XIFENG YAN |
home | publications | tutorials | dataset | software |
|
|||
|
Professor
Ph.D. (2006) |
Computer Science Department
Direction |
I am a professor at the University of California at Santa Barbara, the
co-inventor of Mica (and its
studio PromptDialog), an agentic AI
approach to conversational AI assistants. The primary objective of my
research is to explore foundation
models in artificial intelligence, leveraging these models for knowledge
discovery, and developing cross-disciplinary applications in areas like
finance, healthcare, and science. We have extensively innovated on graph mining, graph
data management and pioneered
transformer-based time series
forecasting. News (March 2025): Our agentic AI approach to conversational AI assistants, Mica (Multiple Intelligent Conversational Agents) is now live. After years of exploring the ideal architecture for customer service bots, we finally arrived at a solution we are satisfied with. It is a kind of YAML version of OpenAI's Swarm. Similar to SQL statements, it abstracts away implementation details, providing a succinct development experience, making systematic prompting engineering, testing and debugging possible in the future. News (Feb 2025): Our alumnus, Dr. Yu Su received 2025 Sloan Research Fellowship. News (Aug 2024): PhD. student Zekun Li received JP Morgan PhD. Fellowship. News (Sept 2023): Our GauchoChat team has won the overall first-place prize for the Alexa Prize SocialBot Grand Challenge 5 (SGC5) . We found the weakness of the current large language models (LLMs) including ChatGPT and developed techniques, GauchoChat: Towards proactive, controllable, and personalized social conversation that turns a reactive LLM to a proactive one. News (June 2023): Our GauchoAI team has won the second-place prize for the Alexa Simbot Challenge. Our approach is summarized in GauchoAI: A proactive modular embodied AI that collaborates with humans. |
||||
RESEARCH INTERESTS | ||||
|
||||
|
|
||||||||||
|
Last Modified: July 10, 2022