BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:BSU Seminar: &quot\;Derivative-Based Neural Modelling of Cumulativ
 e Distribution Functions for Survival Analysis&quot\; - Prof Christopher Y
 au\, University of Oxford 
DTSTART:20221110T140000Z
DTEND:20221110T150000Z
UID:TALK189569@talks.cam.ac.uk
CONTACT:Alison Quenault
DESCRIPTION:Survival models — particularly those able to account for pat
 ient comorbidities via competing risks analysis — offer valuable prognos
 tic information to clinicians making critical decisions and represent a gr
 owing area of application for machine learning approaches. However\, curre
 nt methods typically involve restrictive parameterisations\, discretisatio
 n of time or the modelling of only one event cause. In this talk\, I highl
 ight how general cumulative distribution functions can be naturally expres
 sed via neural network-based ordinary differential equations and how this 
 can be utilised in survival analysis. In particular\, we present DeSurv\, 
 a neural derivative-based approach capable of avoiding the aforementioned 
 restrictions and flexibly modelling competing-risk survival data in contin
 uous time. We apply DeSurv to both single-risk and competing-risk syntheti
 c and real-world datasets and obtain results which compare favourably with
  current state-of-the-art models.
LOCATION:Virtual Seminar 
END:VEVENT
END:VCALENDAR