A review of explanation methods for Bayesian networks

By Carmen Lacave, Francisco Javier Díez. 2002

285 citations recorded by the source corpus.

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

Citations in the corpus

Explanation in Human-AI Systems: A Literature Meta-Review, Synopsis of Key Ideas and Publications, and Bibliography for Explainable AI (2019, background): “…Konstan, and Riedl, 2000; Khan, Poupart, and Black, 2009; Kisa, Van den Broeck, Choi, and Darwiche, 2014; Kofod-Petersen, Cassens, and Aamodt, 2008; Lacave and Díez, 2002; Letham, Rudin, McCormick, Madigan, and others, 2015; Liang and Van den Broeck, 2017; Nugent, Doyle, and Cunningham, 2009;…”
What do we need to build explainable AI systems for the medical domain? (2017, background): “There is some related work available on how to make such systems explainable (Shortliﬀe and Buchanan, 1975; Swartout et al., 1991; Johnson, 1994; Lacave and Diez, 2002).”
'It's Reducing a Human Being to a Percentage': Perceptions of Justice in Algorithmic Decisions (2018, background): “Early knowledge representation and reasoning systems typically aimed to meet this requirement by producing explanations for their outputs, for instance by presenting derivations from rules in rule-based architecture [69, 8], or graphical representations of nodes and edges in the case of Bayesian net”
Attentive Explanations: Justifying Decisions and Pointing to the Evidence (2016, background): “[39] proposes explanation systems for medical applications, [24, 40, 10] propose explanations as a feedback to improve simulated training for special teaching programs, [28] proposes a model to explain robot actions, [23] evaluates methods developed to explain Bayesian networks and [21] proposes a s”
Explanation and Justification in Machine Learning : A Survey Or (2017, methodology): “[Lacave and D´ıez, 2002] present a survey of methods of explanation for Bayesian networks and an excellent analysis of the methods in terms of several properties of explanation.”
A generalized taxonomy of explanations styles for traditional and social recommender systems (2012, background): “Due to this virtue, explanations have been studied in many different research areas, such as Human Computer Interaction (Nielsen and Molich 1990; Pu and Chen 2006), User Modelling (O’Sullivan et al. 2004; Paramythis et al. 2001; Sinha and Swearingen 2002), E-Learning systems (McCarthy et al. 2004) a”
Causability and explainability of artificial intelligence in medicine (2019, methodology): “This was the basis for explanation, and there is some early related work available, for example, (Johnson, 1994; Lacave & Diez, 2002; Swartout, Paris, & Moore, 1991).”
Aspects of Intelligent Systems Explanation (2013, methodology): “Non-mathematically inclined users may have difficulty in understanding explanations based on this technique [43].”
Minimal Sufficient Explanations for Factored Markov Decision Processes (2009, background): “Lacave and Diez (2002) surveyed the existing techniques and their limitation for explanation in Bayesian Networks.”
Interacting meaningfully with machine learning systems: Three experiments (2009, background): “These computationally expensive explanation algorithms for Bayesian networks consist of finding the subset of evidence that most influences the posterior probability and visualizing the flow of evidence through the network during inference (Suermondt, 1992; Chajewska and Draper, 1998; Lacave and Die”
Designing and Evaluating Explanations for Recommender Systems (2011, background): “For further reading, see the following reviews on expert systems with explanatory capabilities for three of the most common inference methods: heuristic-based methods [36], Bayesian networks [35], and case-based reasoning [22].”
Visual Interaction with Deep Learning Models through Collaborative Semantic Inference (2019, background): “However, explanations should also not overwhelm an end-user, and many previous approaches thus choose to select less complex models [52] or aim to reduce the feature space of a trained model [18,87].”
Key challenges for delivering clinical impact with artificial intelligence (2019, background): “…learning methods that build upon a long history of research in traditional symbolic AI techniques to allow for encoding of semantics of data and the use of ontologies to guide the learning process may permit human experts to understand and retrace decision processes more effectively [91, 92].”
Trends and Trajectories for Explainable, Accountable and Intelligible Systems: An HCI Research Agenda (2018, background): “Later research focused on explaining Bayesian Belief Networks [102, 126] and Case-Base Reasoning systems [159], following the trends of research interest in these AI paradigms.”
Explainable AI for Designers: A Human-Centered Perspective on Mixed-Initiative Co-Creation (2018, background): “In a review, Lacave and Dı́ez [10] categorized existing approaches into three main types: explanation of evidence, explanation of the model, and explanation of the reasoning.”
Generating Visual Explanations (2016, background): “In contrast to systems like [13,14,15,16,17,18] which aim to explain the underlying mechanism behind a decision, authors in [1] concentrate on why a prediction is justiﬁable to a user.”
Why Attention is Not Explanation: Surgical Intervention and Causal Reasoning about Neural Models (2020, background): “…linear models (Lakkaraju et al., 2017); that linear models are in turn easier to interpret relative to generalized additive models (Lou et al., 2012) or Bayesian networks (Lacave and D´ıez, 2002); and so on until reaching the almost entirely black-box behavior of deep neural models (Miller, 2018).”
Explanatory Debugging: Supporting End-User Debugging of Machine-Learned Programs (2010, background): “computationally expensive explanations exist for general Bayesian networks [13].”