Domain Generalization using Pretrained Models without Fine-tuning

9 Mar 2022  ·  Ziyue Li, Kan Ren, Xinyang Jiang, Bo Li, Haipeng Zhang, Dongsheng Li ·

Fine-tuning pretrained models is a common practice in domain generalization (DG) tasks. However, fine-tuning is usually computationally expensive due to the ever-growing size of pretrained models. More importantly, it may cause over-fitting on source domain and compromise their generalization ability as shown in recent works. Generally, pretrained models possess some level of generalization ability and can achieve decent performance regarding specific domains and samples. However, the generalization performance of pretrained models could vary significantly over different test domains even samples, which raises challenges for us to best leverage pretrained models in DG tasks. In this paper, we propose a novel domain generalization paradigm to better leverage various pretrained models, named specialized ensemble learning for domain generalization (SEDGE). It first trains a linear label space adapter upon fixed pretrained models, which transforms the outputs of the pretrained model to the label space of the target domain. Then, an ensemble network aware of model specialty is proposed to dynamically dispatch proper pretrained models to predict each test sample. Experimental studies on several benchmarks show that SEDGE achieves significant performance improvements comparing to strong baselines including state-of-the-art method in DG tasks and reduces the trainable parameters by ~99% and the training time by ~99.5%.

PDF Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Domain Generalization DomainNet SEDGE Average Accuracy 46.3 # 12
Domain Generalization DomainNet SEDGE+ Average Accuracy 54.7 # 3
Domain Generalization Office-Home SEDGE+ Average Accuracy 80.7 # 3
Domain Generalization Office-Home SEDGE Average Accuracy 79.9 # 5
Domain Generalization PACS SEDGE+ Average Accuracy 96.1 # 2
Domain Generalization PACS SEDGE Average Accuracy 84.1 # 32
Domain Generalization TerraIncognita SEDGE Average Accuracy 56.8 # 3
Domain Generalization TerraIncognita SEDGE+ Average Accuracy 56.8 # 3
Domain Generalization VLCS SEDGE Average Accuracy 79.8 # 6
Domain Generalization VLCS SEDGE+ Average Accuracy 82.2 # 1

Methods