Addressing Imbalanced Domain-Incremental Learning through Dual-Balance Collaborative Experts

Lan Li  Da-Wei Zhou  Han-Jia Ye  De-Chuan Zhan 

State Key Laboratory for Novel Software Technology, Nanjing University

The code repository for "Addressing Imbalanced Domain-Incremental Learning through Dual-Balance Collaborative Experts" in PyTorch. If you use any content of this repo for your work, please cite the following bib entry:

@article{li2025addressing,
  title={Addressing Imbalanced Domain-Incremental Learning through Dual-Balance Collaborative Experts},
  author={Lan Li and Da-Wei Zhou and Han-Jia Ye and De-Chuan Zhan},
  journal={ICML},
  year={2025}
}

📢 Updates

[07/2025] Code has been released.

[07/2025] arXiv paper has been released.

📝 Introduction

Domain-Incremental Learning (DIL) focuses on continual learning in non-stationary environments, requiring models to adjust to evolving domains while preserving historical knowledge. DIL faces two critical challenges in the context of imbalanced data: intra-domain class imbalance and cross-domain class distribution shifts. These challenges significantly hinder model performance, as intra-domain imbalance leads to underfitting of few-shot classes, while cross-domain shifts require maintaining well-learned many-shot classes and transferring knowledge to improve few-shot class performance in old domains. To overcome these challenges, we introduce the Dual-Balance Collaborative Experts (DCE) framework. DCE employs a frequency-aware expert group, where each expert is guided by specialized loss functions to learn features for specific frequency groups, effectively addressing intra-domain class imbalance. Subsequently, a dynamic expert selector is learned by synthesizing pseudo-features through balanced Gaussian sampling from historical class statistics. This mechanism navigates the trade-off between preserving many-shot knowledge of previous domains and leveraging new data to improve few-shot class performance in earlier tasks. Extensive experimental results on four benchmark datasets demonstrate DCE’s state-of-the-art performance.

🔧 Requirements

Environment

You can create a conda environment and run the following command to install the dependencies.

conda install --file requirements.txt

Dataset

There are 4 datasets involved in the paper, CDDB, CORe50, DomainNet and Office-Home respectively. Follow the two-step guideline to prepare them for the reproduction.

1. Download the datasets mannually according the recommended.

   • CDDB: You can access the dataset at this link.
   • CORe50: You can access the dataset at this link. Please download the core50_imgs.npz file.
   • DomainNet: You can access the dataset at this link. Please download the cleaned version file.
   • Office-Home: You can access the dataset at this link.

2. Check if the dataset has been downloaded properly. The dataset directory is expected to have the following structure:

   CDDB
   ├── biggan
   ├── crn 
   ├── ...
   ├── wild
   
   CORe50
   ├── s1
   ├── ...
   ├── s11
   ├── labels.pkl
   ├── LUP.pkl
   ├── paths.pkl
   
   DomainNet
   ├── clipart
   ├── ...
   ├── sketch
   ├── clipart_test.txt
   ├── clipart_train.txt
   ├── ...
   
   OfficeHome
   ├── Art
   ├── Clipart
   ├── Product
   ├── Real_World
   
   

3. Specify the dataset path in the config file under the configs directory.

   "data_path": "$Your/Dataset/Path/Here$"
   

💡 Running scripts

To prepare your JSON files, refer to the settings in the Configs folder and run the following command. All main experiments from the paper are already provided in the exps folder, you can simply execute them to reproduce the results found in the logs folder.

python main.py --config configs/officehome.json --order 1

🎈 Acknowledgement

This repo is based on CIL_Survey and PyCIL.

💭 Correspondence

If you have any questions, please contact me via email or open an issue.