BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20250705T014122EDT-9218UGb0D5@132.216.98.100
DTSTAMP:20250705T054122Z
DESCRIPTION:Lucy Gao\, PhD\n\nAssistant Professor of Statistics\n	University
  of British Columbia\n\nWHEN: Wednesday\, January 31\, 2024\, from 3:30 to
  4:30 p.m.\n\nWHERE: hybrid | 2001 9IÖÆ×÷³§Ãâ·Ñ College Avenue\, room 1140\; Zoo
 m\n\nNOTE: Dr. Gao will be presenting from British Columbia\n\nAbstract\n
 \n'Double dipping' is the practice of using the same data to fit and valid
 ate a model. Problems typically arise when standard statistical procedures
  are applied in settings involving double dipping. To avoid the challenges
  surrounding double dipping\, a natural approach is to fit a model on one 
 dataset\, and then validate the model on another independent dataset. When
  we only have access to one dataset\, we typically accomplish this via sam
 ple splitting. Unfortunately\, in some problems\, sample splitting is unat
 tractive or impossible. In this talk\, we are motivated by unsupervised pr
 oblems that arise in the analysis of single cell RNA sequencing data\, whe
 re sample splitting does not allow us to avoid double dipping. We first pr
 opose Poisson thinning\, which splits a single observation drawn from a Po
 isson distribution into two independent pseudo-observations. We show that 
 Poisson count splitting allows us to avoid double dipping in unsupervised 
 settings. We next generalize the Poisson thinning framework to a variety o
 f distributions\, and refer to this general framework as 'data thinning'. 
 Data thinning is applicable far beyond the context of single-cell RNA sequ
 encing data\, and is particularly useful for problems where sample splitti
 ng is unattractive or impossible.\n\nSpeaker bio\n\nWebsite Link: https://
 www.lucylgao.com/\n
DTSTART:20240131T203000Z
DTEND:20240131T213000Z
SUMMARY:Data thinning to avoid double dipping 
URL:/epi-biostat-occh/channels/event/data-thinning-avo
 id-double-dipping-353629
END:VEVENT
END:VCALENDAR