Class
Article
College
College of Science
Department
English Department
Faculty Mentor
Kevin Moon
Presentation Type
Poster Presentation
Abstract
The integration of multimodal data presents a challenge in cases when the study of a given phenomena by different instruments or conditions generates distinct but related domains. In this paper, we propose Diffusion Transport Alignment (DTA) a semi-supervised manifold alignment method that exploits prior correspondence knowledge between distinct data views. DTA finds a bijection between two domains, which by assumption, share a similar geometrical structure coming from the same underlying data generating process. We empirically demonstrate the effectiveness of our method to integrate multimodal data, as well as how it can improve the performance of machine learning tasks, otherwise less effective when only one of the domains is considered.
Location
Logan, UT
Start Date
4-8-2022 12:00 AM
Included in
Manifold Alignment With Inter-Domain Diffusion
Logan, UT
The integration of multimodal data presents a challenge in cases when the study of a given phenomena by different instruments or conditions generates distinct but related domains. In this paper, we propose Diffusion Transport Alignment (DTA) a semi-supervised manifold alignment method that exploits prior correspondence knowledge between distinct data views. DTA finds a bijection between two domains, which by assumption, share a similar geometrical structure coming from the same underlying data generating process. We empirically demonstrate the effectiveness of our method to integrate multimodal data, as well as how it can improve the performance of machine learning tasks, otherwise less effective when only one of the domains is considered.