BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:PTED: A multi-dimensional two-sample test for scientific inference
  and generative machine learning
DTSTART;VALUE=DATE-TIME:20261028T182200Z
DTEND;VALUE=DATE-TIME:20261028T182300Z
DTSTAMP;VALUE=DATE-TIME:20261122T134654Z
UID:indico-contribution-3713@nyx.physics.mcgill.ca
DESCRIPTION:Speakers: Connor Stone (University of Toronto)\nTwo-sample tes
 ts are widely applicable in inference and generative modelling\, yet users
  frequently fall back on heuristics and visual inspection due to lack of a
 n accessible test that operates in multiple dimensions. I begin with a dis
 cussion on the 7 places two-sample tests are helpful when inferring astron
 omical parameters from data. Then I present Permutation Test using the Ene
 rgy Distance (PTED)\, a Python implementation of a powerful two-sample tes
 t described in Székely and Rizzo [2004]. Because the statistic depends on
  the data only through pairwise distances\, the test applies in high dimen
 sions\, on learned feature representations\, at large or small or imbalanc
 ed sample sizes\, and to any data type on which a distance can be defined.
  I highlight a number of instructive sensitivity tests on PTED alongside o
 ther multi-dimensional two-sample tests for comparison. PTED is the only m
 ethod that is sensitive to all tests.\n\nhttps://nyx.physics.mcgill.ca/eve
 nt/587/contributions/3713/
LOCATION:Online 
URL:https://nyx.physics.mcgill.ca/event/587/contributions/3713/
END:VEVENT
END:VCALENDAR
