KI-Staging

KI-Staging: Overview

KI-Staging: Overview

(Reference: DFKI GmbH, Illustration: Annika Jenssen)

Mittelgeber

Mittelgeber

(Reference: EU/Land Bremen)

In the KI-Staging project, the DFKI research department Cyber-Physical Systems and the ePhilos AG develop intelligent methods for the analysis of article descriptions in oder to uniformly structure articles and to optimize data inventories of arcticles. For this purpose, methods of natural language processing are utilized which are developed in the field of Artificial Intelligence (AI).

Duration: Jan 1, 2018 - Mar 31, 2020
Partners: ePhilos AG
Research area: Software Systems  
Grant number: FUE0595B
Funding:

Companies are increasingly processing their purchasing processes via electronic catalogs or connected online shops. Therefore, they use so-called e-procurement systems, which optimize operative ordering processes, e.g. by exploiting purchasing advantages, automation and process time reduction. In order to allow for an efficient article search in these shop systems, it is necessary to provide a description of the essential, partly product-specific features and characteristics for each individual article, e.g. size, format, color, etc. However, corresponding descriptions of suppliers often only exist in the form of continuous text, whereby this required information is often missing. Adding them manually is not economically feasible for catalogs with several thousands of regularly changing articles. In the context of KI-Staging, the project partners therefore develop methods for the analysis of article descriptions that allow – at least partially – to automatically recognize and extract the desired features of unstructured continuous text. In addition, methods of natural language processing are utilized. The essential article descriptions, including the features with expressions that are partially only described in the text, flow into an extensive knowledge database as the result of the analysis.

© DFKI GmbH
last updated 22/10/2018
to top