Revolutionizing Quality: The Next-Gen Multimodal AI Evaluation Framework
Track: Multilingual AI | AI4 | Advanced |
Wednesday, June 4, 2025, 4:30pm – 5:15pm
Held in: High Live 1 (Plenary)
Presenters:
Agustín Da Fieno Delucchi - Microsoft
Katerina Gašová - Argos Multilingual
Moderator: Pavel Soukenik
This panel will unveil and demonstrate a newly developed quality framework for evaluating multimodal and AI-generated content. The discussion will explore the integration of text, images, audio, and interactivity, emphasizing the holistic nature of the novel approach to ensure high standards of relevance, accuracy, user engagement, and cultural sensitivity. Attendees will see examples of the new evaluation framework and will be able to provide feedback via live polls on the quality metrics and participate in a Q&A segment to ask questions and share their perspectives on the challenges and opportunities in evaluating multimodal AI-generated content.
Key Takeaways:
- Examples of evaluating multimodal AI-generated content using the proposed comprehensive quality framework
- How to apply a holistic standard for multimodal communication without a source text
- The importance of ethical considerations and responsible AI practices in evaluating AI-generated content
