XSR-OSDA

:fire: Implementation for the ''Interpretable Novel Target Discovery Through Open-Set Domain Adaptation (XSR-OSDA)'' work (under review).

XSR-OSDA is an extension work of the "SR-OSDA" paper published in ICCV 2021 [paper][Github].

Data Preparation

---

• I->AwA: 3D2 and AwA2
• DomainNet -> AwA: DomainNet & AwA2
• DomainNet -> LAD: DomainNet & LAD

Dataset	Domain	Role	#Images	#Attributes	#Classes
DomainNet $\rightarrow$ AwA2	AwA Paint Real	S / T	9,343 / 15,306 3,441 / 5,760 5,251 / 10,047	85	10 / 17
I $\rightarrow$ AwA	I / AwA	S / T	2,970 / 37,322	85	40 / 50
Domain $\rightarrow$ LAD	LAD Paint Real	S / T	13,322 / 19,744 11,714 / 15,311 22,395 / 31,066	253	40 / 56

Dependencies

---

• Python 3.8
• Pytorch 1.10

Training

---

python main.py

Evaluation

---

• Open-set Domain Adaptation Task

$OS^*$: class-wise average accuracy on the seen categories.

$OS^\diamond$: class-wise average accuracy on the unseen categories correctly classified as "unknown".

$OS$: $\frac{OS^* \times C_{shr} + OS^\diamond}{C_{shr} + 1}$

$OS^{H}$: $\frac{ 2 \times OS^* \times OS^\diamond}{OS^* + OS^\diamond}$

• Semantic-Recovery Open-Set Domain Adaptation Task

$S$: class-wise average accuracy on shared classes

$U$: class-wise average accuracy on unknown classes

$H = \frac{2 \times S \times U}{ S + U}$

Citation

---

If you think this work is interesting, please cite:

@InProceedings{Jing_2021_XSROSDA,
author = {Jing, Taotao and Xia, Haifeng and Liu, Hongfu and Ding, Zhengming},
title = {Interpretable Novel Target Discovery Through Open-Set Domain Adaptation},
booktitle = {},
year = {}
}

Contact

---

If you have any questions about this work, feel free to contact

• tjing@tulane.edu