Flexible deep transfer learning by separate feature embeddings and manifold alignment

TL;DR

This paper introduces a deep transfer learning framework that learns separate feature embeddings for different domains and aligns them via manifold and adversarial training, enabling robust transfer across significant domain shifts.

Contribution

It proposes a novel deep learning approach that supports domain changes by learning separate features and aligning them through manifold and adversarial methods.

Findings

Effective on synthetic, measured, and satellite datasets

Outperforms traditional domain adaptation methods

Provides training guidelines for adversarial networks

Abstract

Object recognition is a key enabler across industry and defense. As technology changes, algorithms must keep pace with new requirements and data. New modalities and higher resolution sensors should allow for increased algorithm robustness. Unfortunately, algorithms trained on existing labeled datasets do not directly generalize to new data because the data distributions do not match. Transfer learning (TL) or domain adaptation (DA) methods have established the groundwork for transferring knowledge from existing labeled source data to new unlabeled target datasets. However, current DA approaches assume similar source and target feature spaces and suffer in the case of massive domain shifts or changes in the feature space. Existing methods assume the data are either the same modality, or can be aligned to a common feature space. Therefore, most methods are not designed to support a fundamental domain change such as visual to auditory data. We propose a novel deep learning framework that overcomes this limitation by learning separate feature extractions for each domain while minimizing the distance between the domains in a latent lower-dimensional space. The alignment is achieved by considering the data manifold along with an adversarial training procedure. We demonstrate the effectiveness of the approach versus traditional methods with several ablation experiments on synthetic, measured, and satellite image datasets. We also provide practical guidelines for training the network while overcoming vanishing gradients which inhibit learning in some adversarial training settings.