On the Analysis and Interpretation of Inhomogeneous Quadratic Forms as Receptive Fields

By Pietro Berkes, Laurenz Wiskott. 2006

55 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Cognitive Neuroscience, Model Architectures, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Cognitive Neuroscience

In Model Architectures

In Signal Processing

Citations in the corpus

Visualizing and Understanding Convolutional Networks (2013, methodology): “Motivated by the latter’s short-coming, (Le et al., 2010) (extending an idea by (Berkes & Wiskott, 2006)) show how the Hessian of a given unit may be computed numerically around the optimal response, giving some insight into invariances.”
What do we need to build explainable AI systems for the medical domain? (2017, methodology): “Activation maximization can be used as an analysis framework that searches for an input pattern to produce a maximum model response for a specific quantity of interest (Berkes and Wiskott, 2006; Simonyan and Zisserman, 2014).”
Methods for interpreting and understanding deep neural networks (2017, background): “Activation maximization is an analysis framework that searches for an input pattern that produces a maximum model response for a quantity of interest [9, 17, 44].”
Understanding Deep Features with Computer-Generated Imagery (2015, background): “In addition to the prior work to visualize learned CNN filters [5, 11, 12, 32, 41, 42], there has been work to visualize hand-designed [37] and deep [25] features.”
Causability and explainability of artificial intelligence in medicine (2019, background): “Activation maximization (Montavon et al., 2017) identifies input patterns that lead to maximal activations relating to specific classes in the output layer (Berkes & Wiskott, 2006; Simonyan & Zisserman, 2014).”
Hierarchical Convolutional Deep Learning in Computer Vision (2013, background): “[71] (extending an idea by Berkes and Wiskott [9]) show how the Hessian of a given unit may be computed numerically around the optimal response, giving some insight into invariances.”
VisualBackProp: efficient visualization of CNNs (2018, unknown): “This method extends previous approaches [26].”
VisualBackProp: visualizing CNNs for autonomous driving (2016, methodology): “This method extends previous approaches (Berkes & Wiskott, 2006).”
Slow Feature Analysis for Human Action Recognition (2012, methodology): “11 shows the optimal excitatory stimulus and the optimal inhibitory stimulus of the first six slow feature functions by using the method in [46].”
Slow feature analysis yields a rich repertoire of complex cell properties. (2005, methodology): “However, typically only a few subunits (those corresponding to the largest eigenvalues) are structured like simple cells (see also Berkes & Wiskott, 2005 ).”