BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Bayesian networks for the evaluation of evidence when attributing 
 paintings to painters - Jacob de Zoete (Universiteit van Amsterdam)
DTSTART:20161206T110000Z
DTEND:20161206T120000Z
UID:TALK69477@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:Questions of provenance and attribution have long since motiva
 ted art historical research. Current authentication studies combine tradit
 ional humanities-based methods (for example stylistic analysis\, archival 
 research) with scientific investigation using instrumental analysis techni
 ques like X-ray based methods\, GC-MS\, spectral imaging and metal-isotope
  research. Keeping an overview of information delivered by different speci
 alists and establishing its relative weight is a growing challenge.   &nbs
 p\;  To help clarify complex situations where the relative weight of evide
 nce needs to be established\, the Bayesian framework for interpretation of
  evidence shows great promise. Introducing this mathematical system to cal
 culate the probability of hypotheses based on various pieces of evidence\,
  will strengthen the scientific basis for (art) historical and scientific 
 studies of art. Bayesian networks can accommodate a large variation in dat
 a and can quantify the value of each piece of evidence. Their flexibility 
 allows us to incorporate new evidence and quantify its influence.  &nbsp\;
   <span>In this presentation I will present the first results of a pilot s
 tudy regarding the opportunities and the challenges of implementing Bayesi
 an networks to structure evidence/arguments in painting attribution questi
 ons. This research is based on the painting <i>Sunset at Montmajour</i> th
 at was attributed to Vincent van Gogh in 2013.</span>  <br><br><br><br>
LOCATION:Seminar Room 2\, Newton Institute
END:VEVENT
END:VCALENDAR