According to the record, artificial intelligence is still almost different from the human brain. At best, artificial intelligence has reached the point where it draws inspiration from the human brain, just like learning and changing with experience. According to reports, a new “fluid” form of artificial intelligence is adapting to multiple situations, which makes it an interesting new achievement for many experts.
(Photo: Photo by Paweł Czerwiński on Unsplash)
according to Singular hubResearchers have shown that a new technology called “liquid neural networks” allows machines that learn algorithms to improve their skills even after the initial training period. Experts say that it has a built-in system called “neuroplasticity”, just like the human brain, machines can learn through experience. For example, if AI is used to drive a car, it can learn based on the different experiences it has experienced or even exceeded its initial learning stage.
Liquid neural networks focus on quality rather than size
Although AI is inspired by the human brain, the current human system is still very complex. Therefore, the experts managed to adjust the method to adapt to simpler and smaller models, with results similar to humans. Liquid AI focuses on quality, not the scale of large AI models like OpenAI, Micorosft, and Google.
The study is titled “Flow time constant networkA paper published by researchers from the Massachusetts Institute of Technology and the Austrian Academy of Science and Technology showed that these liquid neuron machine algorithms were inspired by the 302 neurons that make up the Caenorhabditis elegans system.
Supporters of the study believe that it is better to focus on the quality of work provided by AI, rather than the scale of machine learning algorithms. The neural connections in our brain have the same concept. Our brain cells are connected to many other cells. In the case of a neural network, the system will provide the best results according to the situation, but will be hindered by the parameters to feed data where it responds.
Will the robots of the future adapt to the new experience?
The newly discovered liquid network allows parameters to be expanded and changed over time based on experience. This makes it more realistic and practical in the real world undergoing various changes. In the future, robots may have the same adaptability characteristics as humans when facing different experiences. In a world where emergencies are unpredictable, adaptability is essential for humanoid robots.
However, according to Next websiteSome critics say that driving larger AI machines is expensive and wasteful. This new study still needs to be tested in different situations before it can be determined whether it is worth the change. And, given that robots can now think critically like humans, it may still have a negative impact. However, researchers believe that this may be a promising cutting-edge AI technology that can mimic the infinite parameters of the human brain.
This article is owned by Techtimes
Written by Nikki D
Ⓒ2018 TECHTIMES.com reserves all rights. Do not copy without permission.