Disentangling by Factorising
This work addresses the problem of learning independent factors in data for researchers in representation learning, though it is incremental as it builds on prior methods like β-VAE.
The paper tackles unsupervised learning of disentangled representations by proposing FactorVAE, which improves the trade-off between disentanglement and reconstruction quality compared to β-VAE, and introduces a new metric to address issues with existing ones.
We define and address the problem of unsupervised learning of disentangled representations on data generated from independent factors of variation. We propose FactorVAE, a method that disentangles by encouraging the distribution of representations to be factorial and hence independent across the dimensions. We show that it improves upon $β$-VAE by providing a better trade-off between disentanglement and reconstruction quality. Moreover, we highlight the problems of a commonly used disentanglement metric and introduce a new metric that does not suffer from them.