Abstract:Semi-supervised domain adaptation methods leverage information from a source labelled domain with the goal of generalizing over a scarcely labelled target domain. While this setting already poses challenges due to potential distribution shifts between domains, an even more complex scenario arises when source and target data differs in modality representation (e.g. they are acquired by sensors with different characteristics). For instance, in remote sensing, images may be collected via various acquisition modes (e.g. optical or radar), different spectral characteristics (e.g. RGB or multi-spectral) and spatial resolutions. Such a setting is denoted as Semi-Supervised Heterogeneous Domain Adaptation (SSHDA) and it exhibits an even more severe distribution shift due to modality heterogeneity across <a class="link-external link-http" href="http://domains.To" rel="external noopener nofollow">this http URL</a> cope with the challenging SSHDA setting, here we introduce SHeDD (Semi-supervised Heterogeneous Domain Adaptation via Disentanglement) an end-to-end neural framework tailored to learning a target domain classifier by leveraging both labelled and unlabelled data from heterogeneous data sources. SHeDD is designed to effectively disentangle domain-invariant representations, relevant for the downstream task, from domain-specific information, that can hinder the cross-modality transfer. Additionally, SHeDD adopts an augmentation-based consistency regularization mechanism that takes advantages of reliable pseudo-labels on the unlabelled target samples to further boost its generalization ability on the target domain. Empirical evaluations on two remote sensing benchmarks, encompassing heterogeneous data in terms of acquisition modes and spectral/spatial resolutions, demonstrate the quality of SHeDD compared to both baseline and state-of-the-art competing approaches. Our code is publicly available here: <a class="link-external link-https" href="https://github.com/tanodino/SSHDA/" rel="external noopener nofollow">this https URL</a>

Heterogeneous domain adaptation by semantic distribution alignment network

HOMDA: High-Order Moment-Based Domain Alignment for Unsupervised Domain Adaptation

Simultaneous Semantic Alignment Network for Heterogeneous Domain Adaptation

Heterogeneous domain adaptation by class centroid matching and local discriminative structure preservation

Semi-supervised Heterogeneous Domain Adaptation: Theory and Algorithms

Heterogeneous Domain Adaptation by Information Capturing and Distribution Matching.

Heterogeneous Domain Adaptation Through Progressive Alignment.

Learning with Augmented Features for Supervised and Semi-Supervised Heterogeneous Domain Adaptation

Heterogeneous Domain Adaptation With Generalized Similarity and Dissimilarity Regularization

Multisource Heterogeneous Domain Adaptation With Conditional Weighting Adversarial Network

Graph Embedding-Based Heterogeneous Domain Adaptation with Domain-Invariant Feature Learning and Distributional Order Preserving.

Multi-source Heterogeneous Domain Adaptation with Conditional Weighting Adversarial Network

Heuristic Domain Adaptation

Semantic Distribution-aware Contrastive Adaptation for Semantic Segmentation

Semi-Supervised Domain Adaptation via Adaptive and Progressive Feature Alignment

Semi Supervised Heterogeneous Domain Adaptation via Disentanglement and Pseudo-Labelling

Heterogeneous Domain Adaptation Network Based on Autoencoder

Domain Adaptive Classification on Heterogeneous Information Networks

Knowledge Preserving and Distribution Alignment for Heterogeneous Domain Adaptation

Learning with Augmented Features for Heterogeneous Domain Adaptation

Semantic Correlation Transfer for Heterogeneous Domain Adaptation