Large Multimodal Models: From Browsers to Embodied AI
Data Science Conference via YouTube
Overview
Explore groundbreaking advancements in AI models through this conference talk from DSC Adria 2024, where Valentina Zadrija examines the evolution from text-based models like ChatGPT, Mistral, and Claude to their applications in real-world autonomous systems. Learn about the transformative potential of large multimodal models and their expansion beyond browser-based interactions into embodied AI applications such as robotics and autonomous vehicles. Discover the critical aspects of safety alignment and validation in AI development, understanding the importance of incorporating safety by design principles to ensure these powerful systems remain aligned with human values. Gain valuable insights into the systematic approaches needed to address the challenges of implementing AI in real-world scenarios while maintaining a strong focus on human safety and ethical considerations.
Syllabus
Large multimodal models: scaling from browsers to embodied AI | Valentina Zadrija | DSC ADRIA 24
Taught by
Data Science Conference