معرفی
Cyrus Cousins is a Visiting Assistant Professor in the BIGDATA group at Brown University, where he also completed his doctoral studies under Eli Upfal. His research bridges theoretical and practical aspects of machine learning, with a focus on statistical learning theory, fair machine learning, and empirical game theory.
His educational background includes:
- Bachelor's degree in Computer Science, Mathematics, and Biology from Tufts University
- Ph.D. in Computer Science from Brown University
Cousins' research primarily centers on developing novel theoretical bounds for uniform convergence and generalization error, and applying these to real-world problems in data science, empirical game theory, and fair machine learning. He emphasizes non-vacuous bounds in small sample settings and has pioneered fair-PAC learning (minimizing power-means over per-group losses) and adversarial learning from weak supervision. His work draws critical connections between economic theory (e.g., cardinal welfare) and machine learning.
His publication record reveals strong thematic consistency across statistical learning theory and fair machine learning applications. He consistently publishes in top venues (NeurIPS, AAMAS, UAI) with focus on sample complexity analysis, equilibrium computation in games, and interpretable models. His work uniquely integrates rigorous theoretical guarantees with practical implementations, particularly in non-i.i.d. and fairness-constrained settings.
Cousins has received several notable awards:
- NeurIPS 2021 Outstanding Reviewer Award (top 8% of reviewers)
- Dean's Faculty Fellowship at Brown University
- Joukowsky Outstanding Dissertation Prize in physical sciences (2021)
As an educator, Cousins teaches advanced probability courses for computing and data science while mentoring students through graduate teaching assistant roles. His award-winning dissertation established foundational work in concentration of measure for fair ML. He maintains active industry collaborations with Two Sigma Investments and contributes to community initiatives like Queer in AI. His research program demonstrates exceptional synergy between theoretical innovation and societal impact.
Cousins leads research within Brown's BIGDATA group and collaborates extensively through the Empirical Game Theoretic Analysis (EGTA) project with Amy Greenwald and Enrique Areyan Viqueira, developing frameworks for data-driven game analysis and mechanism design.
