BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Halıcıoğlu Data Science Institute - UC San Diego - ECPv6.16.2//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:Halıcıoğlu Data Science Institute - UC San Diego
X-ORIGINAL-URL:https://datascience.ucsd.edu
X-WR-CALDESC:Events for Halıcıoğlu Data Science Institute - UC San Diego
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:America/Denver
BEGIN:DAYLIGHT
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
TZNAME:MDT
DTSTART:20250309T090000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
TZNAME:MST
DTSTART:20251102T080000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
TZNAME:MDT
DTSTART:20260308T090000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
TZNAME:MST
DTSTART:20261101T080000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
TZNAME:MDT
DTSTART:20270314T090000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
TZNAME:MST
DTSTART:20271107T080000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=America/Denver:20260603T130000
DTEND;TZID=America/Denver:20260603T170000
DTSTAMP:20260605T072632
CREATED:20260602T220949Z
LAST-MODIFIED:20260602T220949Z
UID:10000563-1780491600-1780506000@datascience.ucsd.edu
SUMMARY:CVPR Tutorial: Principled Interpretability in Vision Models
DESCRIPTION:CVPR 2026 Tutorial\nIEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2026)\nDate: June 3\, 2026 (start at 1 pm) ·  Location: Mile High 3C\, Denver\, CO\, USA \n\nOverview\nAs deep learning systems are increasingly deployed in high-stakes applications\, understanding their behavior is critical for ensuring trust and safety. Interpretability provides essential tools to explain\, debug\, and improve these models. However\, the field remains fragmented\, spanning a wide range of methods and assumptions\, while lacking standardized evaluation protocols. \n\nThis tutorial aims to provide a unified overview of interpretability in deep learning – bridging post-hoc mechanistic understanding and methods to design inherently interpretable deep learning models.\nBy the end of this tutorial\, attendees will gain a solid understanding of modern interpretability methods for deep learning models\, how to rigorously evaluate them\, and open research directions in this critical area.\n\n\nTutorial Outline\n\nPost-hoc mechanistic interpretability: Methods that analyze model internals at different levels of granularity (neurons\, layers\, circuits)\, with strengths and limitations.\nFaithfulness and reliability evaluation: Protocols and standardized metrics for assessing interpretability methods and producing actionable explanations.\nInterpretable DNN models by design: Concept bottleneck models and related approaches that align internal representations with human-understandable concepts.\nApplications: Debugging\, model editing\, and safety auditing in practical settings.\n\n\nAgenda (New!)\nOur tutorial is on Wednesday\, June 3rd afternoon session\, 1-5 pm: \n\nPart 1: Introduction & Backgrounds\nPart 2: Post-hoc model-level interpretability\nPart 3: Faithful and reliability evaluations\nPart 4: Interpretable DNN models by design\nPart 5: Applications\, Demos\, and Technical Q/A\n\nIntended Audience\nThis tutorial is intended for researchers and practitioners working on computer vision and modern deep learning systems\, as well as graduate students entering interpretability research. No prior experience in interpretability is required. \nMaterials\nPlease stay tuned on the Tutorial schedule and Agenda! Slides and supplementary materials will be posted here after the tutorial. \n\nOrganizers and Contact\n✉️ Lily Weng (lweng@ucsd.edu)\, Tuomas Oikarinen (toikarinen@ucsd.edu)\, Ge Yan (geyan@ucsd.edu)\, Akshay Kulkarni (a2kulkarni@ucsd.edu)
URL:https://datascience.ucsd.edu/event/cvpr-tutorial-principled-interpretability-in-vision-models/
LOCATION:Colorado Convention Center
CATEGORIES:Workshops
END:VEVENT
END:VCALENDAR