Simplicity and probability in causal explanation.

By Tania Lombrozo. 2007

396 citations recorded by the source corpus.

Subfields: Artificial Intelligence, History and Philosophy of Science, Cognitive Science & Psychology

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 History and Philosophy of Science

In Cognitive Science & Psychology

Citations in the corpus

Reframing Human-AI Collaboration for Generating Free-Text Explanations (2021, neutral): “…truthfulness be ascertained or is more information needed?) and factuality, which aligns with “compatibility with receiver’s existing be-liefs” and that the best explanation is the “most explanation from the receiver/user’s perspective (Lombrozo, 2007; Zemla et al., 2017; Sulik et al., 2021).”
Explanation in Human-AI Systems: A Literature Meta-Review, Synopsis of Key Ideas and Publications, and Bibliography for Explainable AI (2019, neutral): “Hilton (1990, 2005); Hilton, McClure, and Sutton (2010); Klein, Rasmussen, Lin, Hoffman, and Case(2014); Krull and Anderson (1997); Lombrozo (2007, 2010, 2010); McGill (1991); Rasmussen, Nixon, and Warner (1990); Samurçay and Hoc (1996); Selfridge, Daniell, and Simmons, D. (1985); Woodward, (2005);…”
Explainable AI: Beware of Inmates Running the Asylum Or: How I Learnt to Stop Worrying and Love the Social and Behavioural Sciences (2017, positive): “Recent work shows that people prefer explanations that are simpler (cite few causes) [Lombrozo, 2007], more general (they explain more events) [Lombrozo, 2007], and coherent (consistent with prior knowledge) [Thagard, 1989].”
Explanation in Artificial Intelligence: Insights from the Social Sciences (2017, neutral): “Lombrozo [105] notes that the experiments on generality and simplicity performed by Read and Marcus-Newhall [152] cannot rule out that participants selected simple”
Conservative Q-Improvement: Reinforcement Learning for an Interpretable Decision-Tree Policy (2019, neutral): “It has been shown that, all else being equal, humans prefer simpler explanations compared to more complex ones [Lombrozo, 2007].”
The challenge of crafting intelligible intelligence (2018, neutral): “According to Lombrozo, humans prefer explanations that are simpler ( i.e. , contain fewer clauses), more general, and coherent ( i.e. , consistent with what the human’s prior beliefs) [26].”
How do Humans Understand Explanations from Machine Learning Systems? An Evaluation of the Human-Interpretability of Explanation (2018, positive): “For example, humans prefer explanations that are both simple and highly probable [Lombrozo, 2007].”
Evaluating everyday explanations (2017, neutral): “These results contrast with previous findings that have shown people prefer explanations that appeal to the fewest number of causes (Lombrozo, 2007; Read & Marcus-Newhall, 1993).”
Combinatorial Methods for Explainable AI (2020, neutral): “It is perhaps not surprising that users prefer simple explanations, with fewer causes or factors, but it was also found that simpler explanations were preferred over more likely explanations [18].”
Evaluating computational models of explanation using human judgments (2013, neutral): “Research on people’s preferences for simplicity bear on this case, which shows that people may choose an explanation with fewer causes even if it is less likely than other more complex alternatives [Lombrozo, 2007].”
What Do We Want From Explainable Artificial Intelligence (XAI)? - A Stakeholder Perspective on XAI and a Conceptual Model Guiding Interdisciplinary XAI Research (2021, neutral): “, information mentioning fewer causes; [185]).”
Explanations in the wild. (2021, neutral): “…explanations, for instance leading to proposals for how Bayesian models of cognition might implement various explanatory virtues (Wojtowicz & DeDeo, 2020), or that test whether people prefer explanations that facilitate prediction (Lombrozo & Carey, 2006) or that are parsimonious (Lombrozo, 2007).”
A Study of Automatic Metrics for the Evaluation of Natural Language Explanations (2021, neutral): “Similarly, Lombrozo (2007) notes that simplicity and generality might be the key to evaluating explanations.”
Understanding Why: The cognitive science of explanation (2008, neutral): “Yet many of the most illuminating explanations are praised for their simplicity and elegance, for the details they leave out (Strevens, 2004; Lombrozo, 2007).”
Explanation and Abductive Inference (2012, positive): “Factors that boost an explanation's perceived quality include simplicity (Bonawitz & Lombrozo, in press; Lagnado, 1994; Lombrozo, 2007; Read & Marcus-Newhall, 1993; Thagard, 1989), breadth (Preston & Epley, 2005; Read & Marcus-Newhall, 1993; Thagard, 1989), consistency with prior knowledge (Thagard,”
Explainability fact sheets: a framework for systematic assessment of explainable approaches (2019, positive): “Therefore, any explainable system should be consistent with the explainee’s prior knowledge [9, 33, 35], which can only be achieved when the explainee’s mental model is part of the explainability system (U10) as otherwise there is nothing to be coherent with.”
Explanation and categorization: how "why?" informs "what?". (2009, neutral): “For example, simple mechanistic explanations are typically preferred over complex alternatives (Lombrozo, 2007), and should thus generate correspondingly larger effects of causal status.”
Explaining by Removing: A Unified Framework for Model Explanation (2020, neutral): “Richer information may not always be desirable, because psychology research shows that people report higher satisfaction with explanations that cite fewer causes (Thagard, 1989; Read and Marcus-Newhall, 1993; Lombrozo, 2007; Miller, 2019).”
Inference to the Best Explanation (IBE) Versus Explaining for the Best Inference (EBI) (2015, neutral): “, simplicity, scope), and that these criteria in turn constrain learning and inference, in part by leading people to discover and generalize the ‘‘best’’ explanation (for evidence, see Bonawitz and Lombrozo 2012; Legare and Lombrozo 2014; Lombrozo 2007; Walker et al. 2014; Williams and Lombrozo 2010”