BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CREST - ECPv5.1.3//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:CREST
X-ORIGINAL-URL:https://crest.science
X-WR-CALDESC:Events for CREST
BEGIN:VTIMEZONE
TZID:Europe/Helsinki
BEGIN:DAYLIGHT
TZOFFSETFROM:+0200
TZOFFSETTO:+0300
TZNAME:EEST
DTSTART:20210328T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0300
TZOFFSETTO:+0200
TZNAME:EET
DTSTART:20211031T010000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=Europe/Helsinki:20210607T150000
DTEND;TZID=Europe/Helsinki:20210607T161500
DTSTAMP:20260712T182302
CREATED:20210528T075324Z
LAST-MODIFIED:20210528T075324Z
UID:12751-1623078000-1623082500@crest.science
SUMMARY:Samory Kpotufe (Columbia University) - "Some Recent Insights on Transfer-Learning"
DESCRIPTION:Statistical Seminar: Every Monday at 2:00 pm.\nTime: 3:00 pm – 4:15 pm\nDate: 7th of June 2021\nPlace: Visio\nSamory Kpotufe (Columbia University) – “Some Recent Insights on Transfer-Learning” \nAbstract: A common situation in Machine Learning is one where training data is not fully representative of a target population due to bias in the sampling mechanism or due to prohibitive target sampling costs. In such situations\, we aim to ’transfer’ relevant information from the training data (a.k.a. source data) to the target application. How much information is in the source data about the target application? Would some amount of target data improve transfer? These are all practical questions that depend crucially on ‘how far’ the source domain is from the target. However\, how to properly measure ‘distance’ between source and target domains remains largely unclear.\nIn this talk we will argue that much of the traditional notions of ‘distance’ (e.g. KL-divergence\, extensions of TV such as D_A discrepancy\, density-ratios\, Wasserstein distance) can yield an over-pessimistic picture of transferability. Instead\, we show that some new notions of ‘relative dimension’ between source and target (which we simply term ‘transfer-exponents’) capture a continuum from easy to hard transfer. Transfer-exponents uncover a rich set of situations where transfer is possible even at fast rates; they encode relative benefits of source and target samples\, and have interesting implications for related problems such as ‘multi-task or multi-source learning’.\nIn particular\, in the case of transfer from multiple sources\, we will discuss (if time permits) a strong dichotomy between minimax and adaptive rates: no adaptive procedure exists that can achieve the same rates as minimax (oracle) procedures.\nThe talk is based on earlier work with Guillaume Martinet\, and ongoing work with Steve Hanneke. \n  \nOrganizers:\nCristina BUTUCEA (CREST)\, Alexandre TSYBAKOV (CREST)\, Karim LOUNICI (CMAP) \, Zoltan SZABO (CMAP)\nSponsors:\nCREST-CMAP \n\n
URL:https://crest.science/event/samory-kpotufe-columbia-university-some-recent-insights-on-transfer-learning/
CATEGORIES:Statistics
ATTACH;FMTTYPE=:
END:VEVENT
END:VCALENDAR