Visual Explanation of Evidence with Additive Classifiers

By Brett Poulin, Roman Eisner, Duane Szafron, Paul Lu, R. Greiner, David S. Wishart, Alona Fyshe, Brandon Pearcy, Cam Macdonell, John Anvik. 2006

135 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Safety, Robustness & Interpretability

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Safety, Robustness & Interpretability

Citations in the corpus

What do we need to build explainable AI systems for the medical domain? (2017, methodology): “A further example for ante-hoc methods can be seen in Poulin et al. (2006), where they described a framework for visually explaining the decisions of any classiﬁer that is formulated as an additive model and showed how to implement this framework in the context of three models: na¨ıve Bayes, linear…”
Explaining nonlinear classification decisions with deep Taylor decomposition (2015, background): “The relevance of each input variable can be identified as elements of the sum [22]:”
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 Explanations in AI (2018, background): “…we need to determine what sort of records algorithmic systems must retain in order to allow for contesting and post-hoc auditing of abnormal events (Mittelstadt, 2016; Sandvig et al., 2014), for instance through identification of classification errors or inaccurate input data (Poulin et al., 2006).”
An Efficient Explanation of Individual Classifications using Game Theory (2010, methodology): “The ExplainD framework (Szafron et al., 2006) provides explanations for additive classifiers in the form of contributions.”
Explanation and Reliability of Individual Predictions (2013, background): “Other, more complex and often better performing models are non-transparent and require additional explanation.”
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”
Deep Taylor Decomposition of Neural Networks (2016, methodology): “…we can identify two distinct approaches: sensitivity analyses (e.g. Gevrey et al., 2003), that find the input variables that causes the output to vary locally, and decompositions (e.g. Poulin et al., 2006) that seek to redistribute the function output on the input variables in a meaningful way.”
Explainable Artificial Intelligence: a Systematic Review (2020, methodology): “Similarly, ExplainD [180] was designed to explain predictions made by classifiers that use additive evidence, such as linear SVMs and regressors.”
Explaining machine learning models in sales predictions (2017, methodology): “In the ExplainD framework (Poulin et al., 2006), weights of evidence and visualizations similar to the EXPLAIN method are used, but the explanation approach is limited to (linear) additive models, while EXPLAIN and IME can be used for all probabilistic models.”
Explaining Deep Neural Networks and Beyond: A Review of Methods and Applications (2021, background): “, neural networks with an explicit top-level sumpooling structure [25], [28], [111], [143], [206].”