Workshop on Logic-based Methods in Machine Learning

Part of the Federated Logic Conference, FLoC 2022
Pre-FLoC workshop, July 31, 2022

Topic and Aims

The far-reaching success of Machine Learning (ML) motivates an ever-growing range of applications. However, the most successful ML models are opaque (“black-box”) because they do not support the explainability or verifiability of their predictions. Recent years have witnessed the emergence of several successful and promising approaches to overcome these limitations with the help of logic-based techniques, including the well-developed technologies of SAT/CP-assisted reasoning and optimization, including the well-developed technologies of SAT-, MaxSAT-, SMT-, and MIPS-solving, constraint optimization, and Model Counting.

This workshop aims at bringing together researchers from various fields that work on SAT-based methods for

• learning interpretable ML models,
• computation of explanations for black-box ML models,
• verification of black-box ML models.

Organizers

Alexey Ignatiev, Monash University, Melbourne, Australia
Stefan Szeider, TU Wien, Vienna, Austria