Huiyuan Lai

Huiyuan Lai is a Researcher in the Department of Computational Linguistics at the Faculty of Arts, University of Groningen, specializing in natural language processing with emphasis on language modeling, large language models, and figurative language generation. Their work bridges theoretical advances with practical applications across computational linguistics and AI-driven biomedical research.

Education:

• PhD in Computational Linguistics, University of Groningen (2024). Dissertation: "Neural Text Rewriting: Style Transfer, Figurative Language, and Beyond".

Research Interests:

Lai's research centers on advancing NLP through innovative methodologies, with particular focus on:

• Language modeling architectures and large language model optimization
• Automatic generation of figurative language (metaphors, sarcasm)
• Multilingual instruction tuning for reasoning consistency
• Reproducibility frameworks in NLP research
• Human-AI collaboration in dialog systems
• Neural text rewriting and style transfer techniques

Publication Trends:

Analysis of Lai's 2024-2025 output reveals a trajectory focused on leveraging LLMs for complex linguistic tasks while addressing critical challenges in reproducibility and cross-domain adaptation. Key themes include the systematization of figurative language generation, multilingual reasoning consistency (mCoT framework), human-in-the-loop refinement of dialog systems (HED-IT), and biomedical applications of contrastive learning (ContraMAE). Their work demonstrates consistent innovation in adapting language models to specialized domains while maintaining methodological rigor.

Collaborations:

• Primary collaborator: Malvina Nissim (University of Groningen)
• Biomedical research partnerships with S. Wang, Q. Zhang, and international BIBM conference contributors