BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//NextToU//EN
CALSCALE:GREGORIAN
BEGIN:VEVENT
UID:95e15e6b-8f29-4cc5-b2e7-f8a5ce83480f@nexttoyou
DTSTAMP:20261105T154452Z
DTSTART:20261106T200000Z
DTEND:20261106T210000Z
SUMMARY:School of Mathematics and Statistical Sciences Colloquium talk by 
 Min Wang
LOCATION:Mathematics and General Classrooms Building (EMAGC) 2.410\, 1201 
 W University Dr\, Edinburg\, TX 78539
DESCRIPTION:Title of the talk: Low‐Rank Variational Correction Estimation 
 for Multi‐Source Heterogeneous Quantile Linear Regression Models\n\nAbstr
 act: High-dimensional data arising in genomics\, econometrics\, and clini
 cal medicine often exhibit substantial heterogeneity across multiple sour
 ces. While existing methods address multi-source heterogeneity\, they do 
 not adequately accommodate the combined challenges of high dimensionality
  and between-source heterogeneity. To address this gap\, we propose a sca
 lable Bayesian framework for multi-source heterogeneous quantile regressi
 on with spike-and-slab priors for simultaneous parameter estimation and f
 eature selection. To overcome computational challenges\, we combine mean-
 field variational inference with Laplace approximation and introduce a no
 vel low-rank variational correction strategy that substantially improves 
 approximation accuracy and adaptability in high-dimensional heterogeneous
  settings. This low-rank correction effectively captures the underlying d
 ependence structure\, leading to more robust and efficient inference. For
  model assessment and diagnostic analysis\, we further develop a Bayesian
  score test coupled with local influence analysis. Extensive simulation s
 tudies and an analysis of The Cancer Genome Atlas (TCGA) data from four c
 ancer cohorts (ESCA\, PAAD\, PCPG\, and READ) demonstrate the computation
 al efficiency\, scalability\, and practical utility of the proposed metho
 d in high-dimensional heterogeneous applications. The proposed lo\n\nhttp
 ://next-to-u.app/events/95e15e6b-8f29-4cc5-b2e7-f8a5ce83480f
URL:http://next-to-u.app/events/95e15e6b-8f29-4cc5-b2e7-f8a5ce83480f
END:VEVENT
END:VCALENDAR