Case-based explanation of non-case-based learning methods

By R. Caruana, H. Kangarloo, John David N. Dionisio, U. Sinha, David B. Johnson. 1999

Subfields: 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 Safety, Robustness & Interpretability

Citations in the corpus

The mythos of model interpretability (2016, methodology): “Explanation by Example: One post-hoc mechanism for explaining the decisions of a model might be to report (in addition to predictions) which other examples the model considers to be most similar, a method suggested by Caruana et al. (1999).”
Deep k-Nearest Neighbors: Towards Confident, Interpretable and Robust Deep Learning (2018, background): “Instead, the nearest neighbors central to the DkNN are an instance of explanations by example [57].”
Interpreting Deep Classifier by Visual Distillation of Dark Knowledge (2018, background): “…tree, which is easy to interpret (Craven & Shavlik, 1996), retrieving typical inputs so that one can interpret by examples via hidden activations (Caruana et al., 1999) or inﬂuence functions (Koh & Liang, 2017) and generating artifacts or prototypes to explain model behaviour (Ribeiro et al.,…”
Explaining Explanations in AI (2018, neutral): “Case-based explanation methods (Caruana et al., 1999; Kim et al., 2014) for non-cased based machine learning involve using the trained model as a distance metric to determine which cases in the training data set are most similar to the case or decision to be explained.”
Explaining Black Box Predictions and Unveiling Data Artifacts through Influence Functions (2020, background): “Instead of constructing importance scores over the input texts on which the model makes predictions, such methods rank training examples by their influence on the model’s prediction for the test input (Caruana et al., 1999; Koh and Liang, 2017; Card et al., 2019).”
Efficient Search for Diverse Coherent Explanations (2019, methodology): “Another important class of explanations comes from “case-based reasoning” (Caruana et al., 1999; Kim et al., 2014) in which the method justifies the decision/score made by the algorithm by showing data points from the training set that the algorithm found similar in some sense.”
Interpretable R-CNN (2017, background): “In the context of supervised learning, Lipton [49] investigated the motives for interpretability including trust [38, 62], causality [58], transferability [5] and informativeness [38], and properties of interpretable models including transparency in terms of simulatability and decomposability and po”
Definitions, methods, and applications in interpretable machine learning (2019, background): “These methods focus on explicitly capturing and displaying the interactions learned by a neural network (81, 82), alternative forms of interpretations such as textual explanations (83), identifying inﬂuential data points (69), and analyzing nearest neighbors (84, 85).”
Explanation Mining: Post Hoc Interpretability of Latent Factor Models for Recommendation Systems (2018, background): “user acceptance [11, 19]) as well as constraints such as legal requirements [14].”
Cross-Loss Influence Functions to Explain Deep Network Representations (2020, background): “One promising direction is case-based reasoning or explanation by example (Caruana et al., 1999, 2015), in which a model returns an explanatory sample for each model prediction.”
Interpreting Neural Networks with Nearest Neighbors (2018, background): “These training examples can provide interpretations as a form of analogy (Caruana et al., 1999), an intuitive explanation for both machine learning experts and non-experts (Klein, 1989; Kim et al.”
Asking 'Why' in AI: Explainability of intelligent systems - perspectives and challenges (2018, methodology): “For example, Caruana et al. (1999) demonstrated howcase-based reasoning could be used to generate explanations for a neural network by using the latter as a means of computing the distance metric for case retrieval.”
A Survey on Neural Network Interpretability (2020, background): “One basic intuition is to find examples that the model considers to be most similar (in terms of latent representations) [111].”
Contextual Explanation Networks (2017, background): “There are multiple other complementary approaches to interpretability ranging from a variety of visualization techniques [35, 36, 37, 38], to explanations by example [39, 40, 41, 42], to natural language rationales [43].”
How Case-Based Reasoning Explains Neural Networks: A Theoretical Analysis of XAI Using Post-Hoc Explanation-by-Example from a Survey of ANN-CBR Twin-Systems (2019, methodology): “For instance, many recent reviews of XAI make little or no reference to key twin-systems papers in the CBR literature [1, 6, 7], while referencing papers outside CBR canon [20, 21].”
Stakeholders in Explainable AI (2018, methodology): “Approaches here include identifying instances from the training set most signiﬁcant to a particular output (Koh and Liang 2017) and employing case-based reasoning techniques to retrieve similar training examples (Caruana et al. 1999).”
Fair, Transparent, and Accountable Algorithmic Decision-making Processes (2017, background): “These interpretations include natural language explanations [42], visualizations of learned representations or models ( e.g. saliency maps in deep neural nets [61]), and explanations by example ( e.g. a tumor is classiﬁed as malignant because to the model it looks like these other tumors) [11].”