Understanding the nuances behind model predictions is crucial, especially when headlines about Air Canada’s Customer Service AI Chatbot hallucinating a refund policy stand as a burning reminder of the cost when AI goes wrong.
In the fast-evolving landscape of machine learning, various metrics guide our development; like Z-indexes and confidence scores. Although testing doesn’t share the same north star in metrics, QA plays a crucial role in offsetting the inherent unpredictability of models from a “look and feel” point of view.
We’ll explore why models can’t be bug-free, how to tackle classification overfitting, and the delicate balance of training for diverse user objectives.
Whatever your involvement in AI – from interested end user, to data scientist or software developer, this talk organised by Audacia promises insights that will leave you with a deeper understanding of the inner workings of machine learning and NLP models.
Senior Software Engineer, Audacia
