Awesome

Disentangled VAE

Replicating DeepMind's papers "β-VAE: Learning Basic Visual Concepts with a Constrained Variational Framework" and "Understanding disentangling in β-VAE"

2D shape disentaglement

Result by changing latent Z from -3.0 to 3.0 with γ=100.0 and C=20.0

Latent variables with small variances seem extracting "x", "y", "rotation" and "scale" parameters.

(This experiment is using DeepMind's dsprite data set.)

Z	Image	Parameter	Variance
z0			0.9216
z1			0.9216
z2		Rotation	0.0011
z3		Rotation?	0.0038
z4		Pos X	0.0002
z5			0.9384
z6		Scale?	0.0004
z7			0.8991
z8			0.9483
z9		Pos Y	0.0004

Reconstruction result

Left: original Right: reconstructed image