Agents that Learn to Explain Themselves

By M. Rey. 1994

Subfields: 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 Safety, Robustness & Interpretability

In Human-Computer Interaction

Citations in the corpus

Generating Visual Explanations (2016, background): “Explanation systems span a variety of applications including explaining medical diagnosis [13], simulator actions [14,15,16,19], and robot movements [17].”
Explanation in Human-AI Systems: A Literature Meta-Review, Synopsis of Key Ideas and Publications, and Bibliography for Explainable AI (2019, methodology): “…2017; Frank, McGuire, Moses, and Stephenson, 2016; Hayes and Shaw, 2017; Haynes, Cohen, and Ritter, 2008; Jain, Zamir, Savarese, and Saxena, 2016; Johnson, 1994; Kim, Chacha, and Shah, 2013; Lane, Core, Van Lent, Solomon, and Gomboc, 2005; Langley, Meadows, Sridharan, and Choi, 2017; Lomas et…”
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).”
Attentive Explanations: Justifying Decisions and Pointing to the Evidence (2016, background): “planations 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 system that determines the motivation for a decision by recalling the situation in w”
Explainable Agency for Intelligent Autonomous Systems (2017, background): “The most relevant work comes from Johnson (1994) and van Lent et al. (2004), who developed agents for tactical air combat and small unit settings, respectively, that recorded decisions made during missions, provided reasons on request, and dealt with coun-terfactual queries.”
Designs for explaining intelligent agents (2009, methodology): “Examples of mechanistic explanations in intelligent systems include the Debrief system (Johnson, 1994), which uses introspection to retrace and replay the results of a problem-solving trace, and Mycin, Vista (Taylor et al., 2002), and other rule-trace systems that provide a rule firing log.”
Do You Get It? User-Evaluated Explainable BDI Agents (2010, background): “Debrief is the first system that explains agent behavior [10].”
Explainable Artificial Intelligence for Training and Tutoring (2005, background): “Three notable efforts falling into this category are the Program Enhancement Advisor (PEA) for teaching LISP programmers to improve their code [5], the family of successors to MYCIN [1], and another entity-driven explanation system, Debrief [4].”
Self-Explaining Agents in Virtual Training (2008, background): “However, Debrief derives what must have been the agent’s underlying beliefs for an action, but sometimes an action has several possible explanations.”
An Explainable Artificial Intelligence System for Small-unit Tactical Behavior (2004, methodology): “By performing this analysis Debrief can report the critical attribute values that resulted in the agent's decision.”
Interpretable R-CNN (2017, background): “These early rule-based systems can generate theorem proving types of interpretable justifications [64, 68, 32, 41, 71], but were much less effective, too expensive to build and too brittle against the complexities of the real world.”
"Why Did You Do That?" Explainable Intelligent Robots (2017, background): “Intelligent agents that can explain their actions and justify their decisions is fast becoming a popular topic for research (Defense Advanced Research Projects Agency (DARPA) 2016; Ribeiro, Singh, and Guestrin 2016; Johnson 1994).”
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).”
Interacting meaningfully with machine learning systems: Three experiments (2009, background): “The bulk of the literature on explaining intelligent systems has focused on explaining expert systems based on deductive logic or rules (Swartout 1983; Clancey 1983; Wick and Thompson 1992; Johnson 1994).”
Design and Evaluation of Explainable BDI Agents (2010, background): “To generate an answer, Debrief modifies the recalled situation repeatedly and systematically, and observes the effects on the agent's decisions.”