Enhancing Domain Adaptation through Prompt Gradient Alignment
About
Prior Unsupervised Domain Adaptation (UDA) methods often aim to train a domain-invariant feature extractor, which may hinder the model from learning sufficiently discriminative features. To tackle this, a line of works based on prompt learning leverages the power of large-scale pre-trained vision-language models to learn both domain-invariant and specific features through a set of domain-agnostic and domain-specific learnable prompts. Those studies typically enforce invariant constraints on representation, output, or prompt space to learn such prompts. In contrast, we cast UDA as a multiple-objective optimization problem in which each objective is represented by a domain loss. Under this new framework, we propose to align per-objective gradients to foster consensus between them. Additionally, to prevent potential overfitting when fine-tuning this deep learning architecture, we penalize the norm of these gradients. To achieve these goals, we devise a practical gradient update procedure that can work under both single-source and multi-source UDA. Empirically, our method consistently outperforms other vision-language model adaptation methods. The implementation is available at https://github.com/VietHoang1512/PGA.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Unsupervised Domain Adaptation | Office-Home (test) | Average Accuracy73.9 | 332 | |
| Unsupervised Domain Adaptation | Office-Home | Average Accuracy79.4 | 238 | |
| Image Classification | DomainNet (test) | Average Accuracy55.4 | 209 | |
| Image Classification | OfficeHome | Average Accuracy88.4 | 131 | |
| Domain Adaptation | Office-Home (test) | Mean Accuracy89.4 | 112 | |
| Unsupervised Domain Adaptation | DomainNet | Average Accuracy56.2 | 100 | |
| Unsupervised Domain Adaptation | Office-Home 101 (test) | Accuracy (Ar→Cl)56.1 | 17 | |
| Image Classification | ImageCLEF | F1 Score94.2 | 14 | |
| Unsupervised Domain Adaptation | ImageCLEF | Accuracy (Domain C)97.4 | 12 | |
| Image Classification | S2RDA-49 synthetic-to-real (test) | Accuracy74.1 | 9 |