Extending Contrastive Learning to Public Opinion Data: Contrastive Scaling as an Alternative to Bridging

• In Event: Advances in Machine Learning

Brief Overview

Through contrasting data across countries and across time, we demonstrate that the combination of a relatively new machine learning approach, i.e., contrastive learning, and scaling can be considered as an alternative to bridging.

Authors

• Tzu-Ping Liu, University of Taipei

• Sam Fuller, University of California-Davis