BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20250713T032905EDT-1749UWIRPG@132.216.98.100
DTSTAMP:20250713T072905Z
DESCRIPTION:Title: Conditional nonparametric variable screening via neural 
 network factor regression.\n\nAbstract: \n\nHigh-dimensional covariates of
 ten admit linear factor structure. To effectively screen correlated covari
 ates in high-dimension\, we propose a conditional variable screening test 
 based on non-parametric regression using neural networks due to their repr
 esentation power. We ask the question whether individual covariates have a
 dditional contributions given the latent factors. Our test statistics are 
 based on the estimated partial derivative of the regression function of th
 e candidate variable for screening and an observable proxy for the latent 
 factors. Hence\, our test reveals how much predictors contribute additiona
 lly to the non-parametric regression after accounting for the latent facto
 rs. Our derivative estimator is the convolution of a deep neural network r
 egression estimator and a smoothing kernel. We demonstrate that when the n
 eural network size diverges with the sample size\, unlike estimating the r
 egression function itself\, it is necessary to smooth the partial derivati
 ve of the neural network estimator to recover the desired convergence rate
  for the derivative. Moreover\, our screening test achieves asymptotic nor
 mality under the null after finely centering our test statistics that make
 s the biases negligible\, as well as consistency for local alternatives un
 der mild conditions. We demonstrate the performance of our test in a simul
 ation study and a real world application.\n\nSpeaker\n\nYue Zhao is an ass
 istant professor at the University of York in the U.K. He received a Bache
 lor’s degree from Stanford University and Ph.D. degrees in physics and sta
 tistics from Princeton University and Cornell University respectively. He 
 then worked as a postdoctoral researcher at ºÃÉ«TV and KU Leuve
 n in Belgium before joining York in 2019. Yue Zhao is interested in the co
 pula method\, high-dimensional statistics\, and applied empirical processe
 s.\n
DTSTART:20241206T203000Z
DTEND:20241206T213000Z
LOCATION:Room 1104\, Burnside Hall\, CA\, QC\, Montreal\, H3A 0B9\, 805 rue
  Sherbrooke Ouest
SUMMARY:Yue Zhao (University of York)
URL:/mathstat/channels/event/yue-zhao-university-york-
 361744
END:VEVENT
END:VCALENDAR