A study on new perspectives jointly noted by researchers at Carnegie Mellon University and the University of Pittsburgh links machine learning to biological learning, showing that the two approaches are not interchangeable, yet can be exploited , and provides valuable insight into brain function.
“We are constantly evolving how to quantify changes in brain and subject behavior during learning. It has been found that there is a well-established We and others in the field have been thinking about how the brain learns compared to this framework, which was developed to teach artificial agents to learn, said Byron Yu, a professor at Carnegie Mellon University.
The optimization point of view suggests that brain activity should change in a mathematically prescribed way during learning, similar to just as the activity of artificial neurons changes in a certain way when they are trained to lead or play chess.
“One of the things we want to understand is how the learning process unfolds over time, if we don’t just look at a snapshot before and after learning. The article has three main lessons that would be important for people to consider when thinking about why neural activity can change during learning that cannot be easily explained in terms of optimization, ”explained Jay Hennig, Carnegie Mellon of Neural Computing and
Lessons include the inflexibility of neural variability in learning, the use of multiple learning processes even in simple tasks, and the presence of large task-specific activity changes.
“It is tempting to draw from successful examples of artificial learning agents and to assume that the brain has to do what they do. However, one specific difference between artificial and biological learning systems is that the artificial system usually only does one thing and does it very well. The activity in the brain is quite different, many processes take place at once. We and others have observed things happening in the brain that machine learning models can’t yet account for, “said Aaron Batista, a professor of bioengineering at the University of Pittsburgh.
Steve Chase, Carnegie Professor Mellon added: “We see that a theme is being built and we see the direction of the future. By drawing attention to these areas where neuroscience can inform machine learning and vice versa, we aim to link them to the optimization approach so that we finally understand at a deeper level how learning unfolds in the brain. ”
Hardware, software, tests, curiosities and colorful news from the IT world by clicking here