In #AdversarialML, targeted training set attacks are one of the biggest threats to #MachineLearning -- highly effective and hard to detect!
In a #NewPaper at #CCS2022 this week, Zayd Hammoudeh and I show how you can use #InfluenceEstimation to detect, understand, and stop these attacks!
Our methods work against backdoor and poisoning attacks, in vision/test/audio domains, and against adaptive attackers.
#adversarialML #machinelearning #newpaper #CCS2022 #InfluenceEstimation
RT @MultilayerNets
We hope everyone’s enjoying #CCS2022! If you’re an early-career scholar looking at multilayer, multilevel networks, the AccelNet-MultiNet International Exchange Program will help you build new connections across institutions. Details: http://ow.ly/UxwN50LgXYg @IUNetSci @NUnetsi
RT @ccs2022@twitter.com
Invites speaker Elsa Arcaute from @CASAUCL@twitter.com continues our great keynote talks speaking about commuting networks, cycling patterns and socio-demographic flows 🏙🚴🏼🚦
#CCS2022
RT @TiaMazzoli
If you are attending #CCS2022 and are interested in the role of mobility on epidemics, at h18:00 in Sala Magna (Epidemics session), I will talk about a spatial network interpretation of the Reproductive Number that informs surveillance and explains Covid-19 spatial invasion.
RT @F_Sicomoro
Inscríbete en la @ccs2022 , que organiza @IFISC_mallorca del 17 al 21 de octubre. Donde estarán los principales investigadores en #ComplexSystems presentando sus avances. @YR_CSS @sees_lab
@Comsotec @C3UNAM @InstitutoIficc @IF_UNAM
#CCS2022
https://www.ccs2022.org
RT @ccs2022@twitter.com
Another day, another invited speaker at #CCS2022: The leader of the Computational Human Dynamics team, Prof. Marton Karsai, from the Department of Network and Data Science at the Central European University (CEU), Austria.
Deadline for submission: June 7th
@CompSysSoc@twitter.com @YR_CSS@twitter.com