معرفی
Virginia Smith is the Leonardo Associate Professor of Machine Learning at Carnegie Mellon University's School of Computer Science, with a courtesy appointment in the Electrical and Computer Engineering Department. She leads research addressing critical challenges in machine learning systems, particularly focusing on safety and efficiency.
Her research interests span federated and collaborative learning, efficient training methods, data privacy, and AI safety. Recent work has explored topics such as model unlearning, LLM security, and resource-efficient distributed learning systems. She has made significant contributions to understanding how to make machine learning systems more robust, private, and efficient while maintaining performance.
Professor Smith's publication record demonstrates a clear progression toward addressing practical challenges in deploying machine learning systems at scale. Her recent work shows strong emphasis on large language model safety, privacy-preserving techniques, and efficient distributed learning approaches. The research spans theoretical foundations to practical implementations, with numerous papers appearing in top-tier venues including NeurIPS, ICML, ICLR, and MLSys.
- AFOSR Young Investigator Award
- Sloan Research Fellowship
- 2023 Samsung AI Researcher of the Year
- Best Paper Award at ICML 2025 Exploration in AI Workshop
- Outstanding Paper Award at MLSys 2023
As an educator, Professor Smith mentors numerous PhD students and postdocs while teaching advanced machine learning courses at CMU. She serves as Program Chair for ICML 2025 and co-organizes a semester program on Federated and Collaborative Learning at the Simons Institute. Her research group maintains strong collaborations with industry partners including Amazon, where she has received research awards.
