Current automated speech recognition attempts to recognize context-dependent phones. This requires large amounts of training data to be effective and does not model human speech processing. Utilizing advancements in linguistic theory, acoustic analysis, and human auditory perception, the Linguistic Event Extraction and Interpretation project aims to address these shortcomings. Speech data labeled with individual acoustic cues is at the center of this project. A specialized data structure is needed to contain the labeled acoustic data while maintaining the theoretical hierarchal relationship between different linguistic units. Utilizing graph comparison theory and empirical linguistic cue relationships we create such a data structure with methods for comparison and modification.

SuperUROP is a great opportunity for me to apply the skills I have learned in my coursework in a meaningful way. The courses I took as an underclassman provided me with the tools I needed for my research. I hope to learn more about the intersection of humans and technology through this work. Automated speech recognition that utilizes linguistics and human auditory perception sits on that boundary and I am excited to pursue it further.