Dual Contradistinctive Generative Autoencoder

Gaurav Parmar, Dacheng Li, Kwonjoon Lee, Zhuowen Tu

Abstract: We present a new generative autoencoder model with dual contradistinctive losses to improve generative autoencoder that performs simultaneous inference (reconstruction) and synthesis (generation). We name our model dual contradistinctive generative autoencoder (DC-VAE) that integrates an instance-level discriminative loss (maintaining the instance-level fidelity for the reconstruction/synthesis) with a set-level adversarial loss (encouraging the set-level fidelity for the reconstruction/synthesis), both being contradistinctive. There also exists a mathematical connection between the instance-based classification and instance-level conditional distribution. DC-VAE achieves competitive results in three tasks, including image synthesis, image reconstruction, and representation learning. DC-VAE is applicable to various tasks in computer vision and machine learning.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Reviewed Version (pdf): https://openreview.net/references/pdf?id=NDlnyQVaU