ANITI'scientific seminars go on
Let's meet on 4th march From 3pm to 4pm with Stephan Wäldchen about Explaining Neural Network Classifiers: Hurdles and Progress
Le séminaire se déroulera en ligne via zoom
About Stephan Wäldchen
2012: Bachelor Thesis 2012 in Particle Physics at CERN
2015: Master Thesis in Quantum Information Theory at the Group of Jens Eisert,
Berlin
2015-2016: First PhD started with Klaus-Robert Müller at TU Berlin
2017: Second PhD started with Gitta Kutyniok at TU Berlin in Deep Learning Theory
May 2021: Working at the Zuse Institut Berlin with Sebastian Pokutta about
Interpretable AI
| Explaining Neural Network Classifiers: Hurdles and Progress |
| Neural Networks have become the standard tools for high-dimensional decision making, e.g. medical imaging, autonomous driving, playing complex games. Even in high-stakes areas they generally operate as black-box algorithms without a legible decision process. This has birthed the field of explainable artificial intelligence (XAI). The first step for XAI-methods is to discern between the relevant and irrelevant input components for a decision. >> In this talk, we formalise this idea by extending the concept of prime implicants from abductive reasoning to a probabilistic setting. This setting captures what many XAI practitioners intuitively aim for. We show that finding such small implicants, even approximately, is a computationally hard problem. Furthermore, good solutions depend strongly on the underlying probability distribution. We present strategies to overcome both problems and discuss what challenges still remain. |
English 
French