The tool upgrades Data Science Lab for researchers across the school, including new MDS programme.
Students and researchers at the Hertie School are benefitting this fall from a new supercomputing system the school recently procured to support research using data science that informs solutions for major societal problems. The state-of-the-art computing resource, a multi-GPU enterprise deep learning server, is accessible to faculty and researchers across the school’s various master’s and doctoral programmes.
The purchase is part of the Hertie School’s expansion of its data science offerings, including the recent additions of a Data Science Lab and a Master of Data Science for Public Policy (MDS) programme, whose first cohort started their studies in September.
The deep learning and machine learning tool is a powerful and efficient high-performance computing system that allows researchers to experiment and build prototypes for data science-based digital products and services. The server offers one of the most advanced deep learning accelerators and GPU available on the market (NVIDIA A100). It will open up new avenues for scientific exploration in data-driven research at the Hertie School, from Natural Language Processing to Computer Vision, which aims to shed light on major public policy challenges such as climate change or migration.
“The computing unit strengthens the Hertie School’s competitive advantages over other programmes, providing students and researchers with state-of-the-art hardware with which to conduct their fundamental or applied research and build prototypes to solve interesting public policy and data science challenges,” says Huy Ngoc Dang, Manager of the Data Science Lab and Programme Coordinator of the MDS.
Postdoctoral researchers had already been using a similar tool, but the purchase of the new model has considerably upgraded computing resources in deep learning research.
In its mission for sustainability, the Hertie School has also taken steps to offset the heavy electricity demand of supercomputing by utilising completely carbon-neutral energy sources. “We want to make sure we do not harm the environment when we use this resource,” says Huy.