Student at The Institute of Information Engineering, School of Cyber Security at University of Chinese Academy of Sciences, China
Kexin Bao is a focused and innovative researcher currently pursuing her Ph.D. at the Institute of Information Engineering, School of Cyber Security, University of Chinese Academy of Sciences. Her research primarily revolves around machine learning and computer vision, with specialization in few-shot class-incremental learning and weakly supervised small object detection. Through her contributions, she aims to address the challenges of enabling AI models to learn efficiently with minimal data and annotations. Kexin has actively participated in six research projects and authored six peer-reviewed SCI/Scopus-indexed journal publications, with a total citation count of 62. Her work includes the design of the Prior Knowledge-Infused Neural Network (PKI), which balances performance and computational efficiency. She collaborates with esteemed researchers like Shiming Ge and continues to demonstrate a high level of commitment to innovation and scholarly excellence. Kexin Bao’s work holds promise for practical applications in AI and has the potential to impact academia and industry alike.
Professional Profile
Scopus Profile | ORCID Profile
Education
Kexin Bao is currently pursuing her Doctor of Philosophy (Ph.D.) in Cyber Security and Information Engineering at the prestigious University of Chinese Academy of Sciences. She is enrolled at the Institute of Information Engineering, which is known for its excellence in cutting-edge research in computer science and cybersecurity. Her academic focus lies in advanced topics within machine learning and computer vision, particularly in areas such as few-shot learning, incremental learning, and object detection. Prior to her Ph.D., Kexin likely completed a Bachelor’s and Master’s degree in a relevant field, which laid the foundation for her research career, though those details are not explicitly mentioned in her profile. Her academic training has equipped her with the theoretical knowledge and practical skills needed to tackle complex real-world problems in artificial intelligence. Her ongoing doctoral studies not only refine her technical abilities but also enable her to contribute meaningfully to the global research community.
Professional Experience
As a Ph.D. student, Kexin Bao’s professional experience is rooted in academic research, with a strong focus on machine learning and computer vision. Although she does not yet have experience in industry or consultancy projects, she has participated in six significant research initiatives that address challenges in artificial intelligence, particularly in data-efficient learning models. Her work involves both independent and collaborative research, including partnerships with renowned scholars like Shiming Ge, Daichi Zhang, and Fanzhao Lin. While still in the early stages of her professional career, she has already contributed to six SCI/Scopus-indexed publications and one patent submission, reflecting her active role in advancing knowledge and technology. Though she has not yet undertaken formal leadership roles or teaching positions, her ability to carry out complex research projects demonstrates a high level of professionalism and expertise. Her growing research profile suggests that she is well-positioned to transition into impactful academic or industry roles in the future.
Research Interest
Kexin Bao’s research interests lie at the intersection of machine learning, computer vision, and artificial intelligence, with a specific focus on Few-Shot Class-Incremental Learning (FSCIL) and Weakly Supervised Small Object Detection. She is deeply interested in developing intelligent systems that can learn continuously from limited data, which is crucial for real-world applications where large annotated datasets are often unavailable. Her work on the Prior Knowledge-Infused Neural Network (PKI) and its variants (PKIV-1, PKIV-2) demonstrates her commitment to enhancing learning efficiency and minimizing resource consumption. She aims to create models that not only generalize well but also adapt quickly to new tasks with minimal retraining. These interests align closely with future directions in sustainable AI, autonomous systems, and edge computing. Kexin continues to explore methods that combine theoretical advancements with practical deployment possibilities, aiming to bridge the gap between academic research and real-world applications in intelligent automation and perception systems.
Award and Honor
Though early in her academic journey, Kexin Bao has already achieved commendable recognition through her contributions to research in computer vision. She has authored six peer-reviewed journal publications indexed in SCI and Scopus, and her work has been cited 62 times, indicating growing academic impact. Additionally, she has filed one patent based on her original research, a significant milestone for any early-career researcher. These achievements reflect both innovation and practical relevance in her work. She has also collaborated with prominent researchers, which further adds to her credibility and visibility in the research community. While she has not yet received named awards or honors beyond her publication and patent successes, her nomination for the Best Researcher Award is itself a testament to her academic excellence, research contribution, and future potential. With continued progress, she is well-positioned to receive further accolades and recognition at national and international levels in the near future.
Research Skill
Kexin Bao possesses a robust set of research skills that span both theoretical understanding and practical implementation in machine learning and computer vision. She is proficient in developing deep learning models and has a strong command of techniques related to few-shot learning, incremental learning, and weak supervision. Her work demonstrates advanced capabilities in model optimization, neural network design, and experimental benchmarking. Kexin has conducted extensive experiments on recognized datasets, validating her models through comparisons with state-of-the-art techniques. She is adept at using research tools, coding in frameworks such as PyTorch or TensorFlow, and performing data preprocessing and analysis. Her development of the Prior Knowledge-Infused Neural Network and its variants highlights her problem-solving ability and innovation mindset. She is also skilled in academic writing, contributing to multiple peer-reviewed journals. These research skills, combined with her ability to work collaboratively and manage projects independently, position her as a capable and resourceful young researcher.
Publications Top Notes
Title: DB-FSCIL: Few-Shot Class-Incremental Learning Using Dual Bridges
Authors: Kexin Bao, Fanzhao Lin, Ruyue Liu, Shiming Ge
Year: 2025
Type: Book Chapter
DOI: 10.1007/978-981-96-0122-6_7
Title: PKI: Prior Knowledge-Infused Neural Network for Few-Shot Class-Incremental Learning
Authors: Kexin Bao, Fanzhao Lin, Zichen Wang, Yong Li, Dan Zeng, Shiming Ge
Year: 2025 (Expected December)
Type: Journal Article (Neural Networks)
DOI: 10.1016/j.neunet.2025.107724
Title: Divide and Conquer: Static-Dynamic Collaboration for Few-Shot Class-Incremental Learning
Authors: Kexin Bao, Daichi Zhang, Yong Li, Dan Zeng, Shiming Ge
Year: 2025
Type: Conference Paper
DOI: 10.1145/3731715.3733310
Title: Learning Contrast-Enhanced Shape-Biased Representations for Infrared Small Target Detection
Authors: Fanzhao Lin, Kexin Bao, Yong Li, Dan Zeng, Shiming Ge
Year: 2024
Type: Journal Article (IEEE Transactions on Image Processing)
DOI: 10.1109/TIP.2024.3391011
Title: Learning Shape-Biased Representations for Infrared Small Target Detection
Authors: Fanzhao Lin, Shiming Ge, Kexin Bao, Chenggang Yan, Dan Zeng
Year: 2024
Type: Journal Article (IEEE Transactions on Multimedia)
DOI: 10.1109/TMM.2023.3325743
Title: Federated Learning with Label-Masking Distillation
Authors: Jianghu Lu, Shikun Li, Kexin Bao, Pengju Wang, Zhenxing Qian, Shiming Ge
Year: 2023
Type: Conference Paper
DOI: 10.1145/3581783.3611984
Conclusion