This National Science Foundation award supports the NSF-CBMS regional conference on Topological Methods in Machine Learning and Artificial Intelligence, hosted by the College of Charleston in Charleston, South Carolina, during the week of May 13-17, 2019. The conference will feature Professor Gunnar Carlsson of Stanford University and Ayasdi Inc. as the Principal Lecturer. Professor Carlsson will deliver a series of ten lectures introducing participants to the fast-emerging field of Topological Data Analysis, which employs many of the techniques commonly used in topology, the study of shape, to analyze massive and complex data sets across multiple application domains. The conference will benefit a broad group of participants as data science is rapidly establishing itself as an interdisciplinary discipline with many high-impact applications. Main targets of the lecture series and the ensuing monograph will be applications to the medical sciences, including, e.g., better targeting and prediction of diseases and improved patient care, though the lectures will benefit a far larger constituency. The great majority of the NSF-supported participants will be recruited from amongst early career researchers, graduate students, minorities, and women.
Topological Data Analysis refers to the use of topology as a tool for understanding and interacting with large and complex data sets. It should be viewed as another step in the development of Machine Learning. Many of the techniques used extensively in topology - including the combinatorial construction of spaces as simplicial complexes, homology and cohomology, local to global methods for computation and application, and the organizing power of the language of category theory and functoriality - all play important roles in the development of this subject. Professor Carlsson?s lecture series and the resulting monograph will introduce students and researchers to this rapidly emerging field. Topics will include topological modeling of data; machine learning; applications of homology to shape analytic tasks, statistics of image patches, and viral evolution; adapting local-to-global methods from topology to point cloud situations; persistence landscapes and persistence images with applications to drug discovery; clustering; and algorithms as data sources. The conference website is at http://math.cofc.edu/CBMS-TDA2019/
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
