NVIDIA Announces AI Collaboration with Taiwan in Fourth Major International Deep Learning Event in a Month
In its fourth major global AI event in a month, NVIDIA today underscored its commitment to helping Taiwan jump start deep learning in an economy best known for its central role in tech’s global supply chain.
NVIDIA founder and CEO Jensen Huang told a packed house of more than 1,300 in Taipei that the company is teaming up with Taiwan’s Ministry of Science and Technology to further the reach of AI on an island whose tech sector has long been dominated by semiconductor and hardware giants such as TSMC, Foxconn, Quanta and Asus.
“In no time in my experience has there been a tech force as powerful as AI that has the ability to revolutionize every industry we know,” said Huang.
His two-hour keynote, filled with ambitious tech demos, kicked off a day of breakout sessions on AI topics led by specialists from around the region and multinational corporations, as well as a fireside chat he led with Taiwan’s Minister of Science and Technology Chen Liang-Gee.
“We want to inspire Taiwan to jump on the AI train, which is still just leaving the station,” he said.
Extensive AI Partnership Unveiled
Huang and Chen’s ministry, known locally as MOST, announced an extensive partnership centered on creating Taiwan’s first AI-focused supercomputer at the National Center for High-Performance Computing as a platform for accelerating advanced research.
Built on the NVIDIA® DGX™ AI computing platforms and Volta architecture-based GPUs, it aims to reach next year 4 petaflops of performance – placing it among the world’s top 25 fastest supercomputers, and then expanding to 2.5x that size within four years.
Other areas of collaboration include hands-on training of 3,000 developers on leading applications of deep learning; rolling out NVIDIA’s Inception program for incubating AI and big-data startups; providing high-level internship opportunities for Taiwanese post-doctoral students; and supporting MOST’s Project Moon Shot, AI Edge, focused on using AI to sharpen the domestic semiconductor market’s competitive position.
Focusing Local AI Efforts on Manufacturing, Education
Speaking at the fireside chat, Chen acknowledged that Taiwan hasn’t established the entrepreneurial tradition of some of its neighbors. And, like many of them, hasn’t made the kinds of strides in software that they have achieved in other fields of technology.
But, he said, the country will do better at teaching entrepreneurial skills and is now pursuing AI in earnest with its Grand Plan, unveiled in late summer, which will build off of its world-class expertise in manufacturing and education.
“Our Grand Plan is to leverage the semiconductor industry to jump into the AI era,” Chen said. “We need to help train AI engineers for industry and spend budget to help universities establish AI research centers.”
In conversation with reporters, Huang was asked whether Taiwan could successfully catch up, given the lead established by the U.S. and China. He noted that countries with less deep IT ecosystems have made great progress – Canada and the U.K. are major centers of AI research, and nations like Singapore are going all in on AI.
“There are lots of reasons to be optimistic,” he said. Among them are the society’s passion around mathematics, the quality of its workforce and its escalating labor costs, which provide more motivation for automation that can bring down costs. “AI, after all, is the automation of automation.”
Demonstrating some of the progress Taiwan is making, two hours of GTC Taiwan were devoted to the nation’s startup scene. Leaders of six young local companies discussed their efforts to apply AI to such areas as retail, entertainment, healthcare and security. They included:
- Viscovery – Uses AI-based image and video recognition advertise to precisely targeted audiences.
- SkyREC – Uses real-time analytics on retail data such as traffic flow, inventory and customer profiles to create valuable insights.
- Glia Studio – Generates news coverage based on neural networks trained on news databases.
- IronYun – Applies AI to big-data analytics for applications in the surveillance industry.
- UmboCV – Deploys deep learning to build facial recognition systems for public safety.
- DYSK Lab – Applies AI to medical pathology.
GTC Taiwan included hands-on training the day before the keynote from the Deep Learning Institute, which provides advanced instruction worldwide in aspects of AI. More than 150 developers and researchers took the local training.
The Taiwan show follows similar events in NVIDIA’s GPU Technology Conference series held since late September in Beijing, Munich and Tel Aviv. Total attendance – including sessions still to come in Washington and Tokyo, along with an earlier edition in Silicon Valley – is expected to exceed 22,000 developers, entrepreneurs and company execs, plus many times more watching live streaming.
Over the course of the day, Huang repeatedly made it clear why he’s taken to the road to barnstorm for AI.
“Because of AI, computational science is just now entering a renaissance,” he said. He noted the exponential rise in recent of years of GPU-accelerated computing, measured by such standards as published papers, downloads of the CUDA development toolkit, as well as massive attendance growth at the company’s GTC events.