Deep learning has enabled major advances in machine learning. However, the deployment of deep learning frameworks in settings that are safety-critical or that impact society requires their decision-making to be explainable. This is fundamental for building trustworthy and user-oriented machine learning models. The aim of this workshop is to generate awareness around explainability in machine learning which is a topic of growing interest. Furthermore, we aim to encourage interdisciplinary interaction and collaboration between researchers from the University of Tübingen and other international institutions that work on different aspects of explainability, in particular in the context of computer vision.

Our workshop will consider research on approaches to explainability and on applications of explainability. Topics of interest will include (but will not be limited to):

• Computer vision (and visual explanations)
• Natural language processing
• Explanations for medical applications
• Autonomous driving
• Fairness
• Causality

Program

The workshop will contain both keynote talks from known researchers in the field as well as invited talks and spotlight presentations of recent advancements in the field of explainability.

March		28th					March		29th
9:00	-	9:20		Opening remarks			9:00	-	9:50		Invited talk: Bernt Schiele
				"Interpretability of Deep Learning Medical Image Computing Technologies: Insights, Challenges and Opportunities"							"Interpretability for Deep Learning in Computer Vision"
9:20	-	10:10		Invited talk: Wojciech Samek			9:50	-	10:15		Spotlight: Moritz Böhle
				"Concept-Level Explainable AI"
10:10	-	10:40		Coffee break			10:15	-	10:40		Coffee break
10:40	-	11:05		Spotlight: Diego Marcos			10:40	-	11:05		Spotlight: Leon Sixt
11:05	-	11:30		Spotlight: Maximilian Augustin			11:05	-	11:30		Spotlight: Seong Joon Oh
11:30	-	12:20		Invited talk: Ulrike von Luxburg and Sebastian Bordt			11:30	-	12:20		Invited talk: Mara Graziani
				"Explanation and Regulation"							"Concept discovery and Dataset exploration with Singular Value Decomposition"
12:20	-	13:20		Lunch break			12:20	-	13:20		Lunch break
13:20	-	13:45		Spotlight: Stephan Alaniz			13:20	-	13:45		Spotlight: Sara Blanco
							13:45	-	14:10		Spotlight: Julius von Kügelgen
13:45	-	14:35		Invited talk: Ruth Fong			14:10	-	15:00		Invited talk: Mauricio Reyes
				"Directions in Interpretability"							"Interpretability of Deep Learning Medical Image Computing Technologies: Insights, Challenges and Opportunities"
14:35	-	15:05		Coffee break			15:00	-	15:15		Closing remarks
15:05	-	15:30		Spotlight: Katrin Renz
15:30	-	15:55		Spotlight: Roland Zimmermann
16:00	-	16:50		Invited talk (hybrid): Trevor Darrell and Lisa Dunlap
				"Moving from Explainable models to Advisable models"
17:00	-	18:00		Panel discussion

Venue

The workshop will take place at the Alte Aula in Tübingen, 28-29th March 2023. Alte Aula is the old auditorium, a historic building whose construction dates back to the 16th century.

Participants will also have the chance to explore the old university town of Tübingen. The city combines the ancient medieval atmosphere with the vibrant life of a cosmopolitan student town. You can get lost in the narrow alleys and timber-framed houses that lead to the 500-years old castle, take a boat trip in the famous "Stocherkahn" or enjoy the tasteful Swabian cuisine. For more information on the city, visit www.tuebingen.de.

Attending

There is no registration fee, however the venue has limited capacity. If you would like to attend, please register by February 28th using this form. Sorry, registrations are closed now!

Childcare funds

The workshop can offer limited childcare grants. Please indicate if you require childcare support when registering your interest to attend.

Contact

Do you have any questions? Please send an email to eml-workshop@inf.uni-tuebingen.de.

Speakers

Organizers