BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//pyconde-pydata-2026//speaker//NHKXFV
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-LPUC9T@pretalx.com
DTSTART;TZID=CET:20260415T132500
DTEND;TZID=CET:20260415T141000
DESCRIPTION:Multimodal learning - systems that combine vision\, language\, 
 audio\, and other sensory inputs—has moved from a niche research topic t
 o a central paradigm in modern machine learning. Today’s most influentia
 l models no longer operate on a single modality but instead learn rich rep
 resentations by combining language with images\, videos\, sound. This shif
 t has fundamentally changed how we build\, train\, and evaluate current ma
 chine learning systems. Python has played a decisive role in this transfor
 mation. Acting as a unifying layer across modalities\, Python enabled rese
 archers and practitioners to seamlessly combine computer vision\, natural 
 language processing\, and speech within a single ecosystem. Python-based f
 rameworks lowered the barriers between research communities\, and accelera
 ted the rise of large-scale\, weakly supervised\, and foundation models. H
 owever\, this success has also introduced new challenges. The ease of expe
 rimentation masks growing issues around scalability\, reproducibility\, an
 d evaluation. Multimodal systems increasingly depend on complex Python-bas
 ed stacks whose abstractions can obscure underlying assumptions and costs.
 \n...
DTSTAMP:20260412T141734Z
LOCATION:Merck Plenary (Spectrum)  [1st Floor]
SUMMARY:The Multimodal Era of Machine Learning (and How Python Made It Poss
 ible) - Hilde Kühne
URL:https://pretalx.com/pyconde-pydata-2026/talk/LPUC9T/
END:VEVENT
END:VCALENDAR