BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Deep Graph Mapper: Seeing Graphs through the Neural Lens - Cristia
 n Bodnar and Cătălina Cangea
DTSTART:20200303T131500Z
DTEND:20200303T141500Z
UID:TALK139426@talks.cam.ac.uk
CONTACT:Mateja Jamnik
DESCRIPTION:Recent advancements in graph representation learning have led 
 to the emergence of condensed encodings that capture the main properties o
 f a graph. However\, even though these abstract representations are powerf
 ul for downstream tasks\, they are not equally suitable for visualisation 
 purposes\, which we believe to be an important part of the research proces
 s---they do not only help discern the structure of complex graphs\, but\, 
 perhaps most essentially\, provide a means of understanding the models app
 lied to them for solving various tasks.\n\nIn this talk\, we will present 
 our recent work that merges Mapper\, an algorithm from the field of Topolo
 gical Data Analysis\, with the expressive power of graph neural networks t
 o produce hierarchical\, topologically-grounded visualisations of graphs. 
 We further demonstrate the suitability of Mapper as a topological framewor
 k for graph pooling by showing an equivalence with the DiffPool and minCUT
  pooling operators. Building upon this framework\, we introduce a novel po
 oling algorithm based on PageRank\, which obtains competitive results with
  state-of-the-art methods on graph classification benchmarks.\n\nThe "pape
 r":https://arxiv.org/abs/2002.03864\, currently under review for ICML'20\,
  and "code":https://github.com/crisbodnar/dgm are also available.
LOCATION:SS03\, Computer Laboratory\, William Gates Building
END:VEVENT
END:VCALENDAR