Post by Tianlong Xu

Excited to share two recent papers from our work in AI for education accepted at top conferences: AIED2026: https://lnkd.in/dcKHBw9U ACL2026: https://lnkd.in/dTRzDXZS For the past two years, we have been building around a core idea: student interactions contain far richer value than what is visible on the surface. By leveraging multimodal input signals from students, we can move beyond simple performance evaluation and build a much richer understanding of their learning process, reasoning patterns, and opportunities for improvement. What makes this journey especially meaningful to me is that it began as a highly innovative 0-to-1 product idea. As the work progressed, it became clear that its value extended far beyond product impact: it gradually evolved into a research direction with strong academic significance. Seeing it mature into both a production-ready system and a line of work recognized by the research community has been especially rewarding. I have always believed that some of the most impactful AI work happens when product innovation and research innovation reinforce each other. This project is a good example of that belief in practice. Grateful to the interns, collaborators, and teammates who contributed to making this possible (Dingjie Song; Yibo YAN; Qingsong Wen) 👏 🎉 🎊 . #AI4Education #LLM #MultimodalAI #EdTech #AIEducation #AIProduct