Towards Efficient Annotation Databases

Abstract

Recent advances in machine learning have increased the demand for efficient annotation data management for machine learning applications by organizations. In this paper, we address this challenge through an industrial collaboration centered around the unification of data for training and prediction workflows by enabling fast analytical processing through summarization. Beyond this specific solution, we provide a very concrete real-world scenario and solution to the data management community as inspiration for further theoretical and practical research. Finally, we report on the open scientific challenges that remain in this field.