On the Importance of Contrastive Loss in Multimodal Learning

Yunwei Ren, Yuanzhi Li

Keywords: multimodal learning, contrastive learning

Abstract: Recently, contrastive learning approaches (e.g., CLIP (Radford et al., 2021)) have received huge success in multimodal learning, where the model tries to minimize the distance between the representations of different views (e.g., image and its caption) of the same data point, while keep the representations of different data points away from each other. However, from a theoretical perspective, it is unclear how contrastive learning can learn to align the representations from different views efficiently, especially in cases where the data is not isotropic. In this work, we analyze the training dynamics of a simple multimodal contrastive learning model, and show that contrastive pairs are important for the model to efficiently balance the learned representations. In particular, we reveal a stage-wise behavior of the learning process: In the first stage, the model aligns the feature representations using positive pairs and the condition number grows in this stage. Then, in the second stage, the model reduces the condition number of the learned representations using negative pairs.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Unsupervised and Self-supervised learning

TL;DR: We show that contrastive pairs are important for models to learn aligned and balanced representations in multimodal learning.