BEGIN:VCALENDAR
PRODID:-//eluceo/ical//2.0/EN
VERSION:2.0
CALSCALE:GREGORIAN
BEGIN:VEVENT
UID:www.tcs.tifr.res.in/event/633
DTSTAMP:20230914T125932Z
SUMMARY:Algorithms\, Games and Evolution
DESCRIPTION:Speaker: Umesh V. Vazirani (University of California at Berkele
 y\nDepartment of Computer Science\nBerkeley\, CA 94720\nUnited States of A
 merica)\n\nAbstract: \nAbstract: Even the most seasoned students of evolu
 tion\, starting with Darwin himself\, have occasionally expressed amazemen
 t that the mechanism of natural selection has produced all of life as we s
 ee it around us. Or stated from a computational perspective\, what algori
 thm could possibly achieve Nature's solution to the problems of robotics\
 , vision and theorem proving in about 10^12 generations (steps)? We demons
 trate that in the regime of weak selection\, the standard equations of pop
 ulation genetics describing natural selection in the presence of recombina
 tion become identical to those of a repeated game between genes played a
 ccording to multiplicative weight updates (MWUA)\, an algorithm known to
  be surprisingly powerful and versatile. A dual description of MWUA shows 
 that it maximizes a trade-off between cumulative performance and entropy\,
  which suggests a new view on the maintenance of diversity in evolution (
 based on joint work with Erick Chastain\, Adi Livnat and Christos Papadimi
 triou).\n \n
URL:https://www.tcs.tifr.res.in/web/events/633
DTSTART;TZID=Asia/Kolkata:20151116T113000
DTEND;TZID=Asia/Kolkata:20151116T123000
LOCATION:A-212 (STCS Seminar Room)
END:VEVENT
END:VCALENDAR