GAMA++: Disentangled Geometric Alignment with Adaptive Contrastive Perturbation for Reliable Domain Transfer
This addresses domain transfer reliability for computer vision tasks, representing an incremental advancement over prior methods like GAMA.
The paper tackled the problem of insufficient disentanglement of task-relevant dimensions and rigid perturbation schemes in geometry-aware domain adaptation, achieving state-of-the-art results on benchmarks like DomainNet, Office-Home, and VisDA with improvements in class-level alignment fidelity and boundary robustness.
Despite progress in geometry-aware domain adaptation, current methods such as GAMA still suffer from two unresolved issues: (1) insufficient disentanglement of task-relevant and task-irrelevant manifold dimensions, and (2) rigid perturbation schemes that ignore per-class alignment asymmetries. To address this, we propose GAMA++, a novel framework that introduces (i) latent space disentanglement to isolate label-consistent manifold directions from nuisance factors, and (ii) an adaptive contrastive perturbation strategy that tailors both on- and off-manifold exploration to class-specific manifold curvature and alignment discrepancy. We further propose a cross-domain contrastive consistency loss that encourages local semantic clusters to align while preserving intra-domain diversity. Our method achieves state-of-the-art results on DomainNet, Office-Home, and VisDA benchmarks under both standard and few-shot settings, with notable improvements in class-level alignment fidelity and boundary robustness. GAMA++ sets a new standard for semantic geometry alignment in transfer learning.