Johannes Angebauer on quality in AI applications
Last week, our co-founder and mathematician Johannes Angebauer had the opportunity to be part of the warm-up for the DGQ Quality Day 2024 to be. As part of the Chili con Q event series, Johannes presented the most important facts that quality managers need to know about large language models (LLMs) and AI in his talk “The role of the modern quality manager in the AI revolution”.
Decisive quality factors for AI and LLMs
The presentation focused on current language models and their development. These AI models offer innovative ways to optimize processes by analyzing large amounts of data, deriving meaningful insights from it and making them accessible with language. However, attention must be paid to various quality aspects. Data quality, feedback quality and alignment quality are the decisive quality factors for the successful use of artificial intelligence.
Angebauer also gave an overview of the weaknesses of and possible attacks on language models and raised explicit quality issues relating to data and data protection, feedback and the alignment and system settings of the AI model.
Finally, Johannes presented in-context learning together with retrieval augmented generation (RAG) as a secure and effective technology for connecting language models with company data. You can find out more about language models in the world of quality at Johannes’ presentation on November 7 at the DGQ Quality Day.