Approximation Theorems for Keisler Measures

Doctoral Dissertation


This dissertation is concerned with Keisler measures and their approximations. We investigate tame families of Keisler measures in varying contexts. We first restrict ourselves to the local NIP setting. There, we partially generalize a theorem of Hrushvoski, Pillay, and Simon and show that in this context, a measure is definable and finitely satisfiable (dfs) if and only if it is finitely approximated. We then consider generically stable measures outside of the NIP setting. We show that generically stable types correspond to {0,1}-valued frequency interpretation measures, and we give examples of finitely approximated measures which are not frequency interpretation measures and local dfs measures which are not locally finitely approximated (joint with Gabriel Conant). We then introduce and describe “sequential approximations”. We show that measures which are finitely satisifable in a countable model of an NIP theory admit this kind of approximation. We also prove that generically stable types admit a similar (and stronger) approximation. In the final chapter, we restrict ourselves to the group setting and introduce a convolution operation on Keisler measures. We classify all idempotent measures over stable groups and also show that a particular convolution algebra over an NIP group is isomorphic to a natural Ellis semigroup (joint with Artem Chernikov).


Attribute NameValues
Author Kyle Gannon
Contributor Minh Chieu Tran, Committee Member
Contributor Sergei Starchenko, Research Director
Contributor Anand Pillay , Committee Member
Contributor Daniel Hoffmann, Committee Member
Degree Level Doctoral Dissertation
Degree Discipline Mathematics
Degree Name Doctor of Philosophy
Banner Code

Defense Date
  • 2020-04-29

Submission Date 2020-06-14
Record Visibility Public
Content License
  • All rights reserved

Departments and Units
Catalog Record


Please Note: You may encounter a delay before a download begins. Large or infrequently accessed files can take several minutes to retrieve from our archival storage system.