Skip to content

Latent Distribution Matching (LDM)

Home / Concepts / Latent Distribution Matching (LDM)

Latent Distribution Matching (LDM)

Auto-generated stub. Edit this file to add more details.

A theoretical SSL framework that reconciles representation alignment and entropy maximization for latent space modeling.

Why It Matters

Provides a unified theoretical framework for diverse SSL methods including contrastive, non-contrastive, and predictive paradigms.

Evidence

We cast SSL as latent distribution matching (LDM): learning representations that maximize their log-probability under an assumed latent model (alignment), while maximizing latent entropy to prevent collapse (uniformity).

Metadata & Links