Training variational autoencoders to generate disentangled latent factors
View Patent ↗Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training a variational auto-encoder (VAE) to generate disentangled latent factors on unlabeled training images. In one aspect, a method includes receiving the plurality of unlabeled training images, and, for each unlabeled training image, processing the unlabeled training image using the VAE to determine the latent representation of the unlabeled training image and to generate a reconstruction of the unlabeled training image in accordance with current values of the parameters of the VAE, and adjusting current values of the parameters of the VAE by optimizing a loss function that depends on a quality of the reconstruction and also on a degree of independence between the latent factors in the latent representation of the unlabeled training image.
1. A method performed by one or more computers for training a variational auto-encoder (VAE) to generate disentangled latent factors on a plurality of unlabeled training images,
wherein the VAE has a plurality of parameters and is configured to receive an input image, process the input image to determine a latent representation of the input image that includes a plurality of latent factors, and to process the latent representation to generate a reconstruction of the input image, and
wherein the method comprises:
receiving the plurality of unlabeled training images, and, for each unlabeled training image:
processing the unlabeled training image using the VAE to determine the latent representation of the unlabeled training image and to generate a reconstruction of the unlabeled training image in accordance with current values of the parameters of the VAE, and
adjusting current values of the parameters of the VAE by determining a gradient of a loss function with respect to the parameters of the VAE, wherein the loss function depends on a quality of the reconstruction of the unlabeled training image and also on a degree of independence between the latent factors in the latent representation of the unlabeled training image.
2. The method of claim 1 , wherein the loss function is of the form L=Q−B(KL), where Q is a term that depends on the quality of the reconstruction of the unlabeled training image, KL is a term that measures the degree of independence between the latent factors in the latent representation of the unlabeled training image and an effective capacity of the latent bottleneck, and B is a tunable parameter.
3. The method of claim 2 , wherein B is a value in the range between 2 exclusive and 250 inclusive.
4. The method of claim 3 , wherein B is four.
5. The method of claim 3 , wherein the value of B is dependent on a number of latent factors in the latent representation of the input image.
6. The method of claim 1 , wherein the generative factors are densely sampled from their respective continuous distributions among the plurality of unlabeled training images.
7. The method of claim 1 , wherein the degree of independence between the latent factors is computed within a latent bottleneck with restricted effective capacity.
8. The method of claim 1 , wherein the VAE includes a latent bottleneck layer and a capacity of the latent bottleneck layer is adjusted simultaneously with the degree of independence.
9. A system comprising one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations for training a variational auto-encoder (VAE) to generate disentangled latent factors on a plurality of unlabeled training images, wherein the VAE has a plurality of parameters and is configured to receive an input image, process the input image to determine a latent representation of the input image that includes a plurality of latent factors, and to process the latent representation to generate a reconstruction of the input image, and wherein the operations comprise:
receiving the plurality of unlabeled training images, and, for each unlabeled training image:
processing the unlabeled training image using the VAE to determine the latent representation of the unlabeled training image and to generate a reconstruction of the unlabeled training image in accordance with current values of the parameters of the VAE, and
adjusting current values of the parameters of the VAE by determining a gradient of a loss function with respect to the parameters of the VAE, wherein the loss function that-depends on a quality of the reconstruction of the unlabeled training image and also on a degree of independence between the latent factors in the latent representation of the unlabeled training image.
10. The system of claim 9 , wherein the loss function is of the form L=Q−B(KL), where Q is a term that depends on the quality of the reconstruction of the unlabeled training image, KL is a term that measures the degree of independence between the latent factors in the latent representation of the unlabeled training image and an effective capacity of the latent bottleneck, and B is a tunable parameter.
11. The system of claim 10 , wherein B is a value in the range between 2 exclusive and 250 inclusive.
12. The system of claim 11 , wherein B is four.
13. The system of claim 11 , wherein the value of B is dependent on a number of latent factors in the latent representation of the input image.
14. The system of claim 9 , wherein the generative factors are densely sampled from their respective continuous distributions among the plurality of unlabeled training images.
15. The system of claim 9 , wherein the degree of independence between the latent factors is computed within a latent bottleneck with restricted effective capacity.
16. The system of claim 9 , wherein the VAE includes a latent bottleneck layer and a capacity of the latent bottleneck layer is adjusted simultaneously with the degree of independence.
17. A computer storage medium encoded with instructions that, when executed by one or more computers, cause the one or more computers to perform operations for training a variational auto-encoder (VAE) to generate disentangled latent factors on a plurality of unlabeled training images, wherein the VAE is configured to receive an input image, process the input image to determine a latent representation of the input image that includes a plurality of latent factors, and to process the latent representation to generate a reconstruction of the input image, and wherein the operations comprise:
receiving the plurality of unlabeled training images, and, for each unlabeled training image:
processing the unlabeled training image using the VAE to determine the latent representation of the unlabeled training image and to generate a reconstruction of the unlabeled training image in accordance with current values of the parameters of the VAE, and
adjusting current values of the parameters of the VAE by determining a gradient of a loss function with respect to the parameters of the VAE, wherein the loss function depends on a quality of the reconstruction of the unlabeled training image and also on a degree of independence between the latent factors in the latent representation of the unlabeled training image.
18. The computer storage medium of claim 17 , wherein the loss function is of the form L=Q−B(KL), where Q is a term that depends on the quality of the reconstruction of the unlabeled training image, KL is a term that measures the degree of independence between the latent factors in the latent representation of the unlabeled training image and an effective capacity of the latent bottleneck, and B is a tunable parameter.
19. The computer storage medium of claim 18 , wherein B is a value in the range between 2 exclusive and 250 inclusive.
20. The computer storage medium of claim 18 , wherein B is four.