Naver AI Research Paper Cited Over 30,000 Times... Emerging as a Global AI Company

Naver Headquarters. Photo by Jinhyung Kang aymsdream@

Naver and Naver Cloud's artificial intelligence (AI) research achievements have been recognized worldwide, establishing them as global AI companies.

On the 20th, Naver and Naver Cloud announced that 347 AI research papers were accepted at top-tier conferences over five years from 2019 to this year. The number of accepted AI research papers increased sharply from 29 in 2019 to about 100 each in last year and this year, proving rapid growth in AI technological capabilities over the five years.

The impact of each research is also significant. AI papers published by Naver Cloud and Naver have been cited more than 10,000 times in 2023 alone and over 30,000 times in total according to Google Scholar. Based on the ratio of the top 100 most influential research papers in 2022, they ranked 6th worldwide among AI companies, ahead of Google and Intel.

Notably, they achieved excellent results in the increasingly important field of AI safety research. A study proposing a tool to detect potential personal information leaks in large-scale language models was selected as a ‘top 10% research’ based on paper review scores at NeurIPS (Neural Information Processing Systems) 2023, one of the world’s top three machine learning conferences.

Through collaborative research between the University of T?bingen-Naver Trusted AI Joint Research Center in Germany and Naver Cloud, Naver and the University of T?bingen have been researching and developing methods to eliminate AI bias and improve interpretability since 2021. Naver is also conducting research on ‘Trusted Large-scale AI’ as a core topic at the ‘Seoul National University-Naver Large-scale AI Research Center’ established with Seoul National University. Research on enhancing the reliability of large-scale language models in collaboration with Seoul National University was accepted at ACL (Association for Computational Linguistics) 2023, a global natural language processing conference.

Research for ‘Sustainable and Responsible AI’ also attracted attention. An interview study demonstrating the social value creation potential of large-scale AI through Naver AI’s welfare call service for elderly living alone, ‘Clova Care Call,’ was selected as a ‘Best Paper’ in the top 1% of research at CHI 2023, the premier conference in human-computer interaction. Clova Care Call was also introduced as an exemplary case in a report published last November by the Asia-Pacific Economic Cooperation (APEC) to explain technologies strengthening the ‘Bio-Circular-Green Economy’ model. A study proposing a method to build training datasets to mitigate biased remarks by large-scale AI on socially divisive issues was nominated for Best Paper at ACL 2023.

In August, Naver unveiled ‘HyperCLOVA X,’ an advanced model of its third globally developed proprietary large-scale AI, and began business targeting individuals and enterprises, making AI a hot topic among tech companies. As the only domestic company, Naver participated as a committee member of the NeurIPS and ICML (International Conference on Machine Learning) conferences. In November, Naver was officially invited as one of two Korean companies to the AI Safety Summit hosted by the UK government, where it engaged in discussions on building safe AI with leading AI technology developers such as OpenAI, DeepMind, and Anthropic.

Hajungwoo, head of Naver Cloud AI Innovation Center, said, “Overseas researchers who always read papers published by Naver can easily encounter them, and the number of accepted papers and citations shows that Team Naver’s AI technological capabilities have a significant global impact. Especially as AI begins to be used in everyday life and work, discussions on safe AI become more important. We will strengthen AI safety research and lead the global competition for technological leadership.”