no code implementations • 22 Dec 2022 • Simon Odense, Artur d'Avila Garcez
The field of neural-symbolic AI attempts to exploit this asymmetry by combining neural networks and symbolic AI into integrated systems.
no code implementations • 19 Mar 2020 • Simon Odense, Artur d'Avila Garcez
We apply this method to a variety of deep networks and find that in the internal layers we often cannot find rules with a satisfactory complexity and accuracy, suggesting that rule extraction as a general purpose method for explaining the internal logic of a neural network may be impossible.
no code implementations • ICLR 2019 • Simon Odense, Artur d'Avila Garcez
In this paper we examine this question systematically by proposing a knowledge extraction method using \textit{M-of-N} rules which allows us to map the complexity/accuracy landscape of rules describing hidden features in a Convolutional Neural Network (CNN).