US-China Collaboration Shapes Global NeurIPS AI Research

WIRED analyzed 5,290 papers presented at last month's NeurIPS and found 141 papers (about 3%) coauthored by researchers at US and Chinese institutions, a similar rate in 2024 (134 of 4,497). It also reports cross-adoption of models — transformers feature in 292 China-author papers, Meta's Llama in 106, and Alibaba's Qwen in 63 — indicating continued cross-border research and shared tooling.
Key Points
- 1Identify collaboration: 141 of 5,290 NeurIPS papers involve US-China coauthorship (~3%).
- 2Document cross-adoption: transformers in 292 China-authored papers; Llama in 106; Qwen in 63.
- 3Advise practitioners: expect ongoing cross-border knowledge transfer affecting model selection and collaboration practices.
Scoring Rationale
Highlights measurable US-China research ties and model cross-adoption, but is based on one conference analysis from a single news outlet.
Sources
Public references used for this report.
Practice with real Logistics & Shipping data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Logistics & Shipping problems
