BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Representation Learning for Latent Risk Dynamics in Online Gamblin
 g
DTSTART:20260513T130000Z
DTEND:20260513T140000Z
DTSTAMP:20260516T082900Z
UID:indico-event-9717@indico.fysik.su.se
DESCRIPTION:Speakers: Sam Andersson (Karolinska Institutet)\n\nMany digita
 l systems generate behavioural time series that are heavy-tailed\, bursty\
 , and only selectively labelled\, making it difficult to define a stable r
 isk state from routine data. In this talk\, I use online gambling as a cas
 e study and present a modelling framework that combines tail-focused excee
 dance features\, representation learning\, and dynamic Bayesian state mode
 lling to derive an auditable operational proxy definition of gambling-rela
 ted risk. The aim is not clinical diagnosis\, but a practical and transpar
 ent way to represent time-varying behavioural regimes and evaluate early-w
 arning policies under realistic review-capacity constraints.\nZoom link: h
 ttps://stockholmuniversity.zoom.us/j/69445559626\nAbout the speaker: Sam A
 ndersson is a doctoral researcher at Karolinska Institutet focusing on com
 putational statistics\, probabilistic modelling and machine learning for l
 ongitudinal behaviour data.\n\nhttps://indico.fysik.su.se/event/9717/
LOCATION:AlbaNova A5:1041 - CoPS grupprum (AlbaNova Main Building)
URL:https://indico.fysik.su.se/event/9717/
END:VEVENT
END:VCALENDAR