SEMIC 2024 Pre-Conference Workshops: AI in Support of Interoperability

SEMIC 2024 Pre-Conference Workshops: AI in Support of Interoperability

Introduction

The SEMIC 2024 pre-conference workshops delved deep into the symbiotic relationship between AI and interoperability. The dual focus was on how AI can enhance interoperability and how interoperability can further aid AI applications.

Morning Session: AI for Interoperability

The morning session was dedicated to exploring how AI applications can drive interoperability initiatives. Rasan Raru, Deputy Head of Unit at the Publication Office Portal in Luxembourg, and Florian Hartl from the SEMIC team, inaugurated the event. They highlighted the significance of the event in fostering knowledge sharing, especially around interoperability and AI.

Key Presentations

1. Finnish Colleagues: Practical Implementation of Interoperability Using AI
   • Speakers: Auntie Helen and Annamarie Wenber
   • Focus: Challenges and practical solutions in using AI in Finnish public administration, covering aspects like legal considerations and data quality.
2. Alberto Abella: Agile Standardization and AI in Open Data
   • Focus: Challenges in data standardization and the role of AI in efficient, agile standardization processes.
3. PWC Belgium: Tools for Data Modelers
   • Speaker: Emilian Cod
   • Focus: AI tools to aid data modelers, including the use of chatbots for data model creation and interrogation.
4. William Vander: EU Reference Data and AI Enhancements
   • Focus: Integration of AI for improving the quality and usability of EU reference data, including automated translation, tagging, and vocabulary expansion.

Afternoon Session: Interoperability Supporting AI

The afternoon session addressed how establishing strong interoperability frameworks can enhance AI applications.

Key Presentations

1. Jeremy Ranem: Enhancing AI Trust with Knowledge Graphs
   • Focus: Integrating Knowledge Graphs and llms to build trust in AI systems, offering richer context, transparency, and dynamic learning.
2. PWC Belgium: Discoverability and Reusability of AI Models with MLdcat
   • Speakers: Imedio Stanley and Arthur Schulz
   • Focus: Addressing the difficulties in discovering and reusing AI models by proposing a standardized machine learning data catalog approach.
3. Rasan Petru: Interoperability of Public Administration Chatbots
   • Focus: Framework for interconnecting public administration chatbots, ensuring smooth, effective, and secure user experiences across different administrative domains.

Panel Discussions

The interactive panel session allowed speakers and participants to discuss various challenges, from the speed of AI standardization to practical uses and risk management.

Key Takeaways

1. Trust and Transparency: AI models need robust frameworks like Knowledge Graphs to be trustworthy.
2. Speed of Adoption: Agile but structured approaches to standardization and interoperability are essential for keeping up with rapid technological advancements.
3. User Experience: Successful AI applications in public administration must prioritize user experience, particularly in conversational agents.

Keywords

• AI for Interoperability
• Machine Learning Data Catalog
• Knowledge Graph
• Semantic Interoperability
• Public Administration Chatbots
• Data Standardization
• User Experience
• Trust in AI

FAQ

Q: What were the main focus areas of the SEMIC 2024 pre-conference workshops?

A: The main focus areas were how AI can bolster interoperability and how interoperability frameworks can enhance AI applications.

Q: Who were some of the key speakers at the event?

A: Notable speakers included Rasan Raru, Auntie Helen, Annamarie Wenber, Alberto Abella, Emilian Cod, William Vander, and Jeremy Ranem.

Q: What are some practical examples of AI supporting interoperability discussed at the event?

A: Examples include Finnish public administration's use of AI for summarizing consultation materials and agile standardization in open data by Alberto Abella.

Q: How is interoperability essential to AI applications?

A: Interoperability ensures seamless data exchange and system integration, essential for building robust, trustworthy AI applications.

Q: What are some of the challenges in AI standardization discussed?

A: Challenges include the need for agile processes, data quality, and the necessity to involve domain experts for practical and effective standardization.

Q: What solutions were proposed for integrating AI in public administration chatbots?

A: Solutions include frameworks for secure data exchange, maintaining user experience, and handling data privacy across interconnected chatbots.