Data Centric AI and foundation models

Both selecting the right AI technique and iteratively training the right AI model are critical to reaching the desired accuracy for an AI use case. At the same time, building the right training data set and using it effectively proves to be equally important in raising AI model accuracy. In this talk, Nicolas Deruytter, CEO of AI expert firm ML6, uses his hands-on experience to show how training data issues can be tackled, combining practical suggestions on best practice way of working with more advanced tooling and modelling approaches to further your raise your AI model accuracy. Nicolas will focus on two aspects of data centric AI: 1. Unlocking the full potential of your data: creating a data set that is sufficiently complete and representative, with data and data labels of good quality 2. Unblocking ML projects that lack (usable) data: especially in case of confidential and/or sensitive data with limitations on cloud storage.