Scholar
Xifeng Yan
Google Scholar ID: XZV2eogAAAAJ
Professor, Computer Science, Univ. of California at Santa Barbara
Citations & Impact
Citations
11,937
H-index
47
i10-index
134
Publications
20
Co-authors
23
list available
Contact
Publications
Resume (English only)
Academic Achievements
- Published multiple papers and delivered talks at various international conferences, such as an invited talk at Nvidia on 'Adaptive Inference in Pretrained LLMs,' a spotlight talk at the Conference on Language Modeling (COLM) about 'Adaptive Layer-skipping in Pre-trained LLMs.' The training dataset of the multimodal language model Open-Qwen2VL has been downloaded over 90,000 times on HuggingFace.
Research Experience
- Professor at the Computer Science Department, University of California, Santa Barbara, focusing on foundation models in AI, graph mining, graph data management, and transformer-based time series forecasting.
Education
- Ph.D. (2006), University of Illinois at Urbana-Champaign
Background
- Research Interests: Foundation models in artificial intelligence, knowledge discovery, and cross-disciplinary applications (e.g., finance, healthcare, and science). Innovated extensively on graph mining, graph data management, and transformer-based time series forecasting. Co-inventor of ADL/Mica, an agent-first approach to conversational AI assistants.
Miscellany
- Personal interests: Attended the 6th International Conference on Data-Driven Plasma Science in Santa Fe, noting that reading legendary stories from Los Alamos and articles about complexity science at the Santa Fe Institute during childhood influenced his career interests.
