Despite being major rivals in the field of artificial intelligence, the US and China have surprisingly close collaboration on cutting-edge research. According to a recent analysis by WIRED, which examined over 5,000 papers presented at the premier AI conference, Neural Information Processing Systems (NeurIPS), there is a notable amount of joint work between researchers from these two countries.
The study found that about 3% of the total papers involve collaboration between authors affiliated with US institutions and those from Chinese institutions. This percentage remains relatively constant over time, indicating a steady level of cooperation between the two nations' AI teams.
Researchers have also discovered that popular algorithms and models developed in one country are frequently shared and adapted across the Pacific. For instance, Google's transformer architecture was featured in 292 papers with authors from Chinese institutions. Meta's Llama family of models appeared in 106 papers featuring researchers from US organizations, while Alibaba's Qwen large language model was mentioned in 63 papers co-authored by researchers from both countries.
According to Jeffrey Ding, an assistant professor at George Washington University who tracks China's AI landscape, the level of collaboration is not surprising. "The US and Chinese AI ecosystems are inextricably enmeshedโand both benefit from the arrangement," he said.
While the analysis may oversimplify the extent of cooperation between researchers, it highlights the complexities of international collaboration in the field of AI. Many Chinese-born researchers study in the US, forging long-lasting bonds with colleagues that transcend their research work.
The analysis also underscores the importance of professional networks and past collaborations in facilitating cooperation across industries. Katherine Gorman, a spokesperson for NeurIPS, noted that "NeurIPS itself is an example of international collaboration and a testament to its importance in our field."
Ultimately, this level of collaboration between US and Chinese AI researchers serves as a reminder that both nations have much to gain from working together, despite current tensions over trade and technology.
The study found that about 3% of the total papers involve collaboration between authors affiliated with US institutions and those from Chinese institutions. This percentage remains relatively constant over time, indicating a steady level of cooperation between the two nations' AI teams.
Researchers have also discovered that popular algorithms and models developed in one country are frequently shared and adapted across the Pacific. For instance, Google's transformer architecture was featured in 292 papers with authors from Chinese institutions. Meta's Llama family of models appeared in 106 papers featuring researchers from US organizations, while Alibaba's Qwen large language model was mentioned in 63 papers co-authored by researchers from both countries.
According to Jeffrey Ding, an assistant professor at George Washington University who tracks China's AI landscape, the level of collaboration is not surprising. "The US and Chinese AI ecosystems are inextricably enmeshedโand both benefit from the arrangement," he said.
While the analysis may oversimplify the extent of cooperation between researchers, it highlights the complexities of international collaboration in the field of AI. Many Chinese-born researchers study in the US, forging long-lasting bonds with colleagues that transcend their research work.
The analysis also underscores the importance of professional networks and past collaborations in facilitating cooperation across industries. Katherine Gorman, a spokesperson for NeurIPS, noted that "NeurIPS itself is an example of international collaboration and a testament to its importance in our field."
Ultimately, this level of collaboration between US and Chinese AI researchers serves as a reminder that both nations have much to gain from working together, despite current tensions over trade and technology.