Piotr Indyk, Professor in the MIT Department of Electrical Engineering and Computer Science (EECS), CSAIL member, and co-director of Foundations of Data Science Institute (FODSI), was recently elected as a member of the American Academy of Arts & Sciences. Indyk was honored for his work on algorithms that focus on geometry and massive data problems.
The MIT professor focuses on efficient, sublinear, and streaming algorithms, with innovative research addressing issues in large data and high-dimensional geometry, which relates to the geometry of spaces with more than three dimensions. In his work, Indyk has addressed the nearest neighbor search, a problem related to finding highly similar data points without needing to scan the entire database, with locality-sensitive hashing. He co-developed this technique for efficient approximate algorithms that can reduce the computational time required to retrieve similar data points, making it useful for data mining, computer vision, and machine learning, among other applications.
Additionally, Indyk has worked on sublinear algorithms that can use limited time and space to navigate massive data streams, such as those he co-developed for the sparse Fourier transform. While the fast Fourier Transform can separate signals into their individual frequencies, Indyk’s algorithm can perform this task up to a hundred times faster when few such frequencies are present, speeding up tasks such as GPS synchronization.
Indyk’s previous recognitions include a National Science Foundation Career Award In 2002, a Packard Fellowship and a Sloan Fellowship in 2003, and the Association for Computing Machinery (ACM)’s Paris Kanellakis Theory and Practice Award in 2012. He has also been named a Simons Investigator and a Fellow of the ACM.