Andrew Jacobsen

Andrew Jacobsen is a Postdoctoral Research Fellow affiliated with the Polytechnic University of Milan and the University of Milan, working under the supervision of Nicolò Cesa-Bianchi. His research focuses on theoretical aspects of machine learning, particularly online learning and adaptive algorithms for dynamic environments.

His educational background includes:

• PhD in Computing Science (2024) from the University of Alberta, supervised by Ashok Cutkosky and Martha White.
• MSc and BSc in Computing Science from the University of Alberta, supervised by Martha White and Adam White.
• Music program at Grant Macewan University, majoring in composition and minoring in music technology.

Andrew's primary research interest lies in online learning theory, with a focus on developing algorithms that require no hyperparameter tuning and can adapt to time-varying objectives. He is also broadly interested in machine learning theory, reinforcement learning, and bandits. His work bridges theoretical guarantees with practical algorithmic design, creating robust systems for non-stationary environments through parameter-free optimization and regret minimization frameworks.

His publication trends reveal a strong emphasis on dynamic regret analysis, equivalence proofs between static and dynamic regret, and hyperparameter-free methods across online optimization. Recent work extends into reinforcement learning applications and continual learning systems, consistently targeting theoretical foundations with practical implementations.

Scientific recognition includes:

• Outstanding PhD Thesis Award for "Adapting to Non-stationarity in Online Learning" (2024).

Andrew has no documented student advising or grant leadership in the provided materials. He maintains active research collaborations at his institutions while pursuing interdisciplinary interests in music composition and physical activities like skateboarding and grappling.