Dr. Andrea E. Martin: "Towards a model of language processing in a neurophysiological system"

Описание: Human language is a fundamental biological signal with computational properties that differ from other perception-action systems: hierarchical relationships between sounds, words, phrases, and sentences, and the unbounded ability to combine smaller units into larger ones, resulting in a "discrete infinity" of expressions that are often compositional. These properties have long made language hard to account for from a biological systems perspective and within models of cognition. In this talk, I synthesize insights from the language sciences, computation, and neuroscience that center on the idea that time can be used to combine and separate representations. I describe how a well-supported computational model from a related area of cognition capitalizes on time and rhythm in computation, and how neuroscientific experiments can then be instrumentalized to determine the computational bounds on artificial neural network models. I offer examples of the approach from cognitive neuroimaging data and computational simulations, including leveraging other existing models. I outline a developing a theory of how language is represented in the brain that integrates basic insights from linguistics and psycholinguistics with the currency of neural computation.

This recording is from an invited talk at Columbia University on February 26th, 2021.