(Jul 2025) Thanks Thomson Reuters and Naver for generous gift and key research collaboration.
(Apr 2025) Zae Myung Kim won the Doctoral Dissertation Fellowship, focusing on meta-scaffolding of LLMs.
(Jan 2025) It's grateful to receive the NSF NAIRR award (280,000 GPU Hours) to build scientific foundational models for trustworthy scientific writing and discovery.
(Dec 2024) James Mooney won the Amazon MLSys Fellowship for 2025-2026, focusing on Algorithmic and Agentic efficiency.
Karin de Langis won the Doctoral Dissertation Fellowship, focusing on Cognitively Informed Natural Language Generation.
(Mar 2024) Some of our recent work on writing assistant are featured in UMN CS (link)
(Dec 2023) Excited to receive a grant from UMN CTSI for the development of AI Assistant for Non-English Speaking Clinical Trial Partipants (/w Jennifer Needle and Gwenyth Fischer)
(Dec 2023) Thrid In2Writing workshop (focused on "Dark Sides: Envisioning, Understanding, and Preventing Harmful Effects of Writing Assistants") will be held at CHI 2024 in Hawaii.
(Oct 2023) We are grateful to OpenAI, Cohere, Google, Amazon, and Oracle for their generous support of our computing needs
I build human-centric language technologies, focusing on cognitively aligning human and machine thinking, advancing AI as thinking partners.
My work bridges language and cognition to develop novel algorithms, benchmarks, and interaction frameworks that support experts in complex, real-world cognitive workflows.
I design AI that augments humans: models that learn from how people plan, reason, and create; anticipate cognitive bottlenecks; scaffold difficult tasks; and adapt dynamically to expert strategies. See my research statement (technical, non-technical). My current research focuses on three areas:
Cognitive scaffolding: Enhancing LLMs with human cognition data to improve planning, abstraction, and reasoning.
Expert-level AI: Building interactive systems for collaboration with scientists, lawyers, journalists, and educators, and benchmarking their workflows and thought processes.
Societal alignment: Designing inclusive NLP systems that reflect pluralistic perspectives.