BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Adaptive Gaussian Processes on Graphs via Spectral Graph Wavelets 
 - Felix Opolka
DTSTART:20220208T131500Z
DTEND:20220208T141500Z
UID:TALK168911@talks.cam.ac.uk
CONTACT:Mateja Jamnik
DESCRIPTION:"Join us on Zoom":https://zoom.us/j/99166955895?pwd=SzI0M3pMVE
 kvNmw3Q0dqNDVRalZvdz09\n\nGraph-based models require aggregating informati
 on in the graph from neighbourhoods of different sizes. In particular\, wh
 en the data exhibit varying levels of smoothness on the graph\, a multi-sc
 ale approach is required to capture the relevant information. In this work
 \, we propose a Gaussian process model using spectral graph wavelets\, whi
 ch can naturally aggregate neighbourhood information at different scales. 
 Through maximum likelihood optimisation of the model hyperparameters\, the
  wavelets automatically adapt to the different frequencies in the data\, a
 nd as a result our model goes beyond capturing low frequency information. 
 We achieve scalability to larger graphs by using a spectrum-adaptive polyn
 omial approximation of the filter function\, which is designed to yield a 
 low approximation error in dense areas of the graph spectrum. Synthetic an
 d real-world experiments demonstrate the ability of our model to infer sca
 les accurately and produce competitive performances against state-of-the-a
 rt models in graph-based learning tasks.
LOCATION:Zoom
END:VEVENT
END:VCALENDAR