NonFoodKG: A Non-Food to Food Product Knowledge Graph for Consumer Applications in Retail Environments

Tracking #: 3180-4394

This paper is currently under review
Authors: 
Michaela Kümpel
Michael Beetz

Responsible editor: 
Guest Editors Interactive SW 2022

Submission type: 
Tool/System Report
Abstract: 
The Web offers plenty of product information that is valuable for supporting customer decision processes. On the one hand, food product knowledge has been shown to be useful for consumer applications like recipe or dietary recommendation, for example. On the other hand, product knowledge can be extremely useful for applications in retail stores when it is linked to a semantic environment map. That way, it can be used to help customers find products they are searching for or highlight interesting product information like awarded labels or contained ingredients and allergens. Unfortunately, the Semantic Web lacks a semantically enhanced knowledge graph unifying non-food product information while the existing shopping applications lack a standardised connection to environment information. This work proposes NonFoodKG, an open-source non-food to food product knowledge graph integrating modular product information from the Web as well as accurate environment information from sensor data that can be customised for different applications and used devices. We describe the design process and modularity of the knowledge graph as well as example applications of it, including an Augmented Reality shopping assistant to highlight all products containing a given preference like an ingredient/ allergen or label and a dietary recommendation. We show how a routing application to find product destinations can easily be used by different agents like smartphone or robot. Thus, NonFoodKG can be seen as an example of one point of data access in digitisation of retail stores towards omnichannel applications. NonFoodKG links to environment information, which we create using the KnowRob knowledge processing framework. This connection enables agents to link product to action knowledge and thereby generalise the product information to be used in different domains while the modular ontologies enable personalisation of applications. NonFoodKG is publicly available and will be maintained and extended over time in order to facilitate various applications such as in the retail and household domain.
Full PDF Version: 
Tags: 
Under Review