Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SmartEx: A Framework for Generating User-Centric Explanations in Smart Environments

Mersedeh Sadeghi, Lars Herbold, Max Unterbusch, Andreas Vogelsang

TL;DR

SmartEx tackles the lack of user-centric explainability in smart environments by providing a formal model for explanations and a reference architecture for a context-aware explanation engine. It combines causal path reasoning over rule-based systems with contextual information to generate personalized explanations at multiple granularity levels via a view-based framework. The authors implement a Java RESTful prototype integrated with Home Assistant and demonstrate feasibility on a TV-muting scenario, using a Context Manager, Inference Function, and NL transformation. The work lays groundwork for broader evaluation, richer context modeling, and extension to other AI systems and interactive explanations.

Abstract

Explainability is crucial for complex systems like pervasive smart environments, as they collect and analyze data from various sensors, follow multiple rules, and control different devices resulting in behavior that is not trivial and, thus, should be explained to the users. The current approaches, however, offer flat, static, and algorithm-focused explanations. User-centric explanations, on the other hand, consider the recipient and context, providing personalized and context-aware explanations. To address this gap, we propose an approach to incorporate user-centric explanations into smart environments. We introduce a conceptual model and a reference architecture for characterizing and generating such explanations. Our work is the first technical solution for generating context-aware and granular explanations in smart environments. Our architecture implementation demonstrates the feasibility of our approach through various scenarios.

SmartEx: A Framework for Generating User-Centric Explanations in Smart Environments

TL;DR

SmartEx tackles the lack of user-centric explainability in smart environments by providing a formal model for explanations and a reference architecture for a context-aware explanation engine. It combines causal path reasoning over rule-based systems with contextual information to generate personalized explanations at multiple granularity levels via a view-based framework. The authors implement a Java RESTful prototype integrated with Home Assistant and demonstrate feasibility on a TV-muting scenario, using a Context Manager, Inference Function, and NL transformation. The work lays groundwork for broader evaluation, richer context modeling, and extension to other AI systems and interactive explanations.

Abstract

Explainability is crucial for complex systems like pervasive smart environments, as they collect and analyze data from various sensors, follow multiple rules, and control different devices resulting in behavior that is not trivial and, thus, should be explained to the users. The current approaches, however, offer flat, static, and algorithm-focused explanations. User-centric explanations, on the other hand, consider the recipient and context, providing personalized and context-aware explanations. To address this gap, we propose an approach to incorporate user-centric explanations into smart environments. We introduce a conceptual model and a reference architecture for characterizing and generating such explanations. Our work is the first technical solution for generating context-aware and granular explanations in smart environments. Our architecture implementation demonstrates the feasibility of our approach through various scenarios.
Paper Structure (11 sections, 1 equation, 3 figures, 3 tables, 2 algorithms)

This paper contains 11 sections, 1 equation, 3 figures, 3 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. User-Centric Explanations Conceptualization
  4. Motivating Scenario
  5. Explanation Conception
  6. A User-centric Explanation Engine
  7. Explanations for Rule-based Systems
  8. Algorithmic Explanation
  9. User-centric Explanation
  10. Implementation and Feasibility Check
  11. Conclusion

Figures (3)

  • Figure 1: Reference architecture for the Smart Explanation Engine (SmartEx)
  • Figure 2: Context model (left) and its usage at runtime. The Context Manager provides the Inference Function with the current values of relevant contexts. In this example, the Inference Function infers that the most appropriate view to present to the user is the Fact Explanation View.
  • Figure 3: Explanation generation workflow

Theorems & Definitions (4)

  • Definition 1: Explanation $\boldsymbol\Psi$
  • Definition 2: Views of $\Psi$
  • Definition 3: Inference Function
  • Definition 4: Transformation Function