Revealing Multimodal Contrastive Representation Learning through Latent Partial Causal Models
Multimodal contrastive representation learning methods have proven successful across a range of domains, partly due to their ability to generate meaningful shared representations of complex phenomena. To enhance the depth of analysis and understanding of these acquired representations, we introduce a unified causal model specifically designed for multimodal data. By examining this model, we show that multimodal contrastive representation learning excels at identifying latent coupled variables within the proposed unified model, up to linear or permutation transformations resulting from different assumptions. Our findings illuminate the potential of pre-trained multimodal models, eg, CLIP, in learning disentangled representations through a surprisingly simple yet highly effective tool: linear independent component analysis. Experiments demonstrate the robustness of our findings, even when the assumptions are violated, and validate the effectiveness of the proposed method in learning disentangled representations.
updated: Fri Feb 09 2024 07:18:06 GMT+0000 (UTC)
published: Fri Feb 09 2024 07:18:06 GMT+0000 (UTC)