SPA++: Generalized Graph Spectral Alignment for Versatile Domain Adaptation

2025-08-07

AI Summary

Addressing the challenge of simultaneously achieving cross-domain transferability and intra-domain structural preservation in domain adaptation, this paper proposes a generalized graph alignment framework. It constructs source and target domain graphs in the feature space and aligns them via spectral regularization, complemented by a neighborhood-aware fine-grained feature propagation strategy. To further enhance robustness, the method integrates data augmentation with consistency regularization. By unifying cross-domain alignment and intra-domain discriminative structure modeling, it significantly improves the discriminability of target-domain representations. Extensive experiments on multiple standard benchmarks demonstrate state-of-the-art performance, particularly under challenging conditions such as large source–target distribution shifts and sparse labeling. The framework achieves consistent and substantial gains over existing methods, validating its effectiveness in preserving both structural fidelity and transferable semantics across domains.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Adversarial Attacks & Robustness
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Domain Adaptation (DA) aims to transfer knowledge from a labeled source domain to an unlabeled or sparsely labeled target domain under domain shifts. Most prior works focus on capturing the inter-domain transferability but largely overlook rich intra-domain structures, which empirically results in even worse discriminability. To tackle this tradeoff, we propose a generalized graph SPectral Alignment framework, SPA++. Its core is briefly condensed as follows: (1)-by casting the DA problem to graph primitives, it composes a coarse graph alignment mechanism with a novel spectral regularizer toward aligning the domain graphs in eigenspaces; (2)-we further develop a fine-grained neighbor-aware propagation mechanism for enhanced discriminability in the target domain; (3)-by incorporating data augmentation and consistency regularization, SPA++ can adapt to complex scenarios including most DA settings and even challenging distribution scenarios. Furthermore, we also provide theoretical analysis to support our method, including the generalization bound of graph-based DA and the role of spectral alignment and smoothing consistency. Extensive experiments on benchmark datasets demonstrate that SPA++ consistently outperforms existing cutting-edge methods, achieving superior robustness and adaptability across various challenging adaptation scenarios.

Problem

• Balancing inter-domain transferability and intra-domain discriminability in Domain Adaptation
• Aligning domain graphs in eigenspaces using spectral regularization
• Enhancing target domain discriminability with neighbor-aware propagation

Innovation

• Graph spectral alignment with novel regularizer
• Neighbor-aware propagation for discriminability
• Data augmentation with consistency regularization