The National Supercomputing Centre (NSSC) and NVIDIA will establish a shared facility to provide training and technical expertise to 400 artificial intelligence (AI) researchers and developers from AI Singapore (AISG).
The NSSC and NVIDIA yesterday (24 October 2017) signed an agreement to establish the AI.Platform@NSSC, which aims to help speed up the deployment of planned projects of AISG.
AISG is a national programme launched last May to drive AI adoption, research, and innovation in the republic. It has brought together tertiary organisations National University of Singapore (NUS), Nanyang Technological University (NTU), Singapore University of Design and Technology (SUTD), and Singapore Management University (SMU), as well as the research institutions under the Agency for Science, Technology and Research (A*STAR).
"AISG aims to deliver 100 meaningful AI projects and proofs-of-concept to solve real-world problems swiftly. As a complementary effort, the training of 400 AI researchers and developers will certainly expedite the realisation of AISG's goals," said Professor Ho Teck Hua, executive chairman of AISG and deputy president (Research & Technology) and Tan Chin Tuan Centennial Professor at NUS.
The AI.Platform@NSSC will be located at NSSC's data centre in Connexis Building, Fusionopolis and launched early next year.
It will be equipped with six units of NVIDIA's DGX-1 deep learning supercomputers, which is designed specifically for AI research. NVIDIA said the supercomputers are packed with 48 Tesla V100 GPUs that can provide up to 99x more throughput than standard CPUs. This thus offers users with up to six Tensor petaflops of AI computing power to run their AI frameworks and algorithm, and conduct proofs of concept.
"NVIDIA is helping industries transform with deep learning using our GPU computing technologies. We are delighted to partner with NSCC to establish this shared facility to support AISG's mission, giving them a powerful platform for their AI projects and innovations in areas such as intelligent video analytics, natural language processing, genomics and life sciences, and cybersecurity," said Raymond Teh, vice president of NVIDIA in Asia Pacific.
Furthermore, NVIDIA will also hold hands-on deep learning training with at least 400 developers, data scientists, and researchers through its Deep Learning Institute (DLI). Participants will be trained on the latest techniques for designing, training, and deploying neural networks across a variety of application domains.