Skip to yearly menu bar Skip to main content


Poster

The Benefits of Balance: From Information Projections to Variance Reduction

Lang Liu · Ronak Mehta · Soumik Pal · Zaid Harchaoui

East Exhibit Hall A-C #2101
[ ]
Thu 12 Dec 11 a.m. PST — 2 p.m. PST

Abstract:

Data balancing across multiple modalities and sources appears in various forms in foundation models in machine learning and AI, e.g. in CLIP and DINO. We show that data balancing across modalities and sources actually offers an unsuspected benefit: variance reduction. We present a non-asymptotic statistical bound that quantifies this variance reduction effect and relates it to the eigenvalue decay of Markov operators. Furthermore, we describe how various forms of data balancing in contrastive multimodal learning and self-supervised clustering can be better understood, and even improved upon, owing to our variance reduction viewpoint.

Live content is unavailable. Log in and register to view live content