BEGIN:VCALENDAR VERSION:2.0 PRODID:researchseminars.org CALSCALE:GREGORIAN X-WR-CALNAME:researchseminars.org BEGIN:VEVENT SUMMARY:Vladimir Vovk (University of London\, U.K.) DTSTART:20210315T093000Z DTEND:20210315T101500Z DTSTAMP:20260421T174209Z UID:BPS/23 DESCRIPTION:Title: Mar tingales in the foundations of statistics\nby Vladimir Vovk (Universit y of London\, U.K.) as part of Bangalore Probability Seminar\n\n\nAbstract \nhttps://www.isibang.ac.in/~statmath/pcm2020/\n\nMartingales in the found ations of statistics\nDate: March 15th\, 2021\nTime: 15:00-15:45 IST\n\n\n Abstract: Traditional methods of testing statistical hypotheses have been developed for the batch setting\, in the terminology of machine learning: given a batch of data\, statisticians typically compute measures of disagr eement\, such as p-values or Bayes factors\, between a null hypothesis and the data. An alternative that is popular in machine learning is the onlin e setting\, in which the items of data (observations ) keep arriving seque ntially. In this introductory lecture I will explain the role of martingal es\, in the form of test martingales\, in online hypothesis testing and di scuss their applications in the foundations of probability and statistics. \n\nMultiple hypothesis testing with e-values\nDate: March 15th\, 2021\nTi me: 16:00-16:45 IST\n\nAbstract: It is interesting that test martingales d o not trivialize in the case of only one observation. In fact\, they provi de a useful alternative\, sometimes called e-values\, to the standard stat istical notion of p-values. The most important mathematical advantage of e -values over p-values is that the average of e-values is always an e-value . This property is valuable in multiple hypothesis testing\, which will be the topic of this lecture.\n\nConformal prediction\nDate: March 19th\, 20 21\nTime: 15:00-15:45 IST\n\nAbstract: Mainstream machine learning\, despi te its recent successes\, has a serious drawback: while its state-of-the-a rt algorithms often produce excellent predictions\, they do not provide me asures of their accuracy and reliability that would be both practically us eful and provably valid. On the other hand\, such measures are commonplace in statistics. Conformal prediction adapts rank tests\, popular in nonpar ametric statistics\, to testing the IID assumption (the observations being independent and identically distributed)\, which is the standard assumpti on made in machine learning. This gives us practical measures\, provably v alid under the IID assumption\, of the accuracy and reliability of predict ions produced by traditional and recent machine-learning algorithms. In th is lecture I will give a brief review of conformal prediction.\n\nConforma l hypothesis testing\nDate: March 19th\, 2021\nTime: 16:00-16:45 IST\n\nAb stract: An interesting application of conformal prediction is the existenc e of exchangeability martingales\, i.e.\, random processes that are test m artingales under any exchangeable probability measure. In particular\, the y are martingales whenever the observations are IID. The topics of this la st lecture in this series will be the construction of exchangeability mart ingales and their use for different kinds of change detection\, including detecting a point at which the IID assumption becomes violated and detecti ng concept shift.\n LOCATION:https://researchseminars.org/talk/BPS/23/ END:VEVENT END:VCALENDAR