Document Analysis and Pattern Recognition Laboratory

Virtue·Conduct·Erudition·Competence

中文

Document Analysis and Pattern Recognition Laboratory

Virtue·Conduct·Erudition·Competence

Machine learning is a core branch of artificial intelligence. It enables computers to make predictions or decisions without explicit programming by automatically learning rules and patterns from data. Its core approaches include supervised learning (training models with annotated data, e.g., classification and regression), unsupervised learning (exploring structures within unlabeled data, e.g., clustering and dimensionality reduction), and reinforcement learning (optimizing decision‑making strategies through interactions with the environment). Key theories cover loss‑function optimization, bias‑variance trade‑off, and regularization techniques for overfitting mitigation. As a subfield of machine learning, deep learning realizes hierarchical automatic extraction of complex features via multi‑layer neural networks and optimizes parameters through back‑propagation. It excels at processing unstructured data such as images and texts yet requires substantial data and computing resources. Together, they drive breakthroughs in artificial‑intelligence domains including image recognition and natural‑language processing.

Yuchen Zheng
Yuchen Zheng
Professor / Master's Supervisor / Doctoral Supervisor

Research More >

Our laboratory conducts research on pattern recognition and intelligent machine learning, focusing on fundamental algorithms, interdisciplinary applications, and AI privacy security. Core research topics include deep learning, few‑shot learning, and cutting‑edge AI for Science technologies. We work on multi‑scenario intelligent document analysis, covering scene‑text recognition, ancient‑document parsing, handwritten‑text recognition, and intelligent handwritten‑signature verification. We also investigate medical‑image processing and imbalanced‑sample learning for healthcare intelligence. Leveraging data mining and graph neural networks, our bioinformatics research involves protein‑structure prediction, drug discovery, biomolecular‑interaction analysis, microbiome analysis, and biomedical knowledge‑graph construction. Additional research directions include intelligent remote‑sensing image interpretation, multilingual speech recognition and semantic analysis, as well as AI privacy‑security technologies such as federated learning and adversarial attack‑and‑defense. We aim to develop few‑shot, robust, secure intelligent‑analysis technologies for complex scenarios and advance artificial‑intelligence applications in cultural heritage, healthcare, biology, and remote sensing.

Lab News More >

Publications More >

Corresponding author*
Co-first author#