notesum.ai
Published at December 6Latent Space Characterization of Autoencoder Variants
cs.LG
cs.CV
cs.IT
math.IT
Released Date: December 6, 2024
Authors: Anika Shrivastava1, Renu Rameshan1, Samar Agnihotri1
Aff.: 1University
| Noise levels | CAE (latent shape: 7x7x128) | DAE (latent shape: 7x7x128) | VAE (latent shape: 7x7x128) |
|---|---|---|---|
| zero | S1: (5, 7), S2: (6, 7), S3: (29, 48) | S1: (6, 7), S2: (6, 7), S3: (29, 48) | S1: (7, 7), S2: (7, 7), S3: (48, 48) |
| 0.01 | S1: (6, 7), S2: (7, 7), S3: (30, 48) | S1: (7, 7), S2: (6, 7), S3: (42, 48) | S1: (7, 7), S2: (7, 7), S3: (48, 48) |
| 0.02 | S1: (5, 7), S2: (7, 7), S3: (30, 48) | S1: (6, 7), S2: (7, 7), S3: (42, 48) | S1: (7, 7), S2: (7, 7), S3: (48, 48) |
| 0.03 | S1: (6, 7), S2: (7, 7), S3: (32, 48) | S1: (6, 7), S2: (7, 7), S3: (42, 48) | S1: (7, 7), S2: (7, 7), S3: (48, 48) |
| 0.04 | S1: (6, 7), S2: (7, 7), S3: (32, 48) | S1: (7, 7), S2: (7, 7), S3: (43, 48) | S1: (7, 7), S2: (7, 7), S3: (48, 48) |
| 0.05 | S1: (6, 7), S2: (7, 7), S3: (31, 48) | S1: (7, 7), S2: (7, 7), S3: (44, 48) | S1: (7, 7), S2: (7, 7), S3: (48, 48) |
| 0.06 | S1: (7, 7), S2: (7, 7), S3: (31, 48) | S1: (7, 7), S2: (7, 7), S3: (43, 48) | S1: (7, 7), S2: (7, 7), S3: (48, 48) |
| 0.07 | S1: (7, 7), S2: (7, 7), S3: (32, 48) | S1: (7, 7), S2: (7, 7), S3: (45, 48) | S1: (7, 7), S2: (7, 7), S3: (48, 48) |
| 0.08 | S1: (7, 7), S2: (7, 7), S3: (33, 48) | S1: (7, 7), S2: (7, 7), S3: (46, 48) | S1: (7, 7), S2: (7, 7), S3: (48, 48) |
| 0.09 | S1: (7, 7), S2: (7, 7), S3: (33, 48) | S1: (7, 7), S2: (7, 7), S3: (47, 48) | S1: (7, 7), S2: (7, 7), S3: (48, 48) |
| 0.1 | S1: (7, 7), S2: (7, 7), S3: (34, 48) | S1: (7, 7), S2: (7, 7), S3: (47, 48) | S1: (7, 7), S2: (7, 7), S3: (48, 48) |