Alessio Mazzetto
amazzett [at] cs [dot] brown [dot] edu
PhD student in Computer Science
Advisor: Eli Upfal || Brown University
I am a fifth year Computer Science Ph.D. student at Brown University with a research focus in Theoretical Computer Science, where I am fortunate to be advised by Eli Upfal. In Spring 2024, my work is supported by the Kanellakis Fellowship. In Fall 2023, I was a co-instrucor for Advanced Introduction to Probability for Computing and Data Science (CS145. In Summer 2023, I was an intern at Yahoo! Research on the Scalable Machine Learning team.

My main research area is Machine Learning Theory. I am broadly interested in learning settings where there is access to a small amount of data for the target task. In my work, I developed theoretically sound methods that can quantify and use the knowledge provided by different sources other than labeled data for weak supervision. Recently, I worked on the problem of learning with distribution drift, where we are given a sequence of samples from a distribution that gradually changes in time, and we want to solve a learning task with respect to the current distribution.

Prior to coming to Brown, I completed a Master in Computer Science and a Bachelors in Information Engineering from University of Padua in Italy.

News
  • January 2024: My first solo-author paper An Improved Algorithm for Learning Drifting Discrete Distributions was accepted at AISTATS 2024!
  • September 2023: Our paper An Adaptive Algorithm for Learning with Unknown Distribution Drift was accepted at NeurIPS 2023!
  • April 2023: Our paper Nonparametric Density Estimation under Distribution Drift was accepted at ICML 2023!

Publications
  • An Improved Algorithm for Learning Drifting Discrete Distributions
    Alessio Mazzetto
    To appear in Artificial Intelligence and Statistics (AISTATS) 2024
  • An Adaptive Algorithm for Learning with Unknown Distribution Drift
    Alessio Mazzetto and Eli Upfal
    Conference on Neural Information Processing Systems (NeurIPS) 2023
    [pdf]
  • Nonparametric Density Estimation under Distribution Drift
    Alessio Mazzetto and Eli Upfal
    International Conference on Machine Learning (ICML) 2023
    [pdf]
  • Tight Lower Bounds on Worst-Case Guarantees for Zero-Shot Learning with Attributes
    Alessio Mazzetto*, Cristina Menghini*, Andrew Yuan, Eli Upfal, and Stephen H. Bach
    Conference on Neural Information Processing Systems (NeurIPS) 2022
    [pdf]
  • Adversarial Multiclass Learning under Weak Supervision with Performance Guarantees
    Alessio Mazzetto*, Cyrus Cousins*, Dylan Sam, Stephen H. Bach, and Eli Upfal
    International Conference on Machine Learning (ICML) 2021
    [pdf][code]
  • Semi-Supervised Aggregation of Dependent Weak Supervision Sources with Performance Guarantees
    Alessio Mazzetto, Dylan Sam, Andrew Park, Eli Upfal, and Stephen H. Bach
    Artificial Intelligence and Statistics (AISTATS) 2021
    [pdf][appendix][code]
  • Accurate MapReduce Algorithms for k-Median and k-Means in General Metric Spaces
    Alessio Mazzetto, Andrea Pietracaprina, and Geppino Pucci
    International Symposium on Algorithms and Computation (ISAAC) 2019
    [pdf]

Manuscripts
  • An Adaptive Method for Weak Supervision with Drifting Data
    Alessio Mazzetto, Reza Esfandiarpoor, Eli Upfal, and Stephen H. Bach
    Preprint. Under submission.
    [pdf][code]

Awards
  • “Best Master Thesis in Theoretical Computer Science” award from Capitolo Italiano of EATCS, 2020. The award was given to the best master thesis in Theoretical Computer Science in Italy for the year 2019.