Main Page

This wiki contains a glossary of terms related to the field of explainable AI within the scope of the TRR projects. The glossary consists of three categories of explanation moves: “Explainer moves”, “Explainee moves” and “General moves”. Additionally, the wiki includes different types of cognitive biases, different reasoning strategies, and other related terms within the scope of TRR 318.

Looking at the glossary from a different perspective, the explanation moves in the glossary include: adult–child interaction moves (such as embedded correction, and repetitions to place a common ground), short answer questions (such as verification and disjunctive questions), long answer questions (such as procedural, expectational, and assertion questions), backchannels (such as phrasal, substantive, and non-lexical ones), and other dialog moves (such as Assert, Reassert, Accept, Accept-part, Reject, Signaling emotions, and Open option).

Moreover, the wiki provides guidelines for annotation of data with explanation moves. In fact, the glossary has been enriched with the extracted examples for each explanation move either from a real-world dataset or from literature. This information is therefore considered as a guideline for the human agents, who will annotate Explainer-Explainee dialogues.

General moves

General moves are the explanation moves, which could happen both from the side of the explainee and the explainer in an explanation dialogue.

• Collected general moves can be seen in one place from here: General moves.

• And the alphabetical list of these moves and corresponding pages can be accessed from this link: General moves category.

Explainee moves

Explainee moves is the category of the explanation moves that happen from the side of explainees in an explanation dialogue.

• Explainee moves can be seen as a whole from here: Explainee moves.

• And the corresponding alphabetical list of moves and pages can be found here: Explainee moves category.

Explainer moves

Explainer moves is the category of the explanation moves that happen from the side of explainers in an explanation dialogue.

• The Explainer moves are collected in one place here: Explainer moves.

• Moreover, the corresponding alphabetical list of moves and pages can be accessed from this link: Explainer moves category.

Biases

The field of explainable AI looks into cognitive factors that bias or compromise decision-making. The wiki contains a list of cognitive biases related to the TRR 318 projects.

• For viewing all the biases and the related definitions in one place refer to this link: Biases.

• An alphabetical list of biases and corresponding pages can be viewed from here: Biases Category

Reasoning

Researchers from the field of explainable AI are also interested in theories that describe patterns of how people reason. The wiki therefore include a list of related concepts to reasoning.

• The related terms to reasoning can be viewed in one place from here: Reasoning.

• In addition, an alphabetical list of corresponding reasoning concepts and pages can be seen in this page: Reasoning categories

Other terms

Apart from the above-mentioned categories, the wiki has classified other related terms to the TRR projects in this category.

• For viewing the concepts of this category in one place, refer to this link: Other terms

• An alphabetical list of these concepts and related pages can also be viewed from here: Other terms category

A Collaborative Annotation Task

As a starting point for making the audience more and more acquainted with the annotation scheme introduced in this wiki, we propose a collaborative annotation task as a part of TRR Activity Afternoon which is held on September 28th, 2022. Please visit this link: A Collaborative Annotation Task

Classification of moves based on domain-related observations

We classify the moves based on the domains in which they have been observed. Additionally, some of the moves are assumed to be cross domain as they might appear in different domains.

• An alphabetical list of the moves, which are observed in the domain of dialog moves can be accessed from this link: Dialog Moves Domain.

• An alphabetical list of the moves, which are observed in the domain of tutoring can be accessed from this link: Observed in Tutoring Domain.

• An alphabetical list of cross domain moves can be accessed from this link: cross domain moves.