The Hardware-Accelerated Learning (HAL) cluster was designed to speed up deep learning research. It focused on data-intensive projects for research in fields such as natural language processing, computer vision, healthcare and more. HAL intended to be transformative in both the system architecture and domain science fields.