Trustworthy Generative AI for Advanced Industrial DigitaliZation (Trust_Gen_Z)
The primary focus area is gAI-based ‘forecasting’. In complex operational scenarios, it will monitor, detect, and forecast industrial equipment and machine conditions considering different data modalities, e.g. image, text, and tabular data demonstrated in the use cases.
Project website
Start
2024-09-10
Planned completion
2027-09-09
Main financing
Collaboration partners
Research area
Research group
Project manager at MDU
Description
The project uses generative AI (gAI) to advance prescriptive analytics in industrial digitization. By developing a multimodal framework, we aim to improve decision-making processes through prescriptive analytics with actionable insights and explanations for their intended outcomes. The project will facilitate inspection, monitoring, optimization and maintenance of industrial equipment and machinery. However, implementing a multimodal framework in the industry faces a challenge: the absence of reliable AI methods that generate predictions while offering prescriptive decisions. While the use of gAI looks promising for this task, there are significant gaps in current explainable AI (XAI) methods, which limit their applicability to prescriptive analytics.
Project goal
The primary objective is to develop a prototypical framework for advancing “prescriptive analytics” in industrial digitalization through gAI.
Project activities
Trust_Gen_Z will address these challenges and support prescriptive analysis to advance digitization in the automotive and telecom industries. The project activities are divided into five work packages:
- Project Management and Dissemination of Results
- Inference to Explanation
- AI for Intelligence
- Analysis for Sustainable Digital Transformation
- Test, Validation and Demonstration
