With machine learning, algorithms are automatically generated from large datasets, speeding the development and reducing the difficulty of creating complex systems, including robotics systems. While data at scale is what makes accurate machine learning ‘go’, the data used to train ML models must also be very accurate and of high quality.