Engineering User-centered Explanations to Query Answers in Ontology-driven Socio-technical Systems

Tracking #: 3193-4407

Authors: 
Carlos Teze
Jose N. Paredes
Maria Vanina Martinez
Gerardo Simari

Responsible editor: 
Guest Editors Ontologies in XAI

Submission type: 
Full Paper
Abstract: 
The role of explanations in intelligent systems has in the last few years entered the spotlight as AI-based solutions appear in an ever-growing set of applications. Though data-driven (or machine learning) techniques are often used as examples of how opaque (also called black box) approaches can lead to problems such as bias and general lack of explainability and interpretability, in reality these features are difficult to tame in general, even for approaches that are based on tools typically considered to be more amenable, like knowledge-based formalisms. In this paper, we continue a line of research and development towards building tools that facilitate the implementation of explainable and interpretable hybrid intelligent socio-technical systems, focusing on features that users can leverage to build explanations to their queries. In particular, we present the implementation of a recently-proposed application framework (and make available its source code) for developing such systems, and explore user-centered mechanisms for building explanations based both on the kinds of explanations required (such as counterfactual, contextual, etc.) and the inputs used for building them (coming from various sources, such as the knowledge base and lower-level data-driven modules). In order to validate our approach, we develop two use cases, one as a running example for detecting hate speech in social platforms and the other as an extension that also contemplates cyberbullying scenarios.
Full PDF Version: 
Tags: 
Reviewed

Decision/Status: 
Minor Revision

Solicited Reviews:
Click to Expand/Collapse
Review #1
Anonymous submitted on 06/Sep/2022
Suggestion:
Major Revision
Review Comment:

The authors have put in effort into addressing both the reviewers comments and have incorporated suggested edits into the paper. The paper reads better now in terms of both content and organization.

Some comments:

- In contributions the authors mention the following, We develop an instantiation of the HEIC Application Framework” → It is not clear to me what the instantiation means, does it mean that the authors developed a new HEIC module for the HEIST framework that incorporates explanation styles. Please make this contribution more clear.

- The paper lacks a discussion section and it is hard to understand the challenges and shortcomings of the work without this. I see some of this in the conclusions, but I request the authors to add a more detailed discussion based on the analysis of what key problems their HEIC framework addresses and what are some challenges.

Finally, I thank the authors for their revisions, but coming from an OWL background I cannot judge how interoperable and reusable a Datalog implementation is. I did read that the authors said the implementation can easily be converted to other languages, I was wondering if this is true for the widely-used ontology language, OWL. Also, I understand that the work borrows from Chari et al’s Explanation Ontology, who also have an updated version which now supports more explanation types (https://www.semantic-web-journal.net/content/explanation-ontology-genera...). But I am not sure if the application of their work to the cybersecurity domain is novel enough, I would recommend that the authors reach out to Chari et al., and have a conversation on whether their implementation might help a sole ontology and if they can benefit from some OWL integration too.

Review #2
Anonymous submitted on 19/Sep/2022
Suggestion:
Accept
Review Comment:

I thank the authors for the effort provided in revising the manuscript.
I do not have any further comments to add by considering the answers to the other reviews as well.
I recommend to accept the manuscript as it is.