Interpreting individual classifications of hierarchical networks

By Will Landecker, Michael D. Thomure, Luís M. A. Bettencourt, Melanie Mitchell, Garrett T. Kenyon, S. Brumby. 2013

74 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Biophysics, Statistical Learning Theory

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 Biophysics

In Statistical Learning Theory

Citations in the corpus

Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges toward Responsible AI (2019, neutral): “Feature relevance explanation Feature Contribution / Statistics: [244] [116, 244]”
A Survey of Methods for Explaining Black Box Models (2018, methodology): “It is worth to mention that perhaps the first step toward LRP was Reference [64], with the Contribution Propagation (CP) method f .”
Explainable Artificial Intelligence: Understanding, Visualizing and Interpreting Deep Learning Models (2017, background): “, it’s representation) [12, 24, 29] as well as techniques for explaining individual predictions [19, 35, 39, 5, 26].”
Methods for interpreting and understanding deep neural networks (2017, background): “[26] to explain the predictions of hierarchical networks, and was also introduced by Bach et al.”
Explaining Recurrent Neural Network Predictions in Sentiment Analysis (2017, background): “Some of these techniques are based on the model’s local gradient information while other methods seek to redistribute the function’s value on the input variables, typically by reverse propagation in the neural network graph (Landecker et al., 2013; Bach et al., 2015; Montavon et al., 2017a).”
Explaining nonlinear classification decisions with deep Taylor decomposition (2015, background): “[36] proposed a contribution propagation approach for hierarchical networks, applying”
RUDDER: Return Decomposition for Delayed Rewards (2018, methodology): “LRP is an extension of the contribution-propagation algorithm [71] based on the contribution approach [94].”
Unmasking Clever Hans predictors and assessing what machines really learn (2019, methodology): “Decomposition methods can be subdivided in three categories: (1) structure approaches, where the function itself has the particular summation structure, and where individual pixel-wise contributions can be identified as the summands of the function [131], (2) analytic approaches, where a local analy”
Explaining Predictions of Non-Linear Classifiers in NLP (2016, background): “In parallel, initiated by the computer vision domain, there is also a trend toward understanding deep learning models through visualization techniques (Erhan et al., 2010; Landecker et al., 2013; Zeiler and Fergus, 2014; Simonyan et al., 2014; Bach et al., 2015; Lapuschkin et al., 2016a) or through ”
Explainable Artificial Intelligence: a Systematic Review (2020, methodology): “Contribution propagation [308] is a per-instance method for hierarchical networks that explain which components of the input were responsible (and to what degree) for its classification.”
Interpreting the Predictions of Complex ML Models by Layer-wise Relevance Propagation (2016, background): “From the multitude of recently proposed methods for explaining predictions, we can identify decomposition approaches [12, 13, 1, 14], which seek to redistribute the value of the prediction function on the input variables, such that the sum of redistributed terms corresponds to the actual function va”
VisualBackProp: visualizing CNNs for autonomous driving (2016, methodology): “An approach similar in spirit to LRP uses contribution propagation [19], where the authors calculate the contributions of high-level features using a previously known technique [20] and then work backwards from the features to the network’s input determining the relative contribution of each node to”
Identifying Individual Facial Expressions by Deconstructing a Neural Network (2016, methodology): “Explanation methods such as [3,26,22,12,13,7,27,25,19] have been able to successfully solve this important open problem.”