Jonathan Bussing

Separately Trained Senses Converge on Shared Latent Structure: An Empirical Check of Identifiability Predictions Without Shared Data

Identifiability theory for contrastive learning predicts that an encoder trained with suitable positive pairs recovers the latent structure behind its inputs, so separately trained networks that perceive a common world should converge on its shared structure without shared training data. This is one proposed explanatio...