Enriching Artificial Intelligence Explanations with Knowledge Fragments

TL;DR

This paper enhances AI explanations in manufacturing by integrating domain knowledge from news, metadata, and knowledge graphs, comparing embedding and semantic methods in demand forecasting.

Contribution

It introduces a novel approach to enrich AI explanations with external knowledge sources, improving interpretability in manufacturing contexts.

Findings

Embedding-based and semantic-based methods are compared in a real-world demand forecasting case.

Enriched explanations provide better insights into model behavior.

Knowledge integration improves interpretability of AI models.

Abstract

Artificial Intelligence models are increasingly used in manufacturing to inform decision-making. Responsible decision-making requires accurate forecasts and an understanding of the models' behavior. Furthermore, the insights into models' rationale can be enriched with domain knowledge. This research builds explanations considering feature rankings for a particular forecast, enriching them with media news entries, datasets' metadata, and entries from the Google Knowledge Graph. We compare two approaches (embeddings-based and semantic-based) on a real-world use case regarding demand forecasting.