Xialei Liu 刘夏雷

Xialei Liu is a Professor at Nankai University. Before that, He was a postdoctoral researcher at the University of Edinburgh working with Prof. Hakan Bilen. He received his Master’s and PhD degrees from the Autonomous University of Barcelona, supervised by Prof. Joost van de Weijer and Prof. Andrew D. Bagdanov. He received a second Master’s degree and a Bachelor’s degree at Northwestern Polytechnical University. He works in the field of computer vision and machine learning. His research interests include multi-modal learning, lifelong learning, self-supervised learning, few-shot learning, object detection, etc.

Google scholar, Linkedin, ResearchGate, Github.

刘夏雷，南开大学计算机学院，教授，博导，国家级青年人才计划入选者。入选南开大学“百名青年学科带头人培养计划”，中国科协青年托举计划。在英国爱丁堡大学从事博士后研究，合作导师： Prof. Hakan Bilen。获得西班牙巴塞罗那自治大学计算机科学博士学位，博士导师：Prof. Joost van de Weijer 和 Prof. Andrew D. Bagdanov。本科、硕士毕业于西北工业大学自动化学院，硕士导师程咏梅教授。主要研究领域为计算机视觉和模式识别，特别是开放环境下机器学习及应用，主要研究方向包括多模态学习、连续学习、小样本学习、自监督学习、目标检测等。

News

招2027年入学硕士、博士若干名
招实习生（本校本科生）：主题包括多模态学习、Agent、具身智能、目标检测、视觉缺陷检测、连续学习
招博士后：包括海外博士后引才计划
2026届毕业生去向：互联网头部企业X2（大模型方向）、初创公司（具身智能方向）、美国名校博士
1 paper accepted at ECCV 2026
1 paper accepted at CVPR 2026
2 papers accepted at NeurIPS 2025
1 paper accepted at ICCV 2025
邀请报告VALSE Webinar：基于图文预训练模型的连续学习方法研究 [视频报告链接] [Panel讨论环节]
2 papers accepted at CVPR 2025
2 papers accepted at ECCV 2024
2 papers accepted at CVPR 2024
第二届粤港澳大湾区(黄埔)国际算法算例大赛“序列任务的持续学习” 冠军（奖金40万）
3 papers accepted at ICCV 2023
1 paper accepted at IJCV, 2023.07
邀请报告：CCIG2023知识引导的自适应感知与结构理解技术论坛（知识引导的连续学习方法探究）
1 paper accepted at CVPR 2023 (Continual Semantic Segmentation)
Served as CVPR 2023 4th CLVISION Workshop organizer (Continual Learning)
Invited talk at ICCPR 2022 (link)
1 paper accepted at TPAMI 2022.10 (arXiv: Survey on Class-incremental Learning, IEEE TPAMI online)
1 paper accepted at ECCV 2022.10 (Long-tailed Class-incremental Learning)
Teaching： Computer Vision 计算机视觉 (研究生)- 2022 秋
Teaching： Global Outlook and Personal Development 国际视野与个人成长 (本科生)- 2022 秋（英）
CSIG交通视频专委会连续学习前沿论坛 (论坛链接，直播回放)
1 paper selected as CVPR 2022 Best Paper Finalists (Multi-task Learning)
VALSE 2022.08 Tutorial Invited Talk (连续学习 / PPT链接 [提取码：e93u])
3 papers accepted at CVPR 2022.06 (Continual learning on Segmentation, Few-shot, Multi-task Learning)
1 paper accepted at TNNLS, 2022.02 (Continual learning on Segmentation)
VALSE 2022.08 组委会成员注册主席 (Registration Chair)
VALSE 2022.08 Workshop 组织者成员 (开放环境下机器学习及应用)
Teaching： Artificial Intelligence 人工智能导论 (本科生)- 2022 春

Recent Publications

Notes: Joint first authors are indicated using # and corresponding authors are indicated using *.

Mind the Gap: Preserving and Compensating for the Modality Gap in CLIP-Based Continual Learning, ICCV 2025 (Highlight). [Paper][Code][中译版]
KAC: Kolmogorov-Arnold Classifier for Continual Learning, Y Hu, Z Liang, F Yang, Q Hou, X Liu*, MM Cheng, CVPR 2025 (Highlight). [Paper][Code][中译版]
GET: Unlocking the Multi-modal Potential of CLIP for Generalized Category Discovery, E Wang, Z Peng, Z Xie, F Yang, X Liu, MM Cheng, CVPR 2025. [Paper][Code][中译版]
Class Incremental Learning for Image Classification with Out-of-distribution Task Identification, X Cao, H Lu, X Liu*, MM Cheng, TMM 2025. [Paper][中译版]
Reformulating classification as image-class matching for class incremental learning, TCSVT 2025. [Paper]
See More, Know More: Richer Prior Knowledge for Novel Class Discovery, Z Peng, E Wang, X Liu*, MM Cheng, CVMJ 2024. [Paper][Code][中译版]
Early Preparation Pays Off: New Classifier Pre-tuning for Class Incremental Semantic Segmentation, Z Xie, H Lu, J Xiao, E Wang, L Zhang, X Liu^*, ECCV 2024. [Paper][Code][中译版]
Class-Incremental Learning with CLIP: Adaptive Representation Adjustment and Parameter Fusion, L Huang, X Cao, H Lu, X Liu^*, ECCV 2024. [Paper][Code][中译版]
Generative Multi-modal Models are Good Class-Incremental Learners, X Cao, H Lu, L Huang, X Liu^*, MM Cheng, CVPR 2024. [Paper][Code][中译版]
Task-Adaptive Saliency Guidance for Exemplar-free Class Incremental Learning, X Liu^#, JT Zhai^#, AD Bagdanov, K Li, MM Cheng^*, CVPR 2024. [Paper][Code][中译版]
Fine-Grained Knowledge Selection and Restoration for Non-Exemplar Class Incremental Learning, JT Zhai, X Liu^*, L Yu, MM Cheng, AAAI 2024. [Paper][Code][中译版]
Masked Autoencoders are Efficient Class Incremental Learners, JT Zhai, X Liu^*, AD Bagdanov, K Li, MM Cheng, ICCV 2023. [Paper][Code][中译版]
Lighting Every Darkness in Two Pairs: A Calibration-Free Pipeline for RAW Denoising, X Jin^#, JW Xiao^#, LH Han, C Guo, R Zhang, X Liu, C Li, ICCV 2023.
Augmented Box Replay: Overcoming Foreground Shift for Incremental Object Detection, Y Liu, Y Cong, D Goswami, X Liu, J Van de Weijer, ICCV 2023.
Endpoints Weight Fusion for Class Incremental Semantic Segmentation, JW Xiao^#, CB Zhang^#, J Feng, X Liu^*, J van de Weijer, MM Cheng, CVPR 2023
Universal Representations: A Unified Look at Multiple Task and Domain Learning, WH Li, X Liu, H Bilen, IJCV, 2023.
Class-incremental Learning: Survey and Performance Evaluation, M Masana, X Liu*, B Twardowski, M Menta, et al., IEEE TPAMI, 2022.
Positive Pair Distillation Considered Harmful: Continual Meta Metric Learning for Lifelong Object Re-Identification, K Wang^#, C Wu^#, AD Bagdanov, X Liu^*, et al., BMVC 2022.
Long-Tailed Class Incremental Learning, X Liu^*^,#, YS Hu^#, XS Cao, AD Bagdanov, K Li, MM Cheng, ECCV 2022.
Representation Compensation Networks for Continual Semantic Segmentation, CB Zhang, JW Xiao, X Liu^*, YC Chen, MM Cheng, IEEE CVPR 2022.
Improving Task Adaptation for Cross-domain Few-shot Learning, WH Li, X Liu^*, H Bilen, IEEE CVPR, 2022.
Learning Multiple Dense Prediction Tasks from Partially Annotated Data, WH Li, X Liu, H Bilen, IEEE CVPR, 2022.
Self-Training for Class-Incremental Semantic Segmentation, L Yu, X Liu^*, J van de Weijer, TNNLS, 2022.
HCV: Hierarchy-Consistency Verification for Incremental Implicitly-Refined Classification, K Wang, X Liu^*, L Herranz, J Van de Weijer, BMVC, 2021.
Universal Representation Learning from Multiple Domains for Few-shot Classification, WH Li, X Liu^*, H Bilen, IEEE ICCV, 2021.
Semantic Drift Compensation for Class-Incremental Learning, L Yu, B Twardowski, X Liu, et al., IEEE CVPR, 2020.
Exploiting Unlabeled Data in CNNs by Self-supervised Learning to Rank, X Liu, J van de Weijer, AD Bagdanov, IEEE TPAMI, 2019.
Learning Metrics from Teachers: Compact Networks for Image Embedding, L Yu, VO Yazici, X Liu, J van de Weijer, et al., IEEE CVPR, 2019.
Memory Replay GANs: Learning to Generate Images from New Categories without Forgetting, C Wu, L Herranz, X Liu, et al., NeurIPS, 2018.
Rotate your Networks: Better Weight Consolidation and Less Catastrophic Forgetting, X Liu^#, M Masana^#, L Herranz, et al., ICPR, 2018.
Leveraging Unlabeled Data for Crowd Counting by Learning to Rank, X Liu, J van de Weijer, AD Bagdanov, IEEE CVPR, 2018.
Rankiqa: Learning from Rankings for No-reference Image Quality Assessment, X Liu, J van de Weijer, AD Bagdanov, IEEE ICCV, 2017.

Useful links

Contact

Email: xialei AT nankai DOT edu DOT cn
Office: 450, College of Computer Science, Nankai University