Does the brain learn the same way machines learn?
Determining how neural activity changes with learning is anything but black and white. Recently, some have argued that learning in the brain, or biological learning, can be thought of in terms of optimization, which is how learning occurs in artificial networks like computers or robots. A new outlook paper co-authored 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, but can be leveraged to provide valuable information on how the brain works.
“The way we quantify the changes we observe in the brain and in a subject’s behavior during learning is constantly evolving,” says Byron Yu, professor of biomedical engineering and electrical and computer engineering. âIt turns out that in machine learning and artificial intelligence, there is a well-developed framework in which something learns, called optimization. We and others in the field have reflected on how the brain learns in relation to this framework, which was developed to train artificial agents to learn. “
The optimization point of view suggests that activity in the brain should change during learning in a mathematically prescribed way, in the same way that the activity of artificial neurons changes in a specific way when they are. trained to drive a robot or play chess.
âOne thing we want to understand is how the learning process unfolds over time, and not just looking at a snapshot before and after learning,â says Jay Hennig, a recent PhD. graduated in Neural Computing and Machine Learning at Carnegie Mellon. “In this outlook article, we offer three takeaways that would be important for people to consider in the context of thinking about why neural activity might change throughout learning and which cannot be easily explained in terms of optimization. “
Points to remember include the inflexibility of neural variability throughout learning, the use of multiple learning processes even during simple tasks, and the presence of large non-task-specific activity changes.
“It’s tempting to take inspiration from successful examples of machine learning agents and assume that the brain has to do whatever it does,” suggests Aaron Batista, professor of bioengineering at the University of Pittsburgh. . “However, a specific difference between artificial learning and biological systems is that the artificial system usually only does one thing and does it very well. The activity in the brain is quite different, with many processes taking place. are happening at the same time. We and others have observed that there are things going on in the brain that machine learning models cannot yet explain. “
Steve Chase, professor of biomedical engineering at Carnegie Mellon and the Neuroscience Institute adds, âWe see a theme building and a direction for the future. By drawing attention to those areas where neuroscience can inform machine learning and vice versa, we aim to connect them to the optimization view to ultimately understand, on a deeper level, how learning takes place in the brain.
This work is co-authored with Emily Oby, Bioengineering Research School at the University of Pittsburgh, and Darby Losey, Ph.D. studying neural computation and machine learning at CMU. The group’s work is ongoing and carried out in collaboration with the Center for Neural Basis of Cognition, an inter-university research and teaching program between Carnegie Mellon and the University of Pittsburgh that leverages the strengths of each institution to study the mechanisms cognitive and neuronal that give rise to biological intelligence and behavior.
Make the link between engagement and learning
Neuron (2021). DOI: 10.1016 / j.neuron.2021.09.005
Carnegie Mellon University, Department of Biomedical Engineering
Does the brain learn the same way machines learn? (2021, October 13)
retrieved on October 13, 2021
This document is subject to copyright. Apart from any fair use for study or private research, no
part may be reproduced without written permission. The content is provided for information only.