Too much, too little, or just right? Ways explanations impact end users' mental models

By Todd Kulesza, Simone Stumpf, Margaret M. Burnett, Sherry Yang, Irwin Kwan, Weng-Keen Wong. 2013

427 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Software, Safety, Robustness & Interpretability, Human-Computer Interaction

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 Software

In Safety, Robustness & Interpretability

In Human-Computer Interaction

Citations in the corpus

Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI) (2018, background): “[161] where they presented their findings regarding how explanations impact end users’ mental models.”
Explain Yourself: A Natural Language Interface for Scrutable Autonomous Robots (2018, background): “This trade-off will vary between personnel with different information needs and expertise and domains as discussed in [9].”
Explanation in Artificial Intelligence: Insights from the Social Sciences (2017, background): “n as well as human explanation. Further, it can be benecial in a more pragmatic way: it can help people construct mental models of systems to improve their understanding. For example, Kulesza et al. [63] describe the tricky problem of producing too much or too little information for explanatory debu”
Deep Weighted Averaging Classifiers (2018, background): “As emphasized by papers that have made use of user studies [16, 22, 36], careful empirical work is required to evaluate the effectiveness of explanations, and we leave such an evaluation of this approach for future work.”
Manipulating and Measuring Model Interpretability (2018, background): “[59] studied several ways in which intelligent agents might explain themselves to stakeholders.”
How do Humans Understand Explanations from Machine Learning Systems? An Evaluation of the Human-Interpretability of Explanation (2018, positive): “Closer to the objectives of the proposed work, Kulesza et al. [2013] performed a qualitative study in which they varied the soundness (nothing but the truth) and the completeness (the whole truth) of an explanation in a recommendation system setting.”
The Role of Explanations on Trust and Reliance in Clinical Decision Support Systems (2015, neutral): “…and Engineering City University London London, United Kingdom Adrian.Bussone.1@city.ac.uk, Simone.Stumpf.1@city.ac.uk, Dympna.O'Sullivan.1@city.ac.uk Abstract— Clinical decision support systems (CDSS) are increasingly used by healthcare professionals for evidence-based diagnosis and treatment…”
Principles of Explanatory Debugging to Personalize Interactive Machine Learning (2015, background): “Principle 1.3: Be Complete Kulesza et al. describe completeness as “the extent to which all of the underlying system is described by the explanation” [25], so a complete explanation does not omit important information about the model.”
Explaining models: an empirical study of how explanations impact fairness judgment (2019, background): “by considering two dimensions of explanation fidelity: soundness (how truthful each element in an explanation is with respect to the underlying system) and completeness (the extent to which an explanation describes all of the underlying system) [20].”
Designing Theory-Driven User-Centric Explainable AI (2019, background): “To support causal attribution, tornado diagrams of vertical bar charts are popularly used for lists of attributions [85], saliency heatmaps for attributing to pixels or super-pixels for image-based models [49, 85], or highlighting on paragraph text [15, 57].”
Explainable Artificial Intelligence: a Systematic Review (2020, methodology): “The explanandum is something unexpected by the user that creates dissonance with his/her beliefs [58] Correctability The capacity of a method for explainability to allow end-users make technical adjustments to an underlying model [60, 61] Simplification The capacity to reduce the number of variables”
To Trust or to Think (2021, background): “Hence, in addition to tuning explanation attributes such as soundness and completeness [39], or faithfulness, sensitivity, and complexity [5], explainable AI researchers should ensure that people will exert effort to attend to those explanations.”