2nd Face to Face Consortia meeting in Sweden
Collaboration is the Key to Successful Innovation of Explainable AI for Industries
In the evolving landscape of industrial applications, Artificial Intelligence (AI) promises to revolutionize efficiency and sustainability. However, many AI initiatives stumble at the implementation stage, with only 15% of projects delivering tangible value. A major hurdle is the opacity of AI systems, which obscures their decision-making processes and breeds mistrust among end users. To bridge this gap, the EXPLAIN project, an EU-funded initiative running from 2022 to 2025, aims to redefine the AI lifecycle for Industry 4.0 by prioritizing explainability and transparency at every stage.
The Challenge: Building Trust in AI
Despite high-quality AI models, end users and domain experts in various industries often rely on their own experience, sidelining AI recommendations. This lack of trust stems from the opaque nature of AI systems. Building trust necessitates transparency and comprehensibility in AI decision-making processes. Human-centered AI has emerged as a solution, but developing such AI for industrial applications is challenging due to limited access to end users and proprietary work-related data.
The EXPLAIN Project: A Collaborative Approach
The EXPLAIN project seeks to overcome these challenges by fostering collaboration across the AI supply chain. The consortium comprises a diverse array of companies and academic experts from Germany, the Netherlands, and Sweden. This collective effort enables user studies to understand end users’ needs and validate concepts on customers’ sites.
Highlights from the Consortium Meeting
From April 15 to 17, 2024, the consortium convened at ABB Corporate Research in Västerås, Sweden, for a pivotal meeting. The event featured engaging demos, workshops, and insightful discussions aimed at advancing the project’s goals.
Day One: The first day focused on lessons learned and showcased seven use cases through captivating demos. These use cases spanned various technical advancements, highlighting the project’s diverse applications.
Day Two: Workshops on the second day delved into the overarching goals of major work packages, identified challenges, and charted the path towards an end-to-end machine learning lifecycle for each use case. Discussions also centered on key performance indicators (KPIs) critical for the upcoming ITEA review.
Day Three: Notable speakers on the third day included ABB AI lead Shiva Sander Tavallaey and RISE Data Scientist and Team Lead Stella Riad. They provided insights into AI for industry and explainable AI (XAI) for predictive maintenance, respectively.
Participants from different partners joined this consortium meeting in the office of ABB Corporate Research in Västerås, Sweden
Led by the UX team at ABB Corporate Research Sweden, the workshops were structured to encourage collaborative discussions aimed at achieving comprehensive project objectives. Key insights from these workshops and demos include:
- There are commonalities and shared insights across various use cases. We have eight use cases from different industrial companies, covering a spectrum from anomaly detection to soft sensors. These use cases differ in appearance and data types, with some involving images and others time series. By showcasing technical solutions from each use case and engaging in cross-functional discussions during workshops, we were able to gain inspiration and enhance our solutions by learning from each other.
- Thoroughly planned and arranged face-to-face meetings are effective in fostering common understanding, addressing existing issues, and facilitating smoother future collaborations.
- Conducting hybrid meetings with a large group, especially during workshops, poses practical challenges. We managed to succeed with several computers and booked meeting rooms, but technical equipment needs to be tested more thoroughly in advance.
The workshops were highly successful and received positive feedback from the participants. The event underscored the importance of face-to-face interaction in enhancing communication and collaboration among diverse project partners. The consortium meeting also offered a chance to display the newly renovated facilities at ABB Corporate Research in Västerås, which hosted the multinational gathering.
As the EXPLAIN project progresses, it highlights the collective dedication to leveraging the transformative potential of AI while maintaining a focus on transparency and trustworthiness.