Equivariant generative prior for inverse problems with unknown rotation

Patent metadata

Field	Value
Publication number	US-12452108-B2
Application number	US-202318100263-A
Country	US
Kind code	B2
Filing date	Jan 23, 2023
Priority date	Jan 27, 2022
Publication date	Oct 21, 2025
Grant date	Oct 21, 2025

Abstract

A processor-implemented method for estimating a channel by a deep generative model includes receiving, at a device, an observation of the channel and mapping, at the device, the observation to a mean value associated with the channel and a covariance matrix associated with the channel. The processor-implemented method also includes reconstructing, at the device, the channel based on the mean value and the covariance matrix.

First claim

What is claimed is:

1. A processor-implemented method for estimating a channel by a deep generative model, comprising:
receiving, at a device, an observation of the channel;
mapping, at the device, the observation to a mean value and a covariance matrix that together define a probability distribution representing an estimate of the channel associated with the observation; and
reconstructing, at the device, the channel based on the mean value and the covariance matrix.

2. The processor-implemented method of claim 1 , in which the deep generative model comprises an equivariant variational autoencoder (VAE).

3. The processor-implemented method of claim 2 , in which the observation is mapped to the mean value and the covariance matrix at an equivariant encoder of the equivariant VAE.

4. The processor-implemented method of claim 2 , in which the channel is reconstructed at an equivariant decoder of the equivariant VAE.

5. The processor-implemented method of claim 1 , further comprising generating a latent space representation based on the mean value and the covariance matrix.

6. The processor-implemented method of claim 5 , in which the channel is reconstructed based on the latent space representation.

7. The processor-implemented method of claim 1 , in which the channel is a wireless communication channel.

8. The processor-implemented method of claim 1 , further comprising training the deep generative model by mapping the observation to a latent code associated with the channel based on gradient descent over a latent space.

9. An apparatus for estimating a channel, comprising:
means for receiving, at a device, an observation of the channel;
means for mapping, at the device, the observation to a mean value and a covariance matrix that together define a probability distribution representing an estimate of the channel associated with the observation; and
means for reconstructing, at the device, the channel based on the mean value and the covariance matrix.

10. The apparatus of claim 9 , in which the apparatus comprises means for equivariant variational autoencoding.

11. The apparatus of claim 10 , in which the means for equivariant variational autoencoding comprises means for equivariant encoding.

12. The apparatus of claim 10 , in which the means for equivariant variational autoencoding comprises means for equivariant decoding.

13. The apparatus of claim 9 , further comprising means for generating a latent space representation based on the mean value and the covariance matrix.

14. The apparatus of claim 13 , in which the means for reconstructing the channel comprises means for reconstructing the channel based on the latent space representation.

15. The apparatus of claim 9 , in which the channel is a wireless communication channel.

16. The apparatus of claim 9 , further comprising means for training a deep generative network associated with the apparatus, the means for training comprising means for mapping the observation to a latent code associated with the channel based on gradient descent over a latent space.

17. An apparatus for estimating a channel via a deep generative model, comprising:
at least one processor; and
at least one memory coupled with the at least one processor and storing instructions operable, when executed by the at least one processor, to cause the apparatus to:
receive an observation of the channel;
map the observation to a mean value and a covariance matrix that together define a probability distribution representing an estimate of the channel associated with the observation; and
reconstruct the channel based on the mean value and the covariance matrix.

18. The apparatus of claim 17 , in which the deep generative model comprises an equivariant variational autoencoder (VAE).

19. The apparatus of claim 18 , in which the observation is mapped to the mean value and the covariance matrix at an equivariant encoder of the equivariant VAE.

20. The apparatus of claim 18 , in which the channel is reconstructed at an equivariant decoder of the equivariant VAE.

21. The apparatus of claim 17 , in which execution of the instructions further cause the apparatus to generate a latent space representation based on the mean value and the covariance matrix.

22. The apparatus of claim 21 , in which execution of the instructions further cause the apparatus to reconstruct the channel based on the latent space representation.

23. The apparatus of claim 17 , in which the channel is a wireless communication channel.

24. The apparatus of claim 17 , in which execution of the instructions further cause the apparatus to train the deep generative model by mapping the observation to a latent code associated with the channel based on gradient descent over a latent space.

25. A non-transitory computer-readable medium having program code recorded thereon for estimating a channel via a deep generative model, the program code executed by a processor and comprising:
program code to receive an observation of the channel;
program code to map the observation to a mean value and a covariance matrix that together define a probability distribution representing an estimate of the channel associated with the observation; and
program code to reconstruct the channel based on the mean value and the covariance matrix.

26. The non-transitory computer-readable medium of claim 25 , in which the deep generative model comprises an equivariant variational autoencoder (VAE).

27. The non-transitory computer-readable medium of claim 26 , in which the observation is mapped to the mean value and the covariance matrix at an equivariant encoder of the equivariant VAE.

28. The non-transitory computer-readable medium of claim 26 , in which the channel is reconstructed at an equivariant decoder of the equivariant VAE.

29. The non-transitory computer-readable medium of claim 25 , in which the program code further comprises program code to generate a latent space representation based on the mean value and the covariance matrix.

30. The non-transitory computer-readable medium of claim 29 , in which the program code further comprises program code to reconstruct the channel based on the latent space representation.

Assignees

• Qualcomm Inc

Inventors

• Behboodi Arash
• Kuzina Anna
• Massoli Fabio Valerio
• Pratik Kumar

Classifications

• H04L25/0242 Primary

  using matrix methods · CPC title

• H04L25/0254 Primary

  using neural network algorithms · CPC title