LLMs in Media
With media increasingly powered by Artificial Intelligence, mastering the underlying technologies is essential for shaping innovative, responsive, and efficient media solutions. This tutorial provides a comprehensive introduction to AI and Large Language Models (LLMs), offering deep insights and practical know-how for media professionals.
We will explore foundational concepts of AIs, examine state-of-the-art LLMs, and practice model training and fine-tuning techniques. Through interactive hands-on sessions, we will gain experiences with real-world applications — such as LLM-based chatbots with intelligent document retrieval systems (RAG), data analysis with vector embedding, AI-based assistant for search platform, vision-language models for video-chatbots, and interview bots for knowledge capture. We’ll also cover how to secure authenticity and provenance in generative AI with C2PA.
Topics:
- AI & LLM Fundamentals: Conception and overview of AI and Large Language Models.
- Model Training & Manipulation: Techniques for training, fine-tuning, and customizing output of LLMs for media applications.
- Vision-Language Models: Overview and practical use-cases combining visual and textual data.
- Hands-on Applications:
- Developing LLM-based chatbots and intelligent document-based conversational agents (RAG systems).
- Advanced data annotation and semantic analysis with vector embeddings.
- Implementation techniques for AI-based assistant of search platform.
- Implementing video-based chatbots using vision-language models.
- Creating interview bots for knowledge extraction and externalization.
- Securing Authenticity in GenAI: Ensuring trust, provenance, and authenticity in AI-generated media using C2PA standards.
,
Media & Data Science Lead
,
Fraunhofer FOKUS
,
Research Associate
,
Fraunhofer FOKUS