BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//euroscipy-2025//speaker//C9GEJ3
BEGIN:VTIMEZONE
TZID:Europe/Warsaw
BEGIN:DAYLIGHT
DTSTART:20240820T000000
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:Recent Developments in Pytensor\, the Successor Package to Theano 
 - Jesse Grabowski\, Ricardo Vieira
DTSTART;TZID=Europe/Warsaw:20250820T133000
DTEND;TZID=Europe/Warsaw:20250820T140000
DTSTAMP:20260911T195013Z
UID:pretalx-euroscipy-2025-LR7S8P@pretalx.com
DESCRIPTION:We present the latest developments in Pytensor\, the successor
  package to Theano. Pytensor is a package for defining\, manipulating\, op
 timizing\, and compiling static computational graphs. We especially focus 
 on full graph-to-graph transformations relevant to the goals of a Bayesian
 /ML workflow. These allow the user to define a single computational graph
 \, which can then be reused in multiple contexts. In the Bayesian workflow
 \, we are able to extract exact expressions for probabilistic inference fr
 om a generative sampling model\, or automatically marginalize discrete ran
 dom variables. In a deep-learning workflow\, we can automatically remove d
 ropout and normalization layers when compiling a prediction function from 
 a training graph\, or replace expensive operations\, such as transformers
 \, with specialized forms at compile time. Finally\, we show how the same 
 machinery leads naturally to transpilation into compiled languages\, via p
 ackages like Numba\, Jax\, and Pytorch
LOCATION:Room 1.38 (Ground Floor)
URL:https://pretalx.com/euroscipy-2025/talk/LR7S8P/
END:VEVENT
END:VCALENDAR