Optimization for Learning Workshop 2018

General Information

  • Dates: 15-18 January, 2018
  • Location: Auditorio Ninslav Bralic, Campus San Joaquín, Pontificia Universidad Católica de Chile, Santiago, Chile
  • Organizing committee: Roberto Cominetti (UAI), Cristóbal Guzmán (PUC), Jorge Vera (PUC)


The purpose of the workshop is to gather experts from different fields related to optimization, learning and game theory, encouraging interaction among them. We expect to have a limited number of survey talks and open problem sessions, reserving a substantial time for discussions among participants.

In this first edition, the main topic of the workshop will be "Acceleration". In the last years there has been profound progress in extending the applicability of acceleration in first-order methods, impacting both theory and practice of problems described by large datasets. Specifically, topics such as fast methods for convex and nonconvex optimization, variance reduction in stochastic approximation, and statistical stability of iterative methods for statistical learning are of utmost relevance, and will be the focus of this event.



  • Gustavo Angulo (PUC)
  • Javiera Barrera (UAI)
  • Germán Cheuque (PUC)
  • Rodrigo Cofré (PUC)
  • Antoine Hochart (UAI)
  • Guido Lagos (UAI)
  • Max Levill (PUC)
  • Diego Morán (UAI)
  • Héctor Ramirez (U. de Chile)
  • Martín Rapaport (U. de Chile)
  • Javiera Riffo (PUC)
  • Carlos Sing-Long (PUC)
  • Felipe Tobar (U. de Chile)
  • Patricio Ulloa (PUC)
  • Adolfo Villalobos (PUC)


Program Workshop OPT4LRN

The full list of abstracts is available here. Slides for some of the talks are available here. And some photos here.


  • Vicerrectoría de Investigación, Pontificia Universidad Católica de Chile
  • Department of Industrial and Systems Engineering, Pontificia Universidad Católica de Chile
  • Iniciativa Científica Milenio, Instituto Milenio ISCI “Sistemas Complejos de Ingeniería”
  • Instituto de Ingeniería Matemática y Computacional, Pontificia Universidad Católica de Chile
  • FONDECYT projects 11160939, 1171501