BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:CIS - Colloquium -  by Dr. Petar Veličković
DTSTART:20230531T110000
DTEND:20230531T120000
DTSTAMP:20260508T162746Z
UID:952b0d33b7ebd28f7e3520210f2c9015ed3eca8cb2ce04dc927d4e0a
CATEGORIES:Conferences - Seminars
DESCRIPTION:Dr. Petar Veličković\nTitle: Reasoning Algorithmically: from
  Toy Experiments to AGI Modules\n\nAbstract: Neural networks that are able
  to reliably execute algorithmic computation may hold transformative poten
 tial to both machine learning and theoretical computer science. On one han
 d\, they could enable the kind of extrapolative generalisation scarcely se
 en with deep learning models. On another\, they may allow for running clas
 sical algorithms on inputs previously considered inaccessible to them.\nOv
 er the past few years\, the pace of development in this area has gradually
  become intense. As someone who has been very active in its latest incarn
 ation\, I have witnessed these concepts grow from isolated 'toy experiment
 s'\, through NeurIPS spotlights\, all the way to helping detect patterns i
 n complicated mathematical objects (published on the cover of Nature) and 
 supporting the development of generalist reasoning agents.\nIn this talk\,
  I will give my personal account of this journey\, and especially how our 
 own interpretation of this methodology\, and understanding of its potentia
 l\, changed with time. It should be of interest to a general audience inte
 rested in graphs\, (classical) algorithms\, reasoning\, and building intel
 ligent systems.\n\nBio: https://petar-v.com/\nI’m Petar\, a Staff Resear
 ch Scientist at DeepMind\, Affiliated Lecturer at the University of Cambri
 dge\, and an Associate of Clare Hall\, Cambridge. I hold a PhD in Computer
  Science from the University of Cambridge (Trinity College)\, obtained und
 er the supervision of Pietro Liò. My research concerns geometric deep lea
 rning—devising neural network architectures that respect the invariances
  and symmetries in data (a topic I’ve co-written a proto-book about). Fo
 r my contributions\, I am recognised as an ELLIS Scholar in the Geometric 
 Deep Learning Program. Particularly\, I focus on graph representation lear
 ning and its applications in algorithmic reasoning (featured in VentureBea
 t). I am the first author of Graph Attention Networks—a popular convolut
 ional layer for graphs—and Deep Graph Infomax—a popular self-supervise
 d learning pipeline for graphs (featured in ZDNet). My research has been u
 sed in substantially improving travel-time predictions in Google Maps (fea
 tured in the CNBC\, Endgadget\, VentureBeat\, CNET\, the Verge and ZDNet)\
 , and guiding intuition of mathematicians towards new top-tier theorems an
 d conjectures (featured in Nature\, Science\, Quanta Magazine\, New Scient
 ist\, The Independent\, Sky News\, The Sunday Times\, la Repubblica and Th
 e Conversation).\n\nThe Center for Intelligent Systems at EPFL (CIS) is a 
 collaboration among IC\, ENAC\, SB\; SV and STI that brings together resea
 rchers working on different aspects of Intelligent Systems. In June 2020\,
  CIS has launched its CIS Colloquia featuring invited notable speakers.\nM
 ore info\n\nOpen to all – Talk followed by a standing lunch. Registratio
 n required\n 
LOCATION:BM 5202 https://plan.epfl.ch/?room==BM%205202 https://epfl.zoom.u
 s/meeting/register/u5EucuCtpjIuHtfyJqAi2ltyHeaBiM939jl4
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR