BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//pyconde-pydata-2025//speaker//RLAACQ
BEGIN:VTIMEZONE
TZID:Europe/Berlin
BEGIN:DAYLIGHT
DTSTART:20240424T000000
TZNAME:CEST
TZOFFSETFROM:+0200
TZOFFSETTO:+0200
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20241027T030000
RDATE:20251026T030000
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20250330T030000
RDATE:20260329T030000
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
SUMMARY:Transformers for Game Log Data - Fabian Hadiji
DTSTART;TZID=Europe/Berlin:20250424T165500
DTEND;TZID=Europe/Berlin:20250424T174000
DTSTAMP:20260816T011826Z
UID:pretalx-pyconde-pydata-2025-9NFHAS@pretalx.com
DESCRIPTION:The Transformer architecture\, originally designed for machine
  translation\, has revolutionized deep learning with applications in natur
 al language processing\, computer vision\, and time series forecasting. Re
 cently\, its capabilities have extended to sequence-to-sequence tasks invo
 lving log data\, such as telemetric event data from computer games.\n\nThi
 s talk demonstrates how to apply a Transformer-based model to game log dat
 a\, showcasing its potential for sequence prediction and representation le
 arning. Attendees will gain insights into implementing a simple Transforme
 r in Python\, optimizing it through hyperparameter tuning\, architectural 
 adjustments\, and defining an appropriate vocabulary for game logs.\n\nRea
 l-world applications\, including clustering and user level predictions\, w
 ill be explored using a dataset of over 175 million events from an MMORPG.
  The talk will conclude with a discussion of the model's performance\, com
 putational requirements\, and future opportunities for this approach.
LOCATION:Palladium
URL:https://pretalx.com/pyconde-pydata-2025/talk/9NFHAS/
END:VEVENT
END:VCALENDAR