Improving Natural Language Understanding through Adversarial Testing

Artificial intelligence is one of the most disruptive technologies of our time. New applications and use cases are being developed at a stunning pace, extending into virtually every industry. But, according to Stanford Professor Christopher Potts, while we are living in a golden age of AI research, major breakthroughs are still ahead of us.

In this webinar Professor Potts gives an overview of his course, Natural Language Understanding (XCS224U). He focuses specifically on the topic of "adversarial testing" in which developers challenge top-performing systems with examples that they suspect will trip up the systems, thus revealing their weaknesses.

We also hear from students who have completed the course as they present original projects that they developed in the course, and describe how the course structure guided their research.