Enriching Artificial Intelligence Explanations with Knowledge Fragments

Artificial intelligence models are increasingly used in manufacturing to inform decision making. Responsible making requires accurate forecasts and an understanding of the models’ behavior. Furthermore, insights into rationale can be enriched with domain knowledge. This research builds explanations considering feature rankings for a particular forecast, enriching them media news entries, datasets’ metadata, entries from Google knowledge graph. We compare two approaches (embeddings-based semantic-based) on real-world use case regarding demand forecasting. The embeddings-based approach measures similarity between relevant concepts retrieved metadata based word movers’ distance embeddings. semantic-based recourses wikification Jaccard instead. leads more diverse when displaying events precise results recommended datasets. conclude that provided further improved information purpose potential actions taken influence provide “what-if” analysis capabilities.