Official implementation of:
Understanding Self-Supervised Learning via Latent Distribution Matching
Fabian A. Mikulasch, Friedemann Zenke
https://arxiv.org/abs/2605.03517
This repository contains code for Latent Distribution Matching (LDM), a probabilistic framework unifying contrastive, non-contrastive, and predictive self-supervised learning methods.
For reproduction, please refer to the README.md in the subfolders.