AI Seminar: Algorithms for Easy and Hard Online Learning Environments
Online learning is a subfield of machine learning modeling situations, where data acquisition is interleaved with decision making and acting. Examples include investment in the stock market, personalized advertising, online ranking, and so on.
Most existing algorithms for online learning make very specific assumptions about the nature of the environment they operate in. The most common are i.i.d. and adversarial environments. For example, weather prediction is an i.i.d. environment and spam filtering is an adversarial environment. However, in many real-world scenarios the environment is neither purely i.i.d. nor fully adversarial, but something in between.
Time:22 March 2019,15:00
Place: Small UP1, DIKU, Universitetsparken 1, University of Copenhagen
Yevgeny Seldin, Associate Professor in the Machine Learning Section at Department of Computer Science, University of Copenhagen, will give a quick introduction into online learning and then survey the field of algorithms that are optimal over ranges of problems and present some new results.
This seminar is a part of the AI Seminar Series organised by SCIENCE AI Centre. The series highlights advances and challenges in research within Machine Learning, Data Science, and AI. Like the AI Centre itself, the seminar series has a broad scope, covering both new methodological contributions, ground-breaking applications, and impacts on society.