Hi, I am Xinyi Lu, a third-year PhD student at University of Michigan, advised by Prof. Xu Wang. Before that, I got my B.S degree from both University of Michigan (majoring in Computer Science), and Shanghai Jiao Tong University (majoring in Electrical and Computer Engineering).

I'm interested in AI in Education, My research focuses on designing assistance for creating high-quality educational content, which ultimately supports students in developing expertise for complex and challenging tasks. Specifically, drawing on theories and methods from Learning Science and AI Interpretability, I design AI assistance that aligns with experts’ cognitive processes, while preserving their full agency throughout the process. I also investigate how effective the human-AI co-created educational content is for learning. To design and evaluate these innovations, I apply both qualitative lab studies and large-scale classroom studies.

News

July, 2025 I will present our work on exploring LLM-Generated feedback on knowledge-intensive essays at AIED 2025
April, 2025 I will attend NCME Annual Meeting and CHI 2025
July, 2024 I will present Generative Student at L@S 2024 in Atlanta
August, 2023 Graduated from SJTU 🎓
April, 2023 ReadingQuizMaker won Best Paper Honorable Mention Awards at CHI 2023 🏆
April, 2023 I will present ReadingQuizMaker at CHI 2023 in Hamburg

Selected Publications

Exploring LLM-Generated Feedback for Economics Essays: How Teaching Assistants Evaluate and Envision Its Use
Xinyi Lu, Aditya Mahesh*, Zejia Shen*, Mitchell Dudley, Larissa Sano, Xu Wang

AIED'2025
paper
Generative students: Using LLM-Simulated Student Profiles to Support Question Item Evaluation
Xinyi Lu, Xu Wang

Learning@Scale'2024
paper
ReadingQuizMaker: A Human-NLP Collaborative System that Supports Instructors to Design High-Quality Reading Quiz Questions
Xinyi Lu, Simin Fan, Jessica Houghton, Lu Wang, Xu Wang

CHI'2023 Best Paper Honorable Mention
paper · video

Games

The Legend of Zelda (Recreate and Upgrade)

Recreated the Level 1 of The Legend of Zelds by NES in the unity game engine, and added a new level with Michigan snow in it.

Play it here
Paws and Claws

A party game of cats and dogs rivalry created with Unity game engin. Players play as a cat and a dog and build their own game map to seize the flag and claim victory. With an intuitive block-building system, players can design and customize their own unique game maps, complete with obstacles, traps, and other challenges.

Download it here