The World’s Largest Multimodal Dataset | Episode 4
Consensus-based evaluation in AI datasets
Frederik, Head of Machine Learning at Encord, explains the third stage of building the world's largest multimodal AI dataset. In this episode, Frederik details the creation of an evaluation set of the utmost quality for AI models - using a five-person consensus workflow - and how similar methods can be used to build other high-quality AI datasets.
Speakers
Frederik Hvilshøj
ML Lead @ Encord
The World's Largest Multimodal AI Dataset
The open-source E-MM1 dataset has 100+ million groups of images, videos, text, audio and 3D point clouds, giving AI teams more training data for their AI models.
