BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:BSU Seminar: "Probabilistic approaches for optimal sequential feat
 ure acquisition" - Dr Christopher Yau\, University of Birmingham and The A
 lan Turing Institute
DTSTART:20190128T140000Z
DTEND:20190128T150000Z
UID:TALK115780@talks.cam.ac.uk
CONTACT:Alison Quenault
DESCRIPTION:In many real-world data-driven problems\, the evidence to supp
 ort decisions is gathered sequentially and not all measurements are availa
 ble immediately. For instance\, in medical diagnosis\, a clinician may ord
 er a series of tests and\, based on their outcomes\, order further tests t
 o determine the disease state of a patient. Each patient disease classific
 ation is therefore associated with a “diagnostic trajectory” charting 
 the series of measurements that were recorded to reach their diagnostic co
 nclusion. Whilst much focus in recent developments in medical artificial i
 ntelligence have focused on predictive modelling and automation of decisio
 n making processes\, in the context of complete data\, there has been rela
 tively less attention paid to the sequential data acquisition processes th
 at operate in reality. In this talk\, I will describe a novel and generic 
 Bayesian optimisation approach that we have developed to integrate sequent
 ial feature acquisition processes into predictive models. I will demonstra
 te the optimality properties of this algorithm and illustrate its use on c
 ritical care data from the publicly available MIMIC-III database. Finally 
 I discuss how the framework can be used to construct personal machine lear
 ning-based diagnostic tools.
LOCATION:Large Seminar Room\, 1st Floor\, Institute of Public Health\, Uni
 versity Forvie Site\, Robinson Way\, Cambridge
END:VEVENT
END:VCALENDAR