On Linear Identifiability of Learned Representations.

Geoffrey Roeder Luke Metz Durk Kingma

Published in: ICML (2021)

Keyphrases

linear systems
causal models
learning algorithm
case study
lower bound
training set
image classification