BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Date iCal//NONSGML kigkonsult.se iCalcreator 2.20.2//
METHOD:PUBLISH
X-WR-CALNAME;VALUE=TEXT:Eventi DIAG
BEGIN:VTIMEZONE
TZID:Europe/Paris
BEGIN:STANDARD
DTSTART:20211031T030000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20220327T020000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:calendar.24026.field_data.0@diag.uniroma1.it
DTSTAMP:20240625T113010Z
CREATED:20220201T174759Z
DESCRIPTION:AbstractGraph Neural Networks (GNNs) are a wide class of connec
tionist models for graph processing. Recent studies have linked the expres
sive power of GNNs to the Weisfeiler--Lehman algorithm\, which is a method
of verifying whether two graphs are isomorphic or not. On the other hand\
, it was also observed that the computational power of GNNs is related to
the unfolding trees\, namely trees that can be constructed by visiting t
he graph from a given node. In this paper\, we unify these two theories an
d prove that the Weisfeiler--Lehman test and the unfolding trees induce t
he same equivalence relationship on the graph nodes: such an equivalencee
xactly explains which nodes can or cannot be distinguished by a GNN. Moreo
ver\, it is proved that GNNs can approximate in probability\, up to any
precision\, any function on graphs that respects the above mentio
ned equivalence relationship. These results provide a more comprehensive u
nderstanding of the computational power of GNNs in node classification/reg
ression tasksYou can also join online at https://meet.google.com/mja-amts-
tze?authuser=0&hs=122
DTSTART;TZID=Europe/Paris:20220216T143000
DTEND;TZID=Europe/Paris:20220216T143000
LAST-MODIFIED:20240108T112834Z
LOCATION:Aula B203 DIAG and online
SUMMARY:The expressive power of Graph Neural Networks - A unifying point of
view - Giuseppe Alessio D'Inverno
URL;TYPE=URI:http://diag.uniroma1.it/node/24026
END:VEVENT
END:VCALENDAR