XAI and Human-in-the-Loop (X-Loop)
Explainability as a lever for the targeted use of AI with human oversight in the metal industry
Project Team
Contact
In this project
- Researchpaper
Despite the enormous potential offered by (partial) automation, artificial intelligence (AI) has so far been scarcely utilised by SMEs in the metal industry. This is the case despite numerous challenges, such as a shortage of skilled workers and supply chain difficulties. One reason for this is that AI is mostly perceived as a black box. In other words, users cannot trace or influence the AI output, and AI predictions cannot be directly validated. This drastically reduces the acceptance of AI. There is therefore still a significant need for transformation towards digitalisation and the use of AI. In the metal industry in particular, automation solutions are urgently needed not only in manufacturing (e.g. robots), but also in so-called indirect areas such as work planning.
The X-Loop research project aims to develop an AI system for forecasting quotations (in particular automated price calculation) for sheet metal and machined components, which is based on component and order information, existing company data and individually set parameters (e.g. material, surface finish and fit dimensions) and employs a transparent and user-influenced approach for users. The project addresses labour-intensive administrative tasks in the metal industry and will use the example of price calculation to investigate how new approaches from the research field of âExplainable AIâ (Explainable Artificial Intelligence, or XAI for short) can be developed as a lever for the targeted Human-in-the-Loop (HITL) deployment of AI systems for (partial) automation in the metal industry, and what added value these bring.
Consequently, the project can be understood as an assistance system that, during price calculation as part of the quotation process, offers evaluation and decision-support functions which simplify, optimise and accelerate the current manual work of users (e.g. work planners, project managers and sales staff). The project results are intended to significantly reduce the workload involved in price calculation within manufacturing companies whilst simultaneously improving the quality of (partially) automated price calculation. In the long term, the aim is also to provide guidance on how the responsible use of AI (keyword: human-machine cooperation) can be successfully implemented in SMEs.
Cooperation partners: goCAD GmbH, Inlinity Technologies GmbH
Application partners: Köbler Zerspan- und Rundschleiftechnik GmbH Co. KG, LST Laser & Schalttechnik GmbH & Co. KG
Project period: July 2024 â June 2027
Transfer
The project is funded by the Federal Ministry of Research, Technology, and Space for the period July 1, 2024, to June 30, 2027, and is coordinated by goCAD GmbH.
The goal of the funding initiative âResearch, Development, and Use of Artificial Intelligence Methods in SMEsâ by the Federal Ministry of Research, Technology, and Space (BMFTR) is to promote the broader use of innovative ML and other AI methods in SMEs so that they can implement new data-driven applications. While AI is already being used (successfully) in many areas within large corporations, active application in SMEs has so far been lacking due to error-prone systems and a lack of traceability.
The X-Loop project serves as a catalyst to increase the use and success of AI in SMEs in the metal industry, as SMEs in particular must establish a âdigital foundationâ to maintain their competitiveness and remain competitive in the market. Furthermore, the project serves as a model for the responsible use of AI and can open up new data-driven applications in SMEs across various industries.