With a focus on human-centricity in rapidly evolving complex production environments, Industry 5.0 further defines intelligent manufacturing that aims to surpass the current state-of-the-art by enhancing production throughput and reliability through data analytics. While algorithm advances have brought new possibilities, the challenge of data quality hinders their successful implementation. Over the past years, research on data curation has attracted increasing attention to ensure high-quality data for meaningful data analytics. This chapter provides an overview of several key techniques in data curation, highlighting breakthroughs in deep learning–based data denoising, annotation, and balancing. These advancements have shown effective in extracting valuable information from noisy, unannotated, and imbalanced data and improve human comprehension to support the next generation of intelligent manufacturing.