John Kirchenbauer

I am motivated by the belief that attempting to teach machines to understand and generate natural language (and sometimes failing) is a good way to learn more about what the real building blocks of general intelligence are along the way. Determining whether current progress in language modeling is simply the result of web-scale memorization in parameter space, or the actual emergence of analogs to reasoning and cognition, is one of the most pressing open questions for the field to pursue.

In Tom’s lab, I recently I spent the better part of a year working on techniques discern whether the thing you’re currently reading or looking at was created by a human or generated by an AI system, because all of a sudden, that has become a real challenge. More generally, my research has explored various aspects of deep learning for discrete data like natural language and graphs.

Before starting my PhD at UMD, I worked at Carnegie Mellon University as a research engineer, and I completed my MS and BS in Computer Science at Washington University in St. Louis. Even before that, I received a diploma in Violin from Oberlin College and Conservatory of Music.