Title: Can AI help developmental science (and vice versa)?
Abstract:
AI models based on deep learning have improved their performance to the point that they can be good candidate first order models of how human perceive objects and process language. Can they also be good candidate first order models of how humans learn such cognitive abilities? Here we examine the case of early language acquisition in infants and review current results on audio representation learning and unsupervised spoken language modeling suggesting a tentative yes, and outline the remaining challenges and roadblocks.
Zoom: https://yale.zoom.us/j/97771780410
If you have any questions or need any further information about the talk, please contact Ka-Fai Yip (kafai.yip@yale.edu).