Variational learning across domains with triplet information

TL;DR

This paper introduces VBTA, a deep generative model that leverages triplet-based metric learning to effectively transfer knowledge across domains, demonstrated on image translation, generation, and cross-lingual classification.

Contribution

The paper proposes the Variational Bi-domain Triplet Autoencoder (VBTA), integrating triplet constraints into deep generative models for improved cross-domain learning.

Findings

Effective in image-to-image translation

Improves bi-directional image generation

Enhances cross-lingual document classification

Abstract

The work investigates deep generative models, which allow us to use training data from one domain to build a model for another domain. We propose the Variational Bi-domain Triplet Autoencoder (VBTA) that learns a joint distribution of objects from different domains. We extend the VBTAs objective function by the relative constraints or triplets that sampled from the shared latent space across domains. In other words, we combine the deep generative models with a metric learning ideas in order to improve the final objective with the triplets information. The performance of the VBTA model is demonstrated on different tasks: image-to-image translation, bi-directional image generation and cross-lingual document classification.