Moving from computer vision back to science R&D could signal where the AI industry will head in the coming years.
Neural discovery platform Zeta Alpha has released 2022’s top 100 most cited papers about artificial intelligence. While the United States still dominates the AI research world, there are some other interesting conclusions from the rankings this past year.
The vast majority of papers had authors based in the United States—68 out of 100—but this represents a drop from previous years. US authors made it to the top 100 78 times in 2020. China and the UK have both gained positions in 2022.
Even more interesting, output from leading AI research organizations features a much more diverse region. Google published the most papers overall from 2020-2022, followed very closely by China’s Tsinghua University. Three other Chinese universities and the University of Singapore also captured spots in the top 20.
This could spell even more involvement in the coming years as more world regions explore AI research and development, invest in top talent and equipment, and seek to publish and deploy on a world stage.
See also: Exploring Artificial Intelligence Variants and Their Uses
Science R&D AI retakes top spots followed by visual modeling
The number one most cited paper was a DeepMind contribution for the second year in a row (AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models). Like 2021’s entry (Highly accurate protein structure prediction with AlphaFold), the paper concerned protein structures. AI analyzing protein structure is crucial for several reasons, primarily because it helps advance our understanding of biological processes and has far-reaching implications in various fields such as medicine, biotechnology, and agriculture. It aids drug discovery, diagnosis speeds, and even helps ensure food security by creating things like disease-resistant crops.
2020’s most cited paper dealt with visual models (An Image is Worth 16×16 Words: Transformers for Image Recognition at Scale). Moving from computer vision back to science R&D could signal where the AI industry will head in the coming years. Whatever happens, AI will continue to dominate research.
