M.S. Final Oral Exam: Yun Bang

Event
Speaker: 
Yun Bang
Monday, November 21, 2022 - 9:00am
Event Type: 

Interpreting Deep Text Quantification Models

Quantification learning is a relatively new deep learning task. Differing from a classic classification problem where the class of a single instance is predicted, a quantification model predicts the distribution of classes within a given set of instances. Quantification learning has applications in various domains. For example, in designing political campaign ads, it is important to know the proportion of different aspects voters care about. QuaNet is a recent deep learning quantification model that was shown to achieve good quantification performance. Like many deep learning models, there is no explanation about the contributions of different input QuaNet uses to predict a class distribution. In this study, we propose a method to provide such an explanation, which is important to increase users' trust towards the model. Our method is the first work on interpreting deep learning quantification models.

Committee: Wallapak Tavanapong (co-major professor), Zhu Zhang (co-major professor), and Qi Li.

Join on Zoom: Please click this URL to start or join. https://iastate.zoom.us/j/91605544815 Or, go to https://iastate.zoom.us/join and enter meeting ID: 916 0554 4815