BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:Algebraic structure in models and data
X-WR-TIMEZONE:Eastern Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20261016T230645Z
UID:tag:localist.com\,2008:EventInstance_39136823439873
DTSTART:20220217T140000Z
DTEND:20220217T151500Z
DESCRIPTION:Data contains algebraic structure\, which can be leveraged to\n
 understand the systems it describes. I will discuss my work to extend\nlin
 ear algebra to the multilinear setting of tensors. We will see how\nmultil
 inear algebra gives interpretable insights from multimodal data\nand how i
 t enables the solving of an inverse problem from analysis of\ntime series 
 data.\n\nAlgebraic structure can be used to interrogate models. I will des
 cribe\nmy work to establish a correspondence between notions from invarian
 t\ntheory and properties of statistical models with a group structure on\n
 their parameter space. This correspondence enables understanding the\npara
 meters in existing models that explain observed data\, as well as\nconstru
 ction of new models that capture otherwise inaccessible\ncharacteristics o
 f\, for example\, biological systems.
LOCATION:Maxwell Dworkin G115
SUMMARY:Algebraic structure in models and data
URL;VALUE=URI:https://events.seas.harvard.edu/event/algebraic_structure_in_
 models_and_data
CATEGORIES:Colloquia / Seminar / Lecture
END:VEVENT
END:VCALENDAR