Zoom link: https://mit.zoom.us/j/92562397534

-----
Speaker: Gasser Elbanna

Affiliation: McDermott Lab

Title: Modeling Continuous Speech Recognition to Understand Contextual Effects in Human Speech Perception

Abstract: Humans excel at transforming acoustic waveforms into meaningful linguistic representations, despite the inherent variability of speech signals. The perceptual and neural mechanisms that enable such robust perception remain unclear. Progress has been limited by the lack of 1) stimulus-computable models that replicate human behavior and 2) large-scale behavioral benchmarks for comparing model and human speech perception. I will present our work on developing candidate models of continuous speech perception along with new behavioral experiments to compare phonemic judgments in humans and models. Our models reproduce patterns of human responses and confusions, and by manipulating the model’s access to past and future speech input, we are testing the role of context in shaping human speech perception.

 Bio: Gasser is a second-year PhD student in the Speech and Hearing, Bioscience and Technology (SHBT) program at Harvard University. He works with Josh McDermott at the Laboratory for Computational Audition at MIT. His research aims to understand how the brain dynamically perceives, encodes, and integrates speech information over time, thereby unraveling the perceptual and neural foundations of auditory intelligence.