About Us

How

The IKEA Knowledge Graph (IKG) is built using semantic web and graph technologies. It houses a controlled vocabulary of all things IKEA, called concepts, lateral connections between them, and enables for powerful reasoning logic. Additionally, it stores the links and rules between the concepts and IKEA's content such as products, images, learning objects, etc.

This allows us to understand the context of content at IKEA and make it actionable to computers. For example, we can provide all the concepts that are broken down from an entire room solution and know which different combinations these concepts can take. This helps us to make home furnishing solutions explicit.

Subject-matter experts, also called domain experts, use various interfaces, such as a spreadsheet, UI editors, or other means to communicate exactly what kind of vocabulary and definitions should be included in the IKG and how they should connect other knowledge domains and finally to content. The domain experts do this work as part of one of the working groups responsible for the IKG's four knowledge areas each:

• Home Furnishing
• Product Range
• Life At Home
• Sustainability

The subject-matter experts are not the only source for knowledge, semi-automated text mining as well as computer vision is used to extract knowledge from existing digital resources, such as product descriptions, product images, videos, and PDFs.

In order to serve all of this knowledge in a performant way it is made available via various APIs on top of the knowledge graph made-to-measure for applications so that they do not need to adjust. This way applications can be rest assured to receive the right kind of information at the right time in the format that they need it. The API product runtime is performant and request for changes are adopted within one month.