IC Monday Seminar - "Metric structures on datasets: design, stability, and classification of algorithms"
Event details
| Date | 21.02.2011 |
| Hour | 16:15 |
| Speaker | Dr Facundo Memoli, Stanford University - IC Faculty candidate |
| Location |
INM202
|
| Category | Conferences - Seminars |
Abstract:
Several methods in data and shape analysis can be regarded as transformations between metric spaces, or in general, as maps from k-tuples of metric spaces into a metric space. Examples are hierarchical clustering methods, the higher order constructions of computational persistent topology, and several techniques that operate within the context of data/shape matching under invariances. Metric geometry, and in particular variants of the Gromov-Hausdorff distance provide a point of view which is applicable in different scenarios. The underlying idea is to regard datasets as metric spaces (or metric measure spaces), and then, crucially, at the same time, regard the collection of all datasets as a metric space in itself. Variations of this point of view give rise to different taxonomies that include several methods, both preexisting and novel, for extracting information from datasets and shapes. I will give and overview of these constructions and describe an ongoing application of one of these new methods to functional classification of chemical compounds by shape properties.
Bio : Facundo Mémoli is a postdoctoral fellow in the Mathematics
Department at Stanford University. His research interests are in
shape and data analysis and in metric geometry. Facundo received his B.Sc. and M.Sc.
in Electrical Engineering from the Universidad de la Republica in
Uruguay, and his Ph.D. in Electrical Engineering from the University of Minnesota
Links
Practical information
- General public
- Free
Contact
- Christine Moscioni