(I am an incoming Ph.D student at Multimedia Lab (MMLAB) in the Chinese University of Hong Kong, working with Prof. Xiangyu Yue.)
I am a third-year master student at Beijing Institute of Technology (BIT), advised by Prof. Shuang Li and Prof. Chi Harold Liu. I received my Bachelor's degree in Software Engineering from BIT (2016 - 2020).
My research interests are in transfer learning and long-tailed learning across classification and object detection. Apart from work, I enjoy running, basketball, movies, badminton.
Email: kaixionggong[at]gmail[dot]com
[Google Scholar] [Github]
News
-
[July 2022] One paper was accepted by ACM Multimedia (ACM MM) 2022!
-
[Apr 2022] Critical Classes and Samples Discovering for Partial Domain Adaptation was accepted to IEEE Transaction on Cybernetics (IF: 11.44)!
-
[July 2021] Pareto Domain Adaptation was accepted to NeurIPS 2021!
-
[Apr 2021] Two papers (One Oral) were accepted to CVPR 2021!
Publications & Preprints
|
Improving Transferability for Domain Adaptive Detection TransformersKaixiong Gong, Shuang Li, Shugang Li, Rui Zhang, Chi Harold Liu, Qiang Chen A novel domain adaptive object detection framework to adapt the DEtection TRansformers (DETR) in both pixel and instance levels across domains. Paper |
|
Metasaug: Meta semantic augmentation for long-tailed visual recognitionShuang Li, Kaixiong Gong, Chi Harold Liu, Yulin Wang, Feng Qiao, Xinjing Cheng A semantic augmentation framework based on meta learning for augmenting the minority classes in long-tailed problems, which learns the augmentation directions automatically. Paper Code |
|
Critical Classes and Samples Discovering for Partial Domain AdaptationIEEE Transaction on Cybernetics (IF: 11.44) Shuang Li, Kaixiong Gong, Binhui Xie, Chi Harold Liu, Weipeng Cao, Song Tian CSDN, which aims to solve PDA problem, identifies the most relevant source classes and critical target samples, such that more precise cross-domain alignment in the shared label space could be enforced by co-training two diverse classifiers. Paper |
|
Transferable semantic augmentation for domain adaptationShuang Li, Mixue Xie, Kaixiong Gong, Chi Harold Liu, Yulin Wang, Wei Li We design TSA to adapt the classifier towards the target domain and alleviate the over-fitting of classifier simultaneously for domain adaptation. Paper Code |
|
Pareto Domain Adaptationfangrui lv, Jian Liang, Kaixiong Gong, Shuang Li, Chi Harold Liu, Han Li, Di Liu, Guoren Wang We develop ParetoDA for handling the encountered conflict when simultaneously optimizing the supervised and alignment loss existing in domain adaptation methods. Paper Code |
|
End-to-End Transferable Anomaly Detection via Multi-spectral Cross-domain Representation AlignmentShuang Li, Shugang Li, Mixue Xie, Kaixiong Gong, Jianxin Zhao, Chi Harold Liu, Guoren Wang A framework for solving anomaly detection in domain adaptation setting, which utilizes multi-spectral representations and adversarial training strategy. Paper |
Invited Talks
-
2021.10, VALSE, Metasaug: Meta semantic augmentation for long-tailed visual recognition.
-
2021.03, BAAI, Metasaug: Meta semantic augmentation for long-tailed visual recognition.
Selected Honors and Awards
-
National Scholarship, Ministry of Education of China (TOP 2%), 2021.
-
Merit Student of Beijing Institute of Technology, 2020.
-
Scholarship of “Huarui Shiji” (Top 5%), 2018.
-
The First Prize Scholarship of Beijing Institute of Technology, 2018.
Contact
-
Email: kaixionggong@gmail.com
-
Address: Room 310, Software building, Beijing Institute of Technology, Beijing