logo

NAVER’s CLOVA and Papago Published 100 Papers on AI Globally with More than 8,000 Citations Last Year … Team NAVER’s AI Technology to Expand Synergies in the “New Cloud”

- Team NAVER, including Search CIC, published 107 papers … rapid growth compared to 2021 (69 papers)

- CLOVA, Papago to collaborate in a cloud-led tech team … to expand research capabilities and synergies on AI technologies- AI technology research that addresses real-world problems to enhance the competitiveness of NAVER’s technology portfolio

 

January 4, 2023

 

CLOVA and Papago of NAVER Corporation (CEO Choi Sooyeon) published more than 100 regular papers at the global top-tier AI conferences last year. This is an academic achievement that is on par with the world’s best industry-based research organizations, and it shows rapid growth in research capabilities compared to 2021 (69 papers), when it was already recognized for global competitiveness in AI technology. The papers published were also cited more than 8,000 times last year, according to Google Scholar, indicating the overall impact of NAVER’s research results on AI research.

 

CLOVA and Papago’s excellent AI competitiveness is expected to make further progress through closer collaboration within the NAVER cloud-led tech organization to be launched this year. As such, NAVER plans to bring together the technological capabilities of CLOVA, Papago, Works Mobile, and Whale into NAVER Cloud this year.

 

AI technology research that addresses real-world problems to enhance the competitiveness of NAVER’s technology portfolio

 

In particular, the fact that research likely to be applied to real-world services was accepted by the world’s most renowned academic conferences adds to the expectation that NAVER’s technology and service portfolio with integrated technological capabilities will constitute a powerful bridge to global partners.

 

For example, the research on how to efficiently apply large-scale language models to services was accepted at the 2022 Conference on Empirical Methods in Natural Language Processing (EMNLP 2022), the world’s leading academic conference in the area of natural language processing. In this research, NAVER proposed a quantization-based additional learning technique that minimizes loss in generation, understanding, and quality of models and reduces memory use while increasing speed. It is a key technology that can cut the cost of using hyperscale AI in services, strengthening the competitiveness of NAVER’s hyperscale AI service.

 

NAVER also presented a paper at the “EMNLP 2022” on the technology for AI to remember and manage previous conversations shared in long-term conversations. This technology is applied to NAVER’s AI call service for senior citizens who live alone, “CLOVA CareCall”, and provides a more personalized and empathetic conversation experience. At Interspeech 2022, a global speech signal processing conference, NAVER also presented studies on methods to apply emotions to speech synthesis, methods to improve speech recognition, and other studies to contribute to the competitiveness of voice AI services, such as CLOVA Dubbing and CLOVA Note.

 

Research covering various areas of AI, including machine learning and computer vision … efforts to lead the discussion of AI ethics

NAVER also demonstrated noticeable research results across various areas of AI technology, including machine learning and computer vision. It became the first company in Korea to present double-digit numbers of papers at the 2022 International Conference on Learning Representations (ICLR), the leading conference on machine learning, the Computer Vision and Pattern Recognition Conference (CVPR), and the 2022 European Conference on Computer Vision (ECCV), presenting 13, 14, and 13 papers, respectively.

 

NAVER also leads the discussion on AI ethics, not only contributing to establishing Korea’s AI ethical standards by participating in the Ministry of Science and ICT’s forum on AI ethical standards but also holding a workshop and sharing insights with experts on hyperscale AI ethics at the ACM [Association for Computing Machinery] Conference on Fairness, Accountability, and Transparency (ACM FAccT), a world-leading academic society on fairness in computer science, in June 2022 attended by about 100 researchers from Google, OpenAI, DeepMind, etc.

 

As the collaboration between various technology organizations becomes closer on the cloud platform, NAVER plans to accelerate the development of service synergies and research capabilities. Ha Jung-woo, the head of NAVER AI Lab, said, “NAVER’s research competitiveness on AI technology recognized globally will create a synergy in the integrated tech organization, stepping to the next level. NAVER’s AI technology portfolio, with its advantage of solving real-world problems, will expand to the global market more actively and provide value to users around the world.”

 

NAVER is also releasing innovative AI research papers via Global AI R&D Belt established with research institutes in Korea and abroad. Last year, it presented multiple papers in collaboration with prominent research groups from Seoul National University, Korea Advanced Institute of Science and Technology (KAIST), New York University, Carnegie Mellon University, and EURECOM at international conferences.

 

NAVER also continues to expand the AI R&D Belt. In April last year, it came up with a joint research space with the KAIST Graduate School of AI at its second office building, “1784,” for effective industry–academia collaboration. NAVER held the first AI hackathon in Vietnam with the Hanoi University of Science and Technology (HUST) in May to strengthen exchanges with AI talents in Vietnam and started working together with Canada’s University of Toronto to develop an AI-based next-generation content platform in October. This research also involves a top global web fiction platform, Wattpad, and the result of the research will be applied to Wattpad’s content service platform to introduce global users to a more innovative experience.

 

-END-