BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//pyconde-pydata-2026//speaker//VX8JGC
BEGIN:VTIMEZONE
TZID:CET
BEGIN:STANDARD
DTSTART:20001029T040000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-pyconde-pydata-2026-MJTQEJ@pretalx.com
DTSTART;TZID=CET:20260416T154500
DTEND;TZID=CET:20260416T161500
DESCRIPTION:What does it take to build an AI that learns to play strategy g
 ames from scratch? Over the past year\, I chose to explore this question o
 ut of personal fascination with game AI — as a seminar project for colle
 ge\, but really as a hobby. The result was a complete reinforcement learni
 ng environment for Antiyoy\, a turn-based strategy game played on hexagona
 l grids.\n\nThe journey raised intriguing challenges: How do you represent
  hexagonal game boards for neural networks? What do you do when your AI ha
 s over 4\,000 possible actions to choose from? How do you design rewards t
 hat teach strategy rather than just reward flailing in the right direction
 ? This talk shares how these problems were approached using Python's moder
 n ML ecosystem—Gymnasium\, PyTorch\, and PPO training—ultimately produ
 cing an agent that wins nine out of ten games against a random opponent. W
 hether that qualifies as "strategic play" is a question the agent and I st
 ill disagree on.\n\nWhether you're curious about building custom RL enviro
 nments\, interested in game AI\, or just wondering what reinforcement lear
 ning actually looks like when it half-works\, you'll leave with practical 
 insights and a healthy dose of realistic expectations.
DTSTAMP:20260412T141743Z
LOCATION:Ferrum [2nd Floor]
SUMMARY:Mastering the Hex: A Case Study in Reinforcement Learning for Strat
 egy Games - Simon Hedrich
URL:https://pretalx.com/pyconde-pydata-2026/talk/MJTQEJ/
END:VEVENT
END:VCALENDAR