GTC Taiwan -- NVIDIA today announced that it is collaborating with Taiwan’s Ministry of Science and Technology (MOST) to accelerate the development of artificial intelligence across Taiwan’s commercial sector in support of its recently announced AI Grand Plan to help foster domestic AI-related industries.
The collaboration — kicked off with a jointly hosted AI Symposium during NVIDIA’s GPU Technology Conference in Taiwan, which is being attended by more than 1,400 scientists, developers and entrepreneurs — calls for NVIDIA to help MOST promote AI across Taiwan through five initiatives.
“Taiwan has been the epicenter of the PC revolution, and it will serve as a key center for the next industry revolution focused on AI,” said NVIDIA founder and CEO Jensen Huang. “We are delighted to be working closely with MOST to ensure that Taiwan fully harnesses the power of this technological wave.”
“AI is the key to igniting Taiwan’s next industrial revolution, building on the long-established strength of our IT manufacturing capabilities,” said Dr. Liang-Gee Chen, Minister of Science and Technology. “Our focus is on drawing academics, industry and young talent into our AI Grand Plan to create an ecosystem based on AI innovation.”
Under the agreement, the National Center for High-Performance Computing will build Taiwan’s first AI-focused supercomputer powered by NVIDIA® DGX™ AI computing platforms and Volta architecture-based GPUs. Its target is to create a platform for accelerating advanced research and industry applications that next year reaches 4 petaflops of performance – placing it in the top 25 fastest supercomputers in the Top500 list – and 10 petaflops within four years.
In other steps:
- MOST and NVIDIA’s Deep Learning Institute will train 3,000 developers over the next four years on the use of deep learning in smart manufacturing, the Internet of Things, smart cities and healthcare. Launched last year, the Deep Learning Institute provides hands-on training for developers, data scientists and researchers through self-paced online labs and instructor-led workshops that use open-source frameworks, as well as NVIDIA’s GPU-accelerated deep learning platforms.
- NVIDIA is rolling out domestically its Inception program to help MOST establish its “Youth Technology Innovation and Entrepreneurship Base” for local AI startups. NVIDIA’s Inception program is a virtual incubator for startups focused on AI and deep learning, providing young companies with hardware grants, marketing support and access to NVIDIA’s larger deep-learning ecosystem. Just last week, it added its 2,000th member company.
- NVIDIA will support MOST’s overseas talent training program for post-doctorates by offering high-level internship programs.
- NVIDIA will provide NVIDIA Deep Learning Accelerator (NVDLA) technology for IoT and SoC devices, plus technical support, to MOST’s Project Moon Shot, AI Edge – its NT$4 billion, four-year program to use AI to increase the competitiveness of the domestic semiconductor industry by focusing on memory, sensors and edge products.
And in a related effort, MOST will provide domestic robotics experts with access to NVIDIA DGX Station™ AI deskside supercomputers and NVIDIA Jetson™ TX2 AI modules through the Central and Southern Taiwan Science Parks. NVIDIA is making available DGX-1 systems for MOST’s Formosa Speech Grand Challenge, in which 150 teams from local universities and high schools will compete at the end of October on creating networks capable of Chinese speech recognition. Taiwan’s AI Grand Plan, which was announced in August, aims to create a strong environment for fostering AI innovations and connect with industrial leadership from around the world.
In 1959, the National Science Council was established to promote overall S&T development, academic research and constructing science parks nationwide. The institution was later reorganized and in 2014 became the Ministry of Science and Technology (MOST), aiming to facilitate stronger connections among the industries and boost Taiwanese competitiveness at the international level.