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:20230326T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0300
TZOFFSETTO:+0200
TZNAME:EET
DTSTART:20231029T010000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=Europe/Helsinki:20230413T103000
DTEND;TZID=Europe/Helsinki:20230413T113000
DTSTAMP:20260711T141332
CREATED:20230407T042928Z
LAST-MODIFIED:20230407T042928Z
UID:14850-1681381800-1681385400@crest.science
SUMMARY:Lorenzo TRAPANI  (University of Leicester School of Business) - "Changepoint Detection in Large Factor Models"
DESCRIPTION:Finance & Financial Econometrics : \nTime: 10.30 am\nDate: 13th of April 2023\nRoom 3001 \nLorenzo TRAPANI (University of Leicester School of Business) “Changepoint Detection in Large Factor Models” \nAbstract :We study changepoint detection in a large factor model\, proposing both offline and online detection methods. In all cases\, we build on the idea that\, in a factor model\, under the maintained assumption of the homoscedasticity of the common factors\, in the presence of a changepoint the second moment matrix of the common factors changes. Hence\, a high-dimensional problem can be conveniently cast into a low-dimensional problem\, based on checking for breaks in the second moment matrix of the estimated common factors. Our results build on a maximal inequality for the partial sums of the second moment matrix of the estimated common factors. By virtue of this\, we are able to derive strong invariance principles for the partial sums of the estimated second moment matrix of the common factors. Hence\, we propose a family of weighted CUSUM statistics for the offline detection of change points\, including the standardised CUSUM process (for which we derive a Darling-Erdos theorem) and even more heavily weighted statistics. Our tests have power versus breaks occurring also close to the sample endpoints. We further study the asymptotics of the MOSUM process\, and of the maximally selected LR statistic. In addition\, we investigate the problem of sequential\, online detection of changepoints\, proposing a family of detectors which ensure procedure-wise size control and short detection delays; as a by-product\, we derive the limiting distribution of the detection delay.\nJoint work M Barigozzi and H Cho\n\n \n\nOrganizers:\n\nJean-Michel ZAKOIAN (CREST) \nSponsors:\nCREST \n
URL:https://crest.science/event/lorenzo-trapani-university-of-leicester-school-of-business-changepoint-detection-in-large-factor-models/
CATEGORIES:Finance-Insurance,Financial Econometrics
ATTACH;FMTTYPE=:
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Helsinki:20230413T113000
DTEND;TZID=Europe/Helsinki:20230413T123000
DTSTAMP:20260711T141332
CREATED:20230407T043140Z
LAST-MODIFIED:20230407T043140Z
UID:14851-1681385400-1681389000@crest.science
SUMMARY:Yuichi GOTO (Kyushu Univ.) "The Existence and Uniqueness of Lagged Spectrum"
DESCRIPTION:Finance & Financial Econometrics: \nTime: 11.30 am\nDate: 13th of April 2023\nRoom 3001 \nYuichi GOTO (Kyushu Univ.) “The Existence and Uniqueness of Lagged Spectrum” \nAbstract : TCoherence is a similarity measure between two time series and takes the form of the time series extension of Pearson’s correlation. However\, only a linear relationship between two time series can be measured by coherence. In this talk\, we introduce a lagged spectrum in order to measure non-linear relationships and show the existence and uniqueness of a lagged spectrum by decomposing the regression model we consider into orthogonal processes. We also propose a test for the existence of the lagged process and apply our test to economic data.\nJoint work : Xuze Zhang (Univ. of Maryland)\, Benjamin Kedem (Univ. of Maryland) and Shuo Chen (Univ. of Maryland) \nOrganizers:\nJean-Michel ZAKOIAN (CREST) \nSponsors:\nCREST \n
URL:https://crest.science/event/yuichi-goto-kyushu-univ-the-existence-and-uniqueness-of-lagged-spectrum/
CATEGORIES:Finance-Insurance,Financial Econometrics
ATTACH;FMTTYPE=:
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Helsinki:20230413T141500
DTEND;TZID=Europe/Helsinki:20230413T151500
DTSTAMP:20260711T141332
CREATED:20230407T081531Z
LAST-MODIFIED:20230407T081531Z
UID:14857-1681395300-1681398900@crest.science
SUMMARY:Matus TELGARSKY (Université Illinois\, Urbana Champaign) - "Searching for the implicit bias of deep learning"
DESCRIPTION:Statistical Seminar: \nTime: 2:15 pm – 3:15 pm\nDate: 13th of April 2023\nPlace: Room 3001 + ZOOM \n  \nMatus TELGARSKY (Université Illinois\, Urbana Champaign) – “Searching for the implicit bias of deep learning” \n  \nAbstract: \nWhat makes deep learning special — why is it effective in so many settings where other models fail? This talk will present recent progress from three perspectives. The first result is approximation-theoretic: deep networks can easily represent phenomena that require exponentially-sized shallow networks\, decision trees\, and other classical models. Secondly\, I will show that their statistical generalization ability — namely\, their ability to perform well on unseen testing data — is correlated with their prediction margins\, a classical notion of confidence. Finally\, comprising the majority of the talk\, I will discuss the interaction of the preceding two perspectives with optimization: specifically\, how standard descent methods are implicitly biased towards models with good generalization. Here I will present two approaches: the strong implicit bias\, which studies convergence to specific well-structured objects\, and the weak implicit bias\, which merely ensures certain good properties eventually hold\, but has a more flexible proof technique. \n  \n  \nLink :  https://zoom.us/j/91051481144?pwd=alF1cjJUZ0pmUlprRmJjUWRDNU9odz09   \nID de réunion : 910 5148 1144\nCode secret : 590133 \n  \nOrganizers:\nCristina BUTUCEA (CREST)\, Alexandre TSYBAKOV (CREST)\, Karim LOUNICI (CMAP) \, Jaouad MOURTADA (CREST)\nSponsors:\nCREST-CMAP \n
URL:https://crest.science/event/matus-telgarsky-universite-illinois-urbana-champaign-searching-for-the-implicit-bias-of-deep-learning/
CATEGORIES:Statistics
ATTACH;FMTTYPE=:
END:VEVENT
END:VCALENDAR