Differences

This shows you the differences between two versions of the page.

Link to this comparison view

missing_values [2017/08/28 10:58]
127.0.0.1 external edit
missing_values [2018/11/17 22:34] (current)
admin
Line 19: Line 19:
 In this paper, we propose a novel approach for view imputation via generative adversarial networks (GANs), which we name as VIGAN. This approach first treats each view as a separate domain and identifies domain-to-domain mappings through a GAN using randomly-sampled data from each view, and then employs a multi-modal denoising autoencoder (DAE) to reconstruct the missing view from the GAN outputs based on paired data from each view. Then, by optimizing the GANs and DAE jointly, our model enables the knowledge integration learned for domain mappings and view correspondences to effectively recover the missing view. Empirical results on benchmark datasets validate the VIGAN approach by comparing against the state of the art, and an evaluation of VIGAN in a genetic study of substance use disorders further proves the effectiveness and utility of this approach in life science. In this paper, we propose a novel approach for view imputation via generative adversarial networks (GANs), which we name as VIGAN. This approach first treats each view as a separate domain and identifies domain-to-domain mappings through a GAN using randomly-sampled data from each view, and then employs a multi-modal denoising autoencoder (DAE) to reconstruct the missing view from the GAN outputs based on paired data from each view. Then, by optimizing the GANs and DAE jointly, our model enables the knowledge integration learned for domain mappings and view correspondences to effectively recover the missing view. Empirical results on benchmark datasets validate the VIGAN approach by comparing against the state of the art, and an evaluation of VIGAN in a genetic study of substance use disorders further proves the effectiveness and utility of this approach in life science.
  
 +https://​arxiv.org/​abs/​1811.04752v1 Learning Representations of Missing Data for Predicting Patient Outcomes