معرفی
Ashok Cutkosky is an Assistant Professor at Boston University in the Electrical & Computer Engineering department, with affiliations in Systems Engineering and Computer Science. His research focuses on machine learning, stochastic optimization, and online learning, emphasizing adaptive and parameter-free algorithms. He holds a PhD from Stanford University (2018) and an AB in Mathematics from Harvard (2013), with a unique mention of a 'Master of Medicine.'
- Primary Appointment: Electrical & Computer Engineering
- Affiliated with: Intelligent, Autonomous & Secure Systems Group
Key areas of research include non-convex optimization, adaptive online learning, and developing algorithms that require minimal prior information. His work bridges theoretical foundations and practical applications in machine learning optimization.
Research Highlights
Dr. Cutkosky's research emphasizes parameter-free online learning algorithms, achieving optimal performance without hyperparameter tuning. Notable contributions include:
- Optimal stochastic non-smooth non-convex optimization
- Mechanics of learning rate tuning (Mechanic algorithm)
- Dynamic balancing for model selection in bandits and reinforcement learning
Awards & Honors
- 2017 COLT Best Student Paper
- NSF Graduate Research Fellowship (2013-2018)
- Stanford Graduate Fellowship
- Hoopes Prize and Captain Jonathan Fay Prize (2013)
Teaching
Recent courses include:
- EC414: Introduction to Machine Learning
- EC525: Optimization for Machine Learning
- EC524: Deep Learning (co-taught)
