Automating Target-Oriented Data Orchestration at Scale (2025-2029)
Abstract
This project involves developing an automated, scalable data orchestration system, i.e., a system that discovers, enriches, and filters high-quality data to meet diverse targets in designing data-driven solutions. The system drastically reduces the amount of data required to adequately train a model. Functionality, efficiency, effectiveness, and scalability are the project¿TM)s priorities. New knowledge will be generated to automate the entire process of preparing training data, including generating a data pool, assembling datasets, and selecting specific data points to meet performance goals. Eliminating the costs of manually preparing training data will have significant benefits ¿ most of all by fostering a modern and resilient data economy.