BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Isaac Newton Institute Seminar Series
SUMMARY:Decision-making under uncertainty: Using subjectiv
e probabilistic judgements for decision support in
pollinator abundance and food security - Martine
Barons (University of Warwick)
DTSTART;TZID=Europe/London:20190704T121000
DTEND;TZID=Europe/London:20190704T123000
UID:TALK126811AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/126811
DESCRIPTION:Hunger and food poverty is on the increase even in
developed nations like the UK\, USA\, Canada &
\; Australia. With a growing population and a fin
ite planet\, there is urgent need for action\, but
in such a large\, complex system identifying the
most effective action requires decision support.
Food security exists when all people\, at all
times\, have physical and economic access to suff
icient\, safe and nutritious food to meet their di
etary needs and food preferences for an active and
healthy life.
In order to provide decision
support\, it is necessary to elicit probability di
stributions to evaluate subjective expected utilit
y scores associated with ameliorating policies tha
t might be enacted. When the underlying process mo
del is extremely large and complex\, this brings i
ts own peculiar challenges. It is first necessary
to elicit the overall\, agreed structure describin
g in broad terms the underlying nature of the syst
em from representatives of all domain experts acro
ss the system as a whole. We have now shown that t
his can be done formally and consistently with pro
bability models if the elicitations concern the el
icitation of dependences &ndash\; formally termed
irrelevances (Smith\, Barons and Leonelli (2016)).
Within a probability model\, these irrelevance st
atements then transform into assertions about vari
ous conditional independence statements. These\, i
n turn\, can be used to determine how the system c
an be divided up into (conditionally) independent
segments. The quantitative expert judgements asso
ciated with each segment of the process can then b
e delegated to a relevant panel of experts. The im
plicit (albeit virtual) owner of beliefs expressed
in the system will be referred to as the supraBay
esian \, meaning that the decision-making group a
cts as a single person would and it is her coheren
ce that we are concerned with. Under suitable cond
itions it can then be shown that the elicited over
arching structure can compose these judgments toge
ther to form a coherent probabilistic model to sco
re different options available to the user\, terme
d an integrating decisions support system (IDSS).
One element of the overarching food pove
rty models is food supply\, and key to parts of th
is is an abundant and healthy population of pollin
ating insects to pollination services for food. In
2014 the UK government undertook a consultation a
nd produced their pollinator strategy for the next
10 years &ldquo\;to see pollinators thrive\, prov
iding essential pollination services and benefits
for food production\, the wider environment and ev
eryone.&rdquo\; However\, the evidence base on th
e complex system driving pollinator vigour and num
bers is patchy and held in disparate domains of ex
pertise\, making the evaluation of policy options
problematic. In this talk I will describe how we
are in the process of developing an IDSS based on
these theoretical developments\, and how a probabi
listic model for pollinator abundance incorporatin
g structured expert elicitation will then form a s
ub-module of this IDSS for policies relating to ho
usehold food insecurity.
J. Q. Smith\, M.J
. Barons\, and M. Leonelli. Coherent inference for
integrating decision support systems. arXiv\, 201
6. http://arxiv.org/abs/150
7.07394.
Co-authors: Jim Q. Smith\, Man
uele Leonelli
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:INI IT
END:VEVENT
END:VCALENDAR