Project abstract for group srivbane

Combinatorial Online Learning and its Applications

These researchers investigate efficient online learning algorithms for combinatorial problems as well as their applications on a variety of real-world domains. Online algorithms maintain adaptive models for data over time without making any statistical assumptions about the data. Problems of particular interest include high-dimensional sparse optimization problems, graph-structured linear programs, and convex programs with semi-definiteness constraints. The corresponding models will be applied to a variety of real world domains including computational finance, climate sciences, forest ecology, social media analytics, and cancer genomics. 

A bibliography of this group’s publications acknowledging MSI is attached.