University of Cambridge > Talks.cam > Machine Learning @ CUED > Efficient Sampling with Kernel Herding

  Browse A-Z

Log in

Cambridge users (raven)details Other usersdetails No account?details

Information on

Subscribing to talksdetails Finding a talkdetails Adding a talkdetails Disseminating talksdetails Help and Documentationdetails

Efficient Sampling with Kernel Herding

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

If you have a question about this talk, please contact Zoubin Ghahramani.

The herding algorithm was proposed as a deterministic dynamic system that integrates learning and inference for discrete Markov Random Fields. In this talk, we take a dual view of herding and extend it to continuous spaces by using the kernel trick. The resulting “kernel herding” is an infinite memory deterministic algorithm that approximates a PDF with a collection of samples. We show that kernel herding decreases the error of expectations of functions in the Hilbert space much faster than the usual iid random samples. If time permits, I’ll also talk about the recent development and applications of the herding algorithm.

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

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

Other lists

Machine Learning Reading Group @ CUED Computer Science Tripos Seminar Series Mackenzie-Stuart Lectures

Other talks

Alzheimer's talks Active Subspace Techniques to Construct Surrogate Models for Complex Physical and Biological Models Modularity, criticality and evolvability of a developmental GRN MEMS Particulate Sensors In search of amethysts, black gold and yellow gold Highly Energy Efficient Key-value Store for In-network Computing BP KEYNOTE LECTURE: Importance of C-O Bond Activation for CO2/COUtilization - An Approach to Energy Conversion and Storage The frequency of ‘America’ in America Café Synthetique: Graduate Talks! Crowding and the disruptive effect of clutter throughout the visual system Existence of Lefschetz fibrations on Stein/Weinstein domains Developing novel methods for interrogating tree ring anatomy for use in modelling carbon sequestration
 

© 2006-2024 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity