Professor stresses cooperation key to advancing AI application

UC Berkeley Professor Pieter Abbeel has the credentials of someone heavily involved in the industrialization of machine learning.

Abbeel is director of UC Berkeley's Robotics Learning Lab, an adviser to Open AI and co-founder and chief scientist at Covariant.ai.

Abbeel shared his thoughts on Saturday at the China-US Entrepreneur and Investment Summit in Santa Clara, California, sponsored by People's Daily Online.

He discussed recent advances and trends in artificial intelligence (AI) and applications in robotic automation, calling for more collaboration worldwide in order to make robots ultimately serve the people.

China in recent years has made serious efforts in pushing forward AI research. According to the Ministry of Industry and Information Technology, China's AI industry output in 2017 was roughly $2.85 billion.

The State Council in July 2017 pledged to make the new generation AI industry a major growth engine, to help improve people's lives by 2020 and make the country the world's center and leader for AI innovation by 2030.

Abbeel said that unlike China, "there is not an overall US strategy here" for AI research.

"In the US, many things are bottom up," he said. "Universities do their open research, it gives lot of opportunities for innovation when you are not constrained to do one or another thing," he explained.

For example, before there was a big breakthrough in deep-learning research in 2005, there were "only a few scholars at a few labs who kept doing their things and believed it was the direction to go," said the AI scientist, adding that the "research environment where individual universities, individual companies and individual professors can decide certain kind of things they want to do is very important for innovation".

Abbeel said he is supervising a group of 25 doctoral students and postdoc fellows plus another 25 undergraduates who also conduct research in his lab. He said the students come from different countries and described the Chinese students as "super strong and always achieving".

Abbeel received a BS/MS in electrical engineering from KU Leuven (Belgium) and his PhD in computer science from Stanford University in 2008.

Besides training PhD students to eventually become tenured professors at universities around the world, Abbeel said regular academic exchanges and scholar visits to Berkeley help disseminate the latest AI research findings and applications.

"It's easier to work locally," he said. "Often we will have visitors from other countries to come working with us for half a year. You always have experts working next to you."

Meanwhile, "we always publish. The articles of our research findings are in archives. Anyone can read it. There are also three important global deep-learning conferences so you can present your work and meet people in the AI field worldwide," he said, emphasizing the importance of continuous collaboration.

Echoing Abbeel, Tang Weihong, vice-president of People's Daily Online, pointed out that cooperation is the only way forward for China and America. She said the summit has focused on the present and future of Chinese and American innovation and venture capital, aiming to provide a platform for the two nations' cooperation in science, technology, innovation and investment.

Regarding China and US cooperation in venture capital, Tang said that it should focus on emerging technologies that benefit the real economy, support early-phase projects and small enterprises and contribute to the establishment of a community with a shared future for mankind.

Contact the writer at junechang@chinadailyusa.com