Variational Bayesian Sequence-to-Sequence Networks for Memory-Efficient Sign Language Translation

TL;DR

This paper introduces a variational Bayesian sequence-to-sequence network architecture for memory-efficient sign language translation, combining Gaussian posteriors and Indian Buffet Process priors to achieve significant weight compression without performance loss.

Contribution

It proposes a novel Stick-Breaking Recurrent network that enhances memory efficiency using Bayesian methods and nonparametric priors for sign language translation.

Findings

Achieves substantial weight compression

Maintains modeling performance despite compression

Introduces a new Bayesian recurrent architecture

Abstract

Memory-efficient continuous Sign Language Translation is a significant challenge for the development of assisted technologies with real-time applicability for the deaf. In this work, we introduce a paradigm of designing recurrent deep networks whereby the output of the recurrent layer is derived from appropriate arguments from nonparametric statistics. A novel variational Bayesian sequence-to-sequence network architecture is proposed that consists of a) a full Gaussian posterior distribution for data-driven memory compression and b) a nonparametric Indian Buffet Process prior for regularization applied on the Gated Recurrent Unit non-gate weights. We dub our approach Stick-Breaking Recurrent network and show that it can achieve a substantial weight compression without diminishing modeling performance.