Chang’an University, China
Xin Hu is a rapidly emerging academic voice in the field of computer vision, multimodal learning, and few-shot learning, currently serving as a Lecturer at the School of Data Science and Artificial Intelligence, Chang’an University, China. With a strong research background in diagram understanding and cross-modal information retrieval, Xin Hu is recognized for bridging the gap between image recognition and language understanding, particularly in educational and knowledge representation contexts. His innovative methodologies address real-world challenges where data scarcity, particularly in educational visual content, hinders effective AI interpretation and deployment. As a researcher with diverse interdisciplinary collaborations, Xin Hu’s work stands at the intersection of artificial intelligence, education technology, and cognitive computing.
Author Profile
ORCID
Education
Xin Hu completed his Bachelor of Engineering in Digital Media Technology and later a Master’s in Computer Technology at Xi’an Shiyou University. He further pursued and earned a Ph.D. in Computer Science and Technology from Xi’an Jiaotong University under the mentorship of Professor Jun Liu. Throughout his academic training, Xin Hu developed foundational skills in artificial intelligence, multimodal signal processing, and machine learning, with a specific focus on visual and linguistic data fusion, eventually applying these to real-world educational datasets and semantic tasks.
Experience
Xin Hu began his formal research journey in 2018 as a Ph.D. candidate, where he contributed to cutting-edge projects under China’s National Key Research and Development Program. He was actively engaged in two major national projects centered on big data knowledge engineering and educational data analysis. These projects aimed to enhance semantic retrieval and intelligent knowledge visualization, particularly in education. His role spanned from system architecture to guiding junior researchers and developing novel few-shot learning frameworks. By late 2023, Xin Hu had joined Chang’an University as a full-time Lecturer, where he continues to explore advanced multimodal learning models with practical educational applications.
Research Interests
Xin Hu’s primary research interests lie in computer vision, particularly few-shot learning, multimodal learning, and visual-linguistic matching tasks. He has demonstrated a unique ability to develop models that operate under limited supervision, focusing on diagrammatic content—an area often overlooked in mainstream AI research. His cross-modal attention frameworks and gestalt-perception-based approaches enable AI systems to better interpret complex visual content, such as diagrams in educational settings. His work in few-shot diagram-sentence matching (Fs-DSM) and gestalt-transformers has further extended AI’s capability to learn from minimal annotated data while preserving semantic integrity.
Awards
While no standalone awards are explicitly listed, Xin Hu has been a consistent contributor to top-tier journals and conferences including IEEE Transactions on Image Processing, Neural Computation, AAAI, and IJCAI. His works have garnered significant citations, demonstrating academic influence and peer validation. As part of national R&D projects, he has also played a key role in transforming applied AI methodologies into deployable knowledge systems. His presence in IEEE and ACM conferences and workshops shows sustained engagement with the global AI research community.
Notable Publications
LFSRM: Few-Shot Diagram-Sentence Matching via Local-Feedback Self-Regulating Memory
Authors: Lingling Zhang, Wenjun Wu, Jun Liu, Xiaojun Chang, Xin Hu, Yuhui Zheng, Yaqiang Wu, Qinghua Zheng
Journal: IEEE Transactions on Pattern Analysis and Machine Intelligence
Year: 2025
Hierarchy-Based Diagram-Sentence Matching on Dual-Modal Graphs
Authors: Wenjun Wu, Lingling Zhang, Jun Liu, Ming Ren, Xin Hu, Jiaxin Wang, Qianying Wang
Journal: Pattern Recognition
Year: 2025
SKFormer: Diagram Captioning via Self-Knowledge Enhanced Multi-Modal Transformer
Authors: Xin Hu, Jiaxin Wang, Tao Gao
Journal: Signal Processing
Year: 2025
Alignment Relation is What You Need for Diagram Parsing
Authors: Xinyu Zhang, Lingling Zhang, Xin Hu, Jun Liu, Shaowei Wang, Qianying Wang
Journal: IEEE Transactions on Image Processing
Year: 2024
Contrastive Graph Representations for Logical Formulas Embedding
Authors: Qika Lin, Jun Liu, Lingling Zhang, Yudai Pan, Xin Hu, Fangzhi Xu, Hongwei Zeng
Journal: IEEE Transactions on Knowledge and Data Engineering
Year: 2023