Towards a General Model of Knowledge for Facial Analysis by Multi-Source Transfer Learning

TL;DR

This paper introduces a multi-source transfer learning framework for facial analysis that combines models into a common embedding and distills this into a lightweight, versatile student model that outperforms its teachers on various tasks.

Contribution

It presents a novel two-step approach using auto-encoder fusion and model distillation to create a compact, generalizable facial analysis model from multiple sources.

Findings

The student model outperforms its teacher on new facial analysis tasks.

The approach achieves state-of-the-art results on 15 facial analysis tasks.

The student model has 75 times fewer parameters than the teacher.

Abstract

This paper proposes a step toward obtaining general models of knowledge for facial analysis, by addressing the question of multi-source transfer learning. More precisely, the proposed approach consists in two successive training steps: the first one consists in applying a combination operator to define a common embedding for the multiple sources materialized by different existing trained models. The proposed operator relies on an auto-encoder, trained on a large dataset, efficient both in terms of compression ratio and transfer learning performance. In a second step we exploit a distillation approach to obtain a lightweight student model mimicking the collection of the fused existing models. This model outperforms its teacher on novel tasks, achieving results on par with state-of-the-art methods on 15 facial analysis tasks (and domains), at an affordable training cost. Moreover, this student has 75 times less parameters than the original teacher and can be applied to a variety of novel face-related tasks.