Theodore Papamarkou - Machine learning with probability generating functions (online)
| Dates: | 14 December 2022 |
|---|
| Times: | 15:00 - 16:00 |
|---|
| What is it: | Seminar |
|---|
| Organiser: | Department of Mathematics |
|---|
|
|
Theodore Papamarkou (University of Manchester) will speak in the Probability seminar. (online)
A relation between activation functions and probability-generating functions (PGFs) exists in the literature. This relation can be used to describe the connection between neural networks and Gaussian processes in the infinite width limit of the networks' hidden layers. It will be demonstrated how results for certain classes of PGFs yield new classes of activation functions and of correlation kernels. Some relevant open research questions will be discussed.
Travel and Contact Information
Find event
https://zoom.us/j/97662591193
