Thiessen, M., & Gärtner, T. (2021). Active Learning Convex Halfspaces on Graphs. In SubSetML @ ICML2021: Subset Selection in Machine Learning: From Theory to Practice. SubSetML: Subset Selection in Machine Learning: From Theory to Practice, Unknown. https://doi.org/10.34726/3901
SubSetML @ ICML2021: Subset Selection in Machine Learning: From Theory to Practice
-
Date (published):
24-Jul-2021
-
Event name:
SubSetML: Subset Selection in Machine Learning: From Theory to Practice
en
Event date:
24-Jul-2021
-
Event place:
Unknown
-
Number of Pages:
17
-
Peer reviewed:
Yes
-
Keywords:
Machine Learning; Graph Theory; Active Learning; Geodesic Convexity
en
Abstract:
We systematically study the query complexity of learning geodesically convex halfspaces on graphs. Geodesic convexity is a natural generalisation of Euclidean convexity and allows the definition of convex sets and halfspaces on graphs. We prove upper bounds on the query complexity linear in the treewidth and the minimum hull set size but only logarithmic in the diameter. We show tight lower bounds corresponding to well-established separation axioms and identify the Radon number as a central parameter bounding the query complexity and the VC dimension. While previous bounds typically depend on the cut size of the labelling, all parameters in our bounds can be computed from the unlabelled graph. We empirically compare our proposed approach with other active learning algorithms and provide evidence that communities in real-world graphs are often convex.