The mathematics seminar for students


The mathematics seminar is a seminar series for students at the bachelor's programme in mathematics. Other students interested in mathematics are also welcome to participate.

In the autumn seminars, we focus on initial insights into the subject of mathematics. This spring's seminars shed light on ongoing mathematical research.

Contact for the mathematics seminar: Martin Herschend

Upcoming seminars, spring 2022

Thursday, 19 May: Carmina Fjellström

Time: 13:15

Place: Room Å4006. It is also possible to participate via Zoom: link to the seminar. For Zoom password, contact Mateusz Stroiński.

Title: Diffusion Maps

Abstract: Have you ever worked with data that have 100 parameters? What about 1 000? What about 100 000? Working with high-dimensionality data is challenging and computationally expensive. Naturally, the question arises as to whether it is possible to reduce the dimension in such a way as to retain important features and still be able to do meaningful interpretation. This is called dimensionality reduction, and it is the cornerstone of high-dimensional data analysis. Many well-known methods for this include the famous Principal Component Analysis, Multidimensional Scaling, and Isometric Feature Map. In this talk, I will present a technique called Diffusion Maps. It is a nonlinear method based on diffusion processes that learns the underlying manifold upon which the data is embedded. Doing so uncovers possibly hidden geometric properties of the dataset. The technique is claimed to be computationally efficient and robust to noise.

Last modified: 2022-05-09