Zhengyuan Liu

Asst. Head of the AI4EDU Programme
Tech Lead of Multimodal Generative AI Group
Agency for Science, Technology and Research (A*STAR), Singapore
IEEE Senior Member

[Email]   [Google Scholar]   [LinkedIn]  

Research Interest
Multimodal and Multilingual Foundation Models.
Generative AI for Education, Healthcare, and Life Science.
Accessibility, Reliability, and Explainability for Human-Centered AI.

Zhengyuan

Current Key Projects
🌟 Human-centered AI | AI for Education Programme
> We conduct interdisciplinary research on how humans and intelligent systems (not limited to LLMs) can effectively and efficiently learn from each other, and explore the world and universe together.
> We are building the Special Interest Group on Human-AI Collaborative Learning, Exploration, and Discovery (InspAIre).

🚀 National Large Language Model Programme (MERaLiON SG)
> Working on improving multilingual and multimodal LLMs' capabilities of knowledge and reasoning.

â–ș “Imagination is more important than knowledge.” – Albert Einstein.

Join Us & Collaboration (📧 just email me)
○ Research Assistant / Student Intern in Human-centered AI. Remote intern is also welcome!
Topic Conversational intelligent tutoring systems, Human-AI collaborative learning, etc.
○ Research Assistant / Student Intern in Generative AI and Multimodal Fundation Models.
Topic Knowledge-enhanced reasoning, Multi-agent collaboration, etc.
○ Research Collaboration in AI for Life Science.
Topic Reach out to me for an online / in-person chat (☕) on next-generation science accelerators, etc.

News [Dec. 2025] Our SingaX team (A*STAR, NTU, NUS) won the 2nd prize 🏅 in the Embodied Agent Interface (EAI) Challenge in NeurIPS 2025. [Award Link]
News [Nov. 2025] Our work “Persuasion Dynamics in LLMs: Investigating Robustness and Adaptability in Knowledge and Safety with DuET-PD” is published in EMNLP 2025. [Paper]
News [Nov. 2025] Our work “CCL-XCoT: An Efficient Cross-Lingual Knowledge Transfer Method for Mitigating Hallucination Generation” is published in EMNLP 2025. [Paper]
News [Jun. 2025] Presented our work “COGENT: A Curriculum-oriented Framework for Generating Grade-appropriate Educational Content” in BEA 2025. [Paper]
News [May. 2025] Presented our work “SingaKids: A Multilingual Multimodal Dialogic Tutor for Language Learning” in ACL 2025 (Oral). [Paper]
News [May. 2025] Our work “Reinforcing Compositional Retrieval: Retrieving Step-by-Step for Composing Informative Contexts” is published in ACL 2025. [Paper]
News [Mar. 2025] Our work “Scaling Up Collaborative Dialogue Analysis: An AI-driven Approach to Understanding Dialogue Patterns in Computational Thinking Education” got published in LAK 2025. [Paper]
News [Jan. 2025] Our paper “CrossIn: An Efficient Instruction Tuning Approach for Cross-Lingual Knowledge Alignment” won the 🏅 Best Paper Award in SumEval@COLING 2025. [Paper]
News [Nov. 2024] Our paper “Learning planning-based reasoning by trajectories collection and process reward synthesizing” won the 🏅 Outstanding Paper Award in EMNLP 2024. [Paper]
News [Nov. 2024] Presented our work “Personality-aware Student Simulation for Conversational Intelligent Tutoring Systems” in EMNLP 2024. [Paper]
News [Sep. 2024] Presented our work “Optimizing Code-Switching in Conversational Tutoring Systems: A Pedagogical Framework and Evaluation” in SIGDIAL 2024. [Paper]
News [Sep. 2024] Our paper “CRAFT: Extracting and Tuning Cultural Instructions from the Wild” won the 🏅 Best Paper Award in C3NLP@ACL2024. [Paper]
News [Feb. 2024] Presented our work “Scaffolding Language Learning via Multi-modal Tutoring Systems with Pedagogical Instructions” in IEEE CAI 2024. [Paper]
News [Dec. 2023] Our paper “Joint Dialogue Topic Segmentation and Categorization: A Case Study on Clinical Spoken Conversations” won the 🏅 Outstanding Paper Award in EMNLP 2023. [Paper]
News [Jul. 2022] Presented our work “Learning from Bootstrapping and Stepwise Reinforcement Reward: A Semi-Supervised Framework for Text Style Transfer” in NAACL 2022. [Paper]
News [Sep. 2021] Our paper “Coreference-Aware Dialogue Summarization” won the 🏅 Best Paper Award in SIGDIAL 2021. [Paper]

Academic Service
Assistant Program Chair in NeurIPS 2025.
Conference Program Committee, Reviewer, and Area Chair:
NeurIPS, ICLR, ACL, EMNLP, NAACL, AAAI, LAK, LERC, ICASSP, Interspeech, etc.
Conference Session Chair: ICASSP and Interspeech.
Journal Reviewer: IEEE TASLP, ACM CSUR, Neurocomputing, Computer Speech & Language, etc.