Piotr Indyk

Piotr Indyk is Thomas D. and Virginia W. Cabot Professor in the Theory of Computation Group at the Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology.

He has made foundational contributions to the design and analysis of efficient approximate algorithms for massive data processing. His work spans algorithms for high-dimensional geometric problems, streaming and sublinear algorithms, fine-grained complexity and learning-augmented algorithms.

Indyk co-developed the notion of locality-sensitive hashing (LSH), a tool for designing efficient algorithms for searching for similar objects in large data sets. Its core idea is to map objects into bins such that two objects that are similar are more likely to be mapped together than objects that are different. Algorithms based on this approach have been adopted and refined by numerous research and industrial entities around the world.

In addition to his work in high-dimensional computational geometry, Indyk developed approximate algorithms that use sub-linear amounts of space or time. His contributions include space-efficient algorithms for estimating frequency moments, distances between data sets, and other key statistics. His work on sublinear-time algorithms led to highly efficient methods for computing the discrete Fourier transform for signals consisting of few frequencies.

Indyk's research has not only enriched the theoretical underpinnings of computer science but has also influenced practical applications and interdisciplinary research in fields such as signal processing, computer vision, and machine learning.

Indyk received his MA from the University of Warsaw and his Ph.D. degree from Stanford University.