Connections between kernels, GPs, and NNs

Add to your list(s) Download to your calendar using vCal

• Matthias Bauer (University of Cambridge); Paul Rubenstein
• Thursday 12 November 2015, 14:30-16:00
• Engineering Department, CBL Room 438.

If you have a question about this talk, please contact Yingzhen Li.

First, we will follow Radford Neal’s PhD thesis and explain that an infinite Neural Network with random weights is equivalent to a GP. We will then look at several examples of Neural Network kernels and discuss their properties.

In the second part, we are going to talk about GPs and Kernels in the context of regression. We will derive Kernel Ridge Regression and show that it is equivalent to MAP inference in a GP regression model. Along the way we will give a brief introduction to the theory of Reproducing Kernel Hilbert Spaces. Time permitting, we will end by introducing Support Vector Regression – another Kernel regression technique, but one that cannot be viewed as performing MAP inference in any GP model.

There is no required reading. However, if you want to read something, feel free to look at Chapter 2 of Radford Neal’s thesis. html pdf

This talk is part of the Machine Learning Reading Group @ CUED series.

This talk is included in these lists:

• All Talks (aka the CURE list)
• Cambridge Centre for Data-Driven Discovery (C2D3)
• Cambridge Forum of Science and Humanities
• Cambridge Language Sciences
• Cambridge University Engineering Department Talks
• Cambridge talks
• Centre for Smart Infrastructure & Construction
• Chris Davis' list
• Computational Continuum Mechanics Group Seminars
• Engineering Department, CBL Room 438
• Featured lists
• Guy Emerson's list
• Hanchen DaDaDash
• Inference Group Journal Clubs
• Inference Group Summary
• Information Engineering Division seminar list
• Interested Talks
• ML
• Machine Learning Reading Group
• Machine Learning Reading Group @ CUED
• Machine Learning Summary
• Quantum Matter Journal Club
• Required lists for MLG
• School of Technology
• Simon Baker's List
• TQS Journal Clubs
• Trust & Technology Initiative - interesting events
• bld31
• custom
• ndk22's list
• ob366-ai4er
• rp587
• yk373's list

Note that ex-directory lists are not shown.