The geometry of efficient codes: How rate-distortion trade-offs distort the latent representations of generative models

doi:10.1371/journal.pcbi.1012952

The geometry of efficient codes: How rate-distortion trade-offs distort the latent representations of generative models

Fig 3

Reconstruction loss, all the models.

For ease of reading, a brief description of each model along with its label is reported in the legend. ( A) Experiment 1. The figure illustrates that increasing capacity reduces reconstruction loss. The trend is similar for the baseline -VAE model that is trained with a balanced dataset and for the two models (E1M1–E1M2) that are trained with unbalanced datasets. ( B) Experiment 2. The figure illustrates that at any given capacity, the baseline model has a smaller reconstruction loss compared to the hybrid models that are additionally trained to solve classification tasks (E2M1–E2M5). Furthermore, at any given capacity, reconstruction loss changes across the different tasks and is worst for the E2M5 model, which addresses four classification tasks simultaneously. See the main text for explanation.

doi: https://doi.org/10.1371/journal.pcbi.1012952.g003