BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//euroscipy-2026//speaker//MJKEJJ
BEGIN:VTIMEZONE
TZID:Poland
BEGIN:DAYLIGHT
DTSTART:20250721T000000
TZNAME:CEST
TZOFFSETFROM:+0200
TZOFFSETTO:+0200
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20251026T030000
RDATE:20261025T030000
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20260329T030000
RDATE:20270328T030000
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
SUMMARY:How to use skrub Data Ops in practice - Guillaume Lemaitre\, Jér
 ôme Dockès
DTSTART;TZID=Poland:20260721T093000
DTEND;TZID=Poland:20260721T100000
DTSTAMP:20260913T112147Z
UID:pretalx-euroscipy-2026-MLEJZS@pretalx.com
DESCRIPTION:Skrub is a package that eases preparing dataframes so they can
  be used in machine-learning tasks. In practice\, data can be spread over 
 multiple tables\, represent various types of information (tabular\, textua
 l\, graphical)\, or be stored on external database systems rather than dat
 aframes. \n\nSkrub Data Ops help with constructing versatile pipelines tha
 t can handle this variety of scenarios\, while at the same time avoiding d
 ata leakage and allowing to build rich hyper-parameter grids that can be e
 xplored to maximize the performance of the final machine learning model. 
 \n\nIn this talk\, we give a brief introduction of the Data Ops framework 
 before presenting three separate use cases highlighting their versatility:
  a traditional machine learning pipeline that uses Optuna to perform hyper
 -parameter tuning\, a pipeline that trains on data stored in a relational 
 database rather than a dataframe\, and an image classification task with P
 ytorch. \n\nBy the end of the talk\, attendees will learn about the skrub 
 Data Ops\,  their main features and how they can be used successfully in d
 ifferent practical scenarios.
LOCATION:Room 1.38 (Ground Floor\, Turing)
URL:https://pretalx.com/euroscipy-2026/talk/MLEJZS/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Deal with imbalanced classification using scikit-learn - Guillaume
  Lemaitre\, Anne Beyer
DTSTART;TZID=Poland:20260722T090000
DTEND;TZID=Poland:20260722T103000
DTSTAMP:20260913T112147Z
UID:pretalx-euroscipy-2026-D9FJAC@pretalx.com
DESCRIPTION:Class imbalance is a common challenge in real-world machine le
 arning. This course explores why standard approaches fail and how to build
  reliable classifiers using scikit-learn's calibration and threshold-tunin
 g tools.\n\nWe cover practical solutions including resampling strategies\,
  probabilistic calibration with `CalibratedClassifierCV`\, and decision th
 reshold optimization using `TunedThresholdClassifierCV`. You'll learn to e
 valuate models appropriately with calibration curves and confusion matrice
 s.\n\nThe course also addresses prevalence shift or in other words when yo
 ur training data doesn't reflect the target population. We demonstrate wei
 ght-based training corrections and post-hoc probability adjustments applic
 able to any binary classifier.\n\nThe material is available here: https://
 github.com/probabl-ai/calibration-cost-sensitive-learning
LOCATION:Room 1.38 (Ground Floor\, Turing)
URL:https://pretalx.com/euroscipy-2026/talk/D9FJAC/
END:VEVENT
END:VCALENDAR