BEGIN:VCALENDAR VERSION:2.0 PRODID:-//Talks.cam//talks.cam.ac.uk// X-WR-CALNAME:Talks.cam BEGIN:VEVENT SUMMARY:ENS Special Event - "\;Big data and structured patterns"\; &\; "\;Bayesian networks in crime evidence analysis"\; - Profe ssor Sofia Olhede (University college London) &\; Dr Leila Schneps (Pie rre and Marie Curie University\, Author of 'Maths on Trial') DTSTART:20161124T160000Z DTEND:20161124T170000Z UID:TALK69293@talks.cam.ac.uk CONTACT:Kasia Warburton DESCRIPTION:The Emmy Noether Society are delighted to announce our special penultimate event of term:\n\nThursday\, 24th November\, at the Isaac New ton Institute at 4pm\nA double talk event followed by complimentary wine r eception\n\nProfessor Sofia Olhede (University college London)\n&\nDr Leil a Schneps (Pierre and Marie Curie University\, Author of 'Maths on Trial') \n\nProfessor Sofia Olhede: Big data and structured patterns\n\nScientists try to understand apparent patterns and structures from data. This is com plicated because patterns can be deceiving\; we are apt to see structure i n noise. I will talk about building statistical models to understand data and how we can model phenomena in time and space\, as well as extract mode l parameters for such data from observations.\n\nDr Leila Schneps: Bayesia n networks in crime evidence analysis\n\nOne of the most difficult tasks f or juries in criminal trials is to properly weigh the impact of different pieces or different kinds of evidence which may be interdependent on each other. Studies have shown that there are several fallacies that jury membe rs fall into on a regular basis: overestimation of the weight of DNA evide nce but underestimation of other types of scientific evidence\, double-cou nting evidence\, for instance taking the fact of being arrested as evidenc e in favor of guilt as well as the reasons for which the person\nwas arres ted\, and the famous prosecutor's fallacy or transposed conditional. Bayes ian networks are a powerful tool for assessing the weight of different kin ds of evidence correctly\, taking all their dependencies into account. We believe that it has the potential to become an invaluable tool in investig ating the real implications of crime evidence. In this talk we will explai n how these networks function and give examples of their use in real cases . LOCATION:Isaac Newton Institute for Mathematical Sciences\, 20 Clarkson Ro ad\, Cambridge END:VEVENT END:VCALENDAR