MIT EECS | CS+HASS Undergraduate Research and Innovation Scholar
Associative Memory Using Quantum Systems
- Natural Language and Speech Processing
Robert C. Berwick
Properties of language shared with those of the human computational system are the ability to label structures with a categorical identity, the ability to concatenate two elements, and the ability to transfer them to the conceptual-intentional interface. Oscillations have been found to be related to the timing of cortical information processing; the synchronization of oscillations plays a central role in cortical computations. We propose that universal grammar is not found in the genome but in brain rhythms instead. The goal of this project is to show how properties of the linguistic and human computational system are enabled by a system of coupled oscillators.
“I’ve been working with Professor Bob Berwick since last fall, completing a project developing sparse representations of parse trees that was accepted to a conference. I also built an oscillatory associative memory model for my nonlinear dynamics class that inspired this project. I hope to learn about the process of forming and testing strong hypotheses. I am excited by the chance of finding a more convincing neural representation for language.”