Home // International Journal On Advances in Software, volume 16, numbers 1 and 2, 2023 // View article

A Graph Matching Algorithm to Extend Software Wise Systems with Human Semantic

Authors:
Abdelhafid Dahhani
Ilham Alloui
Sébastien Monnet
Flavien Vernier

Keywords: statecharts; monitoring systems; adaptive system and control; knowledge-based systems; discrete-event systems; graph matching; semantic

Abstract:
Wise systems refer to distributed communicating software objects, which we termed Wise Objects, able to autonomously learn how they are expected to behave and how they are used. Wise Objects are designed to be associated either with software or physical objects (e.g., home automation) to adapt to end users while demanding little attention from them. This last requirement obeys to the principle of calm technology introduced by Mark Weiser and John Seely Brown in 1995. Wise Objects are endowed with autonomous computing capabilities as they implement the notion of IBM’s 4 state loop Monitor-Analyze-Plan-Execute over a shared Knowledge. However, they suffer from a lack of semantic, which prevents them from communicating effectively with a human. The work presented in this paper aims at extending Wise Objects with the ability to use human semantic to communicate with a user. Construction of such systems requires at least two views: (i) a conceptual view relying on knowledge given by developers to either control or specify the expected system behavior; and (ii) an auto-generated view acquired by wise systems during their learning process. The problem is that, while a conceptual view is understandable by humans (i.e., developers, experts, etc.), a view generated by a software system contains mainly numerical information with mostly no meaning for humans. In this paper, we address the issue of how to relate both views using two state-based formalisms: Input Output Symbolic Transition Systems for conceptual views and State Transition Graphs for views generated by the wise systems. Our proposal is to extend the generated knowledge with the conceptual knowledge using a matching algorithm founded on graph morphism. Target results are twofold: (i) make wise systems’ generated knowledge understandable by humans, (ii) enable human evaluation of wise systems’ outputs. To illustrate the overall process, the construction of two samples of graph matching on a roller shutter and a light bulb are considered.

Pages: 59 to 70

Copyright: Copyright (c) to authors, 2023. Used with permission.

Publication date: June 30, 2023

Published in: journal

ISSN: 1942-2628