Neural Information Processing



New paper accepted at the "ICML 2021 Workshop on Theoretic Foundation, Criticism, and Application Trend of Explainable AI"

Title: "How Well do Feature Visualizations Support Causal Understanding of CNN Activations?" by Roland S. Zimmermann, Judy Borowski, Robert Geirhos, Matthias Bethge, Thomas S. A. Wallis, & Wieland Brendel

To see the whole article please visit our publications page.
Link to the workshop: