About
Sivan Sabato studies the statistical and algorithmic foundations of machine learning, focusing on interactive machine learning. This area of research studies making inferences and predictions from data when the data collection is done during the learning process and is controlled by the learning algorithm. For instance, an algorithm aimed at calculating a prediction model based on lab experiments can direct the lab equipment to run certain experiments based on the results collected so far, so as to get the most accurate predictive model. One of the greatest challenges in this field of research is developing methods to decide which information should be collected. These methods must collect the most useful observations, while avoiding possible biases that can be caused in the process. Sabato addresses this and other challenges using a combination of tools from computer science, statistics and mathematics.
Awards
- Best student paper award (co-author), International Conference on Data Mining (2018)
- Alon scholarship for excellent young faculty, The Israeli Council for Higher Education (2015)
- Google Anita Bork Memorial Scholarship (2011)
- Adams Excellence Scholarship, Israel Academy of Science and Humanities (2008)
Relevant Publications
- Ben-David, N., Sabato, S. (2022, June). A Fast Algorithm for PAC Combinatorial Pure Exploration. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 36, No. 6, pp. 6064-6071).
- Hanneke, S., Kontorovich, A., Sabato, S., Weiss, R. (2021). Universal Bayes consistency in metric spaces. The Annals of Statistics, 49(4), 2129-2150.
- Sabato, S., Yom-Tov, E. (2020, November). Bounding the fairness and accuracy of classifiers from population statistics. In International Conference on Machine Learning (pp. 8316-8325). PMLR.
- Kontorovich, A., Sabato, S., Urner, R. (2017). Active nearest-neighbor learning in metric spaces. The Journal of Machine Learning Research, 18(1), 7095-7132.
- Hsu, D., Sabato, S. (2016). Loss minimization and parameter estimation with heavy tails. The Journal of Machine Learning Research, 17(1), 543-582.
Support Us
CIFAR is a registered charitable organization supported by the governments of Canada, Alberta and Quebec, as well as foundations, individuals, corporations and Canadian and international partner organizations.