Welcome
Home
Courses
Recent news
Selected Talks
Research Papers
All Papers 
Minimally Supervised Learning
(e.g., Active Learning, Transfer Learning)
Unsupervised Learning
(e.g., Clustering, Causality, Geometric Graphs)
Supervised Learning
(e.g., Manifold, Sparsity, Tradeoffs)
Miscellaneous
Software Code

Samory Kpotufe

My photo 

Professor, Department of Statistics, Columbia University


Research Interests

My photo 

I work in statistical machine learning theory which aims to understand performance limits in various problems of modern interest, and pinpoint beneficial aspects of data that ML procedures might leverage. My focus earlier in my career had been on common nonparametric methods (e.g., kNN, trees, kernel averaging, density-based clustering), and has evolved to include traditional parametric methods (e.g., bounded VC classes, regression on finite-dimensional linear spaces) in non-i.i.d. settings where their performance is yet to be fully understood.

My work often deals with adaptivity, i.e., how to automatically leverage beneficial aspects of data as opposed to designing specifically for each scenario, while accounting for modern application constraints (e.g., space and computation time, imperfect data).

Some specific interests over the years: performance limits and adaptivity in active learning, transfer and multi-task learning as they appear in classification, regression, and bandits problems; notions of intrinsic data dimension, statistical and computational benefits (or lack thereof) of sparse or manifold representations; hyperparameter-tuning and guarantees in density-based clustering.

Professional Activities

  • co-Chair: Foundations of Data Science, Columbia Data Science Institute.

  • Associate Editor: Statistical Science (2023 to present).

  • Associate Editor: SIAM Mathematics of Data Science (2021 to present).

  • Associate Editor: Journal of the American Statistical Association (JASA), Theory and Methods (2023 to present).

  • Action Editor: Journal of Machine Learning Research (2020 to 2023).

  • Program co-Chair: COLT (2021).

  • Publication Chair: ICML (2020)

  • Advisory Committee on Equality and Diversity: Institute of Mathematical Statistics (2020).

  • Editorial Board: Journal of Machine Learning Research (2014 to 2018).

  • Senior Area Chair: NeurIPS (2021, 22).

  • Area Chair: ICML (2019), NeurIPS (2015, 16, 18, 19), COLT (2015, 16, 17), ALT (2018, 2020), AISTATS (2017, 18, 19).

  • Reviewing: JMLR, Machine Learning, IEEE Info Theory, IEEE TPAMI, AoS, ESAIM Probability and Statistics, Bernouilli, Biometrika, NeurIPS, SODA, COLT, ICML, AAAI, NeurIPS, AISTATS, …

Short bio

I graduated (Sept 2010) in Computer Science at the University of California, San Diego, advised by Sanjoy Dasgupta. I then was a researcher at the Max Planck Institute for Intelligent Systems. At the MPI I worked in the department of Bernhard Schoelkopf, in the learning theory group of Ulrike von Luxburg. Following this, I spent a couple years as an Assistant Research Professor at the Toyota Technological Institute at Chicago. I then spent 4 years at ORFE, Princeton University as an Assistant Professor. I officially joined Columbia in Spring 2019, followed by a visiting membership at the Institute of Advanced Study in Spring 2020. For various honors over the years, see the 'recent news’ section.