I am a Postdoctoral Fellow in the Department of Computer Science at Tulane University working with Nick Mattei. I obtained my PhD in 2025 from the David R. Cheriton School of Computer Science at the University of Waterloo where I was supervised by Kate Larson.

ben-sweden-2022.jpg

Email: research@<myname>.ca

I am on the job market. I’m interested in academic jobs or research-based industry jobs in countries where I can work in English.

I study the applications of social choice with a particular focus on the intersection of machine learning with social choice.

My broad interest is in collective decision-making: What rules or mechanisms guide how groups (of humans or AIs) make joint actions? The bite-sized chunk of this question where I focused much of my PhD research is on exploring the space of learnable voting rules – learning new voting rules provides us decision-making tools well-adapted for any properties we consider important.

I have also worked in a range of applied topics relating to social choice, machine learning, and artificial intelligence:

  • Delegative Voting: Liquid democracy has promising applications to ensemble systems and continual learning, in addition to being a cool tool for human settings!
  • Peer Review: Classical social choice results guide the “micro review” process we use at the GAIW workshop. This forms the basis of a peer review format that might survive AI.
  • Recommender Systems: Ideas of fairness and welfare from the economic and social choice literature capture ideas of fairness not well measured in current recommender systems.
  • Teaching AI Ethics: I have quite enjoyed teaching about both ethical and policy issues of artificial intelligence at both the graduate and undergraduate level.

news

Dec 09, 2026 “DeepVote” has been accepted to NeurIPS 2026! We show that we can give novel axiomatic properties to learned approximations of standard voting rules.
Aug 16, 2026 I’ll be running the 3rd Workshop on Social Choice and Learning Algorithms at IJCAI 2026. Come say hi!
Jan 20, 2026 Our paper studying the empirical behaviour of voting rules has been accepted at AAAI 2026! We show how we can learn about the axiomatic properties of voting rules in a manner complementary to theoretical analysis.

selected publications

  1. NeurIPS
    2026
    DeepVoting: Learning and Improving Voting Rules with Fine-Tuning
    Leonardo Matone, Ben Abramowitz , Ben Armstrong, Avinash Balakrishnan , and Nicholas Mattei
    40th Conference on Neural Information Processing Systems. 2026.
  2. ICLR
    2026
    Designing Rules to Pick a Rule: Aggregation by Consistency
    Proceedings of 14th International Conference on Learning Representations. 2026.
  3. AAAI
    2026
    What Voting Rules Actually Do: A Data-Driven Analysis of Multi-Winner Voting
    Joshua Caiata, Ben Armstrong, and Kate Larson
    40th Annual AAAI Conference on Artificial Intelligence. 2026.
  4. IJCAI
    2024
    Optimizing Viscous Democracy
    Ben Armstrong, Shiri Alouf-Heffetz , and Nimrod Talmon
    33rd International Joint Conference on Artificial Intelligence. 2024.