Unsupervised Domain Adaptation via Discriminative Manifold Embedding and Alignment

This is the Pytorch demo code for Unsupervised Domain Adaptation via Discriminative Manifold Embedding and Alignment (DRMEA) (AAAI 2020)

Overview

"DRMEA describes the domains by a sequence of abstract manifolds, and develops a Riemannian manifold learning framework to achieve transferability and discriminability consistently. "

Network Architectures

Experiment Result

ImageCLEF	I→P	P→I	I→C	C→I	C→P	P→C	Avg.
ResNet-50	74.8 ± 0.3	83.9 ± 0.1	91.5 ± 0.3	78.0 ± 0.2	65.5 ± 0.3	91.2 ± 0.3	80.7
DAN	74.5 ± 0.4	82.2 ± 0.2	92.8 ± 0.2	86.3 ± 0.4	69.2 ± 0.4	89.8 ± 0.4	82.5
DANN	75.0 ± 0.3	86.0 ± 0.3	96.2 ± 0.4	87.0 ± 0.5	74.3 ± 0.5	91.5 ± 0.6	85.0
JAN	76.8 ± 0.4	88.0 ± 0.2	94.7 ± 0.2	89.5 ± 0.3	74.2 ± 0.3	91.7 ± 0.3	85.8
CDAN	76.7 ± 0.3	90.6 ± 0.3	97.0 ± 0.4	90.5 ± 0.4	74.5 ± 0.3	93.5 ± 0.4	87.1
CDAN+E	77.7 ± 0.3	90.7 ± 0.2	97.7 ± 0.3	91.3 ± 0.3	74.2 ± 0.2	94.3 ± 0.3	87.7
DRMEA (No AL)	78.0 ± 0.1	91.1 ± 0.1	95.6 ± 0.2	88.7 ± 0.3	74.8 ± 0.1	94.8 ± 0.2	87.3
DRMEA (No DS)	78.9 ± 0.1	90.5 ± 0.2	94.0 ± 0.1	87.8 ± 0.1	76.7 ± 0.2	93.0 ± 0.1	86.8
DRMEA	80.7 ± 0.1	92.5 ± 0.1	97.2 ± 0.1	90.5 ± 0.1	77.7 ± 0.2	96.2 ± 0.2	89.1

Requirements

python 3.6
PyTorch 1.0

Dataset

The dataset should be placed in ./Dataset, e.g.,

./Dataset/ImageCLEF
The structure of the datasets should be like

Image-CLEF (Dataset)
|- I (Domain)
|  |- aeroplane (Class)
|     |- XXXX.jpg (Sample) 
|     |- ...
|  |- bike (Class)
|  |- ...
|- P (Domain)
|- C (Domain)

Usage

Download the Image-CLEF dataset from Google Drive
Training with config

python main.py --dset ImageCLEF --mEpo 50 --ExpTime 10 --BatchSize 32
Experiment results refer to Variables:

ACC_Recorder and Total_Result
Best model and epxerimental logs can be found in folder ./Model_Log/...

Citation

If this repository is helpful for you, please cite our paper:

@inproceedings{luo2020unsupervised,
  title={Unsupervised Domain Adaptation via Discriminative Manifold Embedding and Alignment},
  author={You-Wei Luo, and Chuan-Xian Ren, and Pengfei Ge, and Ke-kun Huang, and Yu-Feng Yu},
  booktitle={AAAI},
  year={2020}
}

Contact

If you have any questions, please feel free contact me via luoyw28@mail2.sysu.edu.cn.

h0ngxuanli/DRMEA