans can recognize it the next time around. For machines however, this new machine can replicate this pattern recognition skill in humans but they will need hundreds or even thousands of examples to perform with the same accuracy as humans. 

According to co-author of the study, Joshua Tenenbaum from MIT (Massachusetts Institute of Technology), it is still a long way to go to build a machine as smart as a human child however, this breakthrough marks the first time that a machine is able to learn and utilize a large amount of real world concepts even visual cues like handwritten characters that require a certain amount of time to distinguish in humans.

This new study is published in the journal, Science. 

Real Time Analytics