loading page

An Improved Uncertainty Autoencoder with Blurred Measurements
  • Hongguang Xu,
  • Ke Xu,
  • Weiqiang Wu
Hongguang Xu
Harbin Institute of Technology Shenzhen
Author Profile
Ke Xu
Shenzhen Polytechnic

Corresponding Author:[email protected]

Author Profile
Weiqiang Wu
Shenzhen Polytechnic
Author Profile

Abstract

Compressed sensing (CS) techniques have enabled efficient acquisition and recovery of sparse high-dimensional data via succinct low-dimensional projections, which usually consist of an encoder and a decoder. Unlike conventional CS techniques with the encoding-decoding architecture, the uncertainty autoencoder (UAE) can sample from the learned input data distribution without an explicit likelihood function, hence avoids potential uninformative latent representations. However, existing works on UAE mainly focus on the encoders and maximize the lower bound of the mutual information between input and measurements, rather than the decoders, which brings the shortcoming that the two may not cope well. In this letter, we propose a novel training scheme for UAE that blurs the measurements to learn the encoder and decoder simultaneously. Experimental results show that the proposed method improves the reconstruction performances when applied to UAE.
23 Mar 2023Submitted to The Journal of Engineering
27 Mar 2023Submission Checks Completed
27 Mar 2023Assigned to Editor
06 Apr 2023Reviewer(s) Assigned
30 May 2023Review(s) Completed, Editorial Evaluation Pending
31 May 2023Editorial Decision: Revise Minor
12 Jun 20231st Revision Received
15 Jun 2023Submission Checks Completed
15 Jun 2023Assigned to Editor
25 Jun 2023Reviewer(s) Assigned
17 Aug 2023Review(s) Completed, Editorial Evaluation Pending
21 Aug 2023Editorial Decision: Revise Minor
01 Sep 20232nd Revision Received
01 Sep 2023Submission Checks Completed
01 Sep 2023Assigned to Editor
03 Sep 2023Review(s) Completed, Editorial Evaluation Pending
03 Sep 2023Editorial Decision: Accept