“It would also be interesting to consider the possibility of directly leveraging these findings in human learning and Education. For example, can some of the complex competencies be similarly modeled and understood better? The insights could help us systematically design more sophisticated and personalized learning pathways.“
Human species differentiates itself from its peers in the way we use tools. A defining aspect of this ability is in the way we use objects in novel ways to solve our problems and is considered to be a hallmark of our intelligence. An interesting research conducted at MIT’s Center for Brains, Minds and Machines studies the cognitive components that underlie this improvised tool use.
(Note: In India, this innovative usage of available resources is a way of life and even has a special term for it – Jugaad, a word that has also made into the Oxford English Dictionary.)
This research work essentially breaks down the ability for this innovative tool usage into three separate sub-capabilities – sample, simulate and update. The researchers verified this approach by training Machine Learning Models on these sub-capabilities and then combining them to come up with solutions – with outcomes similar to those in the case of humans.
This approach also raises exciting new possibilities of building more human-like forms of intelligence, essentially by adopting similar ‘reinforcement learning’ approaches.
It would also be interesting to consider the possibility of directly leveraging these findings in human learning and Education. For example, can some of the complex competencies be similarly modeled and understood better? The insights could help us systematically design more sophisticated and personalized learning pathways.
1 comment