POEM explains models that learn hierarchies of concepts, such as Convolutional Neural Networks that detect shapes and objects in images.

We contribute to this body of work with POEM: a tool that produces pattern-oriented explanations of image classification models. POEM explains models that learn ...

To overcome these limitations, we propose POEM: a system that explains image classifier CNNs using concise and informative patterns of concepts. POEM takes in a ...

Jul 1, 2023 · POEM identifies patterns such as "if sofa then living room", meaning that if an image contains a sofa and the model pays attention to the sofa, ...

We contribute to this line of research by proposing POEM, a framework that produces patterns of concepts to explain image classifier CNNs.

For example, POEM may identify a pattern of the form "if bed then bedroom", indicating that if an image contains a bed and the model pays attention to this ...

In this article, we propose a deep spatial–focal convolutional neural network that encodes the correlations between consecutive focused images that are fed to ...

Title, POEM: Pattern-Oriented Explanations of Convolutional Neural Networks. Publication Type, Journal Article. Year of Publication, 2023.

This project includes the source codes of the POEM pipeline, which can be used to explain image classifier CNN models using patterns of semantic concepts.