Towards Unsupervised Open-Set Graph Domain Adaptation via Dual Reprogramming

TL;DR

This paper proposes a novel framework called GraphRTA, which conducts reprogramming on both the graph and model sides, and extends the classifier with an extra dimension for the unknown class, thus eliminating the need of manually specified threshold in open-set recognition.

Abstract

Unsupervised Graph Domain Adaptation has become a promising paradigm for transferring knowledge from a fully labeled source graph to an unlabeled target graph. Existing graph domain adaptation models primarily focus on the closed-set setting, where the source and target domains share the same label spaces. However, this assumption might not be practical in the real-world scenarios, as the target domain might include classes that are not present in the source domain. In this paper, we investigate the problem of unsupervised open-set graph domain adaptation, where the goal is to not only correctly classify target nodes into the known classes, but also recognize previously unseen node types into the unknown class. Towards this end, we propose a novel framework called GraphRTA, which conducts reprogramming on both the graph and model sides. Specifically, we reprogram the graph by modifying target graph structure and node features, which facilitates better separation of known and unknown classes. Meanwhile, we also perform model reprogramming by pruning domain-specific parameters to reduce bias towards the source graph while preserving parameters that capture transferable patterns across graphs. Additionally, we extend the classifier with an extra dimension for the unknown class, thus eliminating the need of manually specified threshold in open-set recognition. Comprehensive experiments on several public datasets demonstrate that our proposed model can achieve satisfied performance compared with recent state-of-the-art baselines. Our source codes and datasets are publicly available at https://github.com/cszhangzhen/GraphRTA.

Figures (4)

• Figure 1: The pipeline of the proposed GraphRAT model. In the model reprogramming module, some nodes and connections are faded, indicating that these weights have been pruned during the reprogramming process. In the graph reprogramming module, node features are updated, and edges are dynamically modified (deleted or added), as indicated by dashed lines. Additionally, domain adversarial learning is incorporated to categorize instances into three distinct groups: source, target-known, and target-unknown, thereby enhancing the model's generalization capacity.
• Figure 2: Visualization of node representations in the target graph for the citation dataset (A$\rightarrow$C) with the unknown class highlighted in orange.
• Figure 3: H-scores under different ratios in reprogramming.
• Figure 4: Accuracy under different layers and dimension.