Explore the Potential of Continuous Learning to Make AI Systems Learn Like Humans

Researchers say that future AI algorithms could learn just like humans.

Machines can have a hard time retaining memories, just as humans can. Electrical engineers at The Ohio State University analyzed the impact of a \”continual-learning\” process on their performance to help explain why artificial agents can develop holes in their cognitive processes.

Continuous learning occurs when a computer learns to continually learn a series of tasks. It uses the knowledge it has accumulated from previous tasks to learn new ones.

One of the biggest challenges scientists face in order to reach such heights, is how to avoid the \”catastrophic\” forgetting that occurs when artificial neural networks learn new tasks. This issue could be problematic for society as it relies more on AI systems, according to Ness Schroff, Ohio Eminent Scholar, professor of computer science, and engineering, at The Ohio State University.

Source:
https://techxplore.com/news/2023-07-future-ai-algorithms-potential-humans.html

发表回复

您的邮箱地址不会被公开。 必填项已用 * 标注