Abductive Inference

Text

This workshop explores the concept of abductive inference in relation to machine learning, focusing on the distinction between weak and strong abduction. Building on earlier discussions of weak abduction—understood as the generation of conjectural models by current AI systems—we’ll consider what strong abduction might entail, particularly in light of debates around digital aesthetics, creativity, and politics. Central to our inquiry is the contested role of explanation in AI: while classical statistics emphasized interpretability, contemporary algorithmic approaches prioritize prediction and the construction of new, data-driven realities. This tension invites critical reflection on how different forms of inference (abduction vs. induction) shape epistemological, political, and artistic possibilities—and limitations—in computational culture. The workshop will follow a dynamic and discussion-driven format. Each of the sessions will begin with a short provocation or prompt (approx. 15 minutes), followed by an open discussion (approx. 30 minutes) to collectively unpack the themes and questions raised. Time will be built in between sessions for informal wrap-up conversations, a

Text

Schedule - Intro: Clemens Apprich 9:30 to 10:00 - Session 1: Luke Stark 10.00 to 10:45 - Session 2: Sun-Ha Hong 10:45 to 11:30 - Session 3: Mikkel Rørbo 11:30 to 12:15 - Wrap-up: 12:15 to 13:00 Lunch Break - Session 4: Alex Quicho 14:30 to 15:15 - Session 5: Daniel Susser 15:15 to 16:00 - Session 6: Shaka McGlotten 16:00 to 16:45 - Short break: 16:45 to 17:00 - Wrap-up & next steps: 17:00 to 18:00

Lecturers