Repository logo

Infoscience

  • English
  • French
Log In
Logo EPFL, École polytechnique fédérale de Lausanne

Infoscience

  • English
  • French
Log In
  1. Home
  2. Academic and Research Output
  3. Conferences, Workshops, Symposiums, and Seminars
  4. Minimizing Regret in Unconstrained Online Convex Optimization
 
conference paper

Minimizing Regret in Unconstrained Online Convex Optimization

Tatarenko, Tatiana
•
Kamgarpour, Maryam  
June 2018
2018 European Control Conference (ECC)
2018 17th European Control Conference (ECC)

We consider online convex optimizations in the bandit setting. The decision maker does not know the time- varying cost functions, or their gradients. At each time step, she observes the value of the cost function for her chosen action. The objective is to minimize the regret, that is, the difference between the sum of the costs she accumulates and that of the optimal action computable had she known the cost functions a priori. We present a novel algorithm in order to minimize the regret in an unconstrained action space. Our algorithm hinges on the idea of introducing randomization to approximate the gradients of the cost functions using only their observed values. We establish an almost sure regret bound for the mean values of actions and an expected regret bound for the actions.

  • Details
  • Metrics
Logo EPFL, École polytechnique fédérale de Lausanne
  • Contact
  • infoscience@epfl.ch

  • Follow us on Facebook
  • Follow us on Instagram
  • Follow us on LinkedIn
  • Follow us on X
  • Follow us on Youtube
AccessibilityLegal noticePrivacy policyCookie settingsEnd User AgreementGet helpFeedback

Infoscience is a service managed and provided by the Library and IT Services of EPFL. © EPFL, tous droits réservés