BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Online Combinatorial Auctions
DTSTART:20230112T140000
DTEND:20230112T150000
DTSTAMP:20260405T204904Z
UID:82a28662d965e6e5393030854acda63ea04499b7745dd64c64227509
CATEGORIES:Conferences - Seminars
DESCRIPTION:Dr. Andrés Cristi\nAbstract:\nIn online combinatorial auction
 s\, a set of indivisible items are to be allocated to a set of agents who 
 arrive sequentially. Agents have random valuations for the different subse
 ts of items and the goal is to allocate the items on the fly so as to maxi
 mize the total value of the assignment. A prophet inequality in this setti
 ng refers to the existence of an online algorithm guaranteed to obtain\, i
 n expectation\, a certain fraction of the expected value obtained by an op
 timal solution in hindsight. Because of their connections to ad auctions a
 nd pricing problems in e-commerce platforms\, the study of prophet inequal
 ities for online combinatorial auctions has been an intensive area of rese
 arch in recent years\, and constant factor prophet inequalities are known 
 when the agents’ valuation functions are submodular or fractionally suba
 dditive. In this talk I will present two innovative ideas in the context o
 f algorithm design for online allocation: a sample-based approach to prote
 ct items from being allocated too early\, and fixed-point arguments to pro
 ve the existence of good online policies. I will explain how we used these
  techniques to find the first constant prophet inequality when valuations 
 are subadditive (a.k.a. complement-free)\, resolving a central open proble
 m in the area\, and to find item prices with the best-possible guarantee w
 hen we bound the number of items any agent can get. I will also discuss co
 nnections with my research in sample-based optimal stopping and future dir
 ections in online resource allocation problems.\n\nBio:\nAndrés Cristi is
  a final year PhD student at Universidad de Chile\, advised by José Corre
 a and Paul Dütting. He previously graduated as a Mathematical Engineer an
 d an MS in Operations Management. His research is focused on the interplay
  between optimization and incentives\, this is\, situations where the outc
 ome depends on the actions of strategic agents. He is particularly interes
 ted in allocation problems with a dynamic aspect\, where decisions are mad
 e on the fly\, and data-driven approaches\, in which decisions are directl
 y made using observations rather than distributional assumptions. Modern p
 latforms like routing apps\, online advertisers\, and online marketplaces 
 face these challenges on a daily basis\, and his work is centered on under
 standing the fundamental aspects that drive decision-making in these setti
 ngs.\n\n 
LOCATION:ODY 4 03 https://plan.epfl.ch/?room==ODY%204%2003
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR