Naver Achieves Success at Global AI Conference with 51 Papers Accepted

[Asia Economy Reporter Buaeri] Naver announced on the 2nd that a total of 51 regular papers were accepted at AI conferences this year.

The company explained that with 43 papers in the first half of the year alone, it caught up with last year's paper publication record in just half a year.

The papers showed achievements in various fields such as computer vision, natural language processing, deep learning, and speech technology.

In the speech field's top conferences, ICASSP and Interspeech, 9 papers were accepted respectively. Most recently, 8 papers were confirmed for presentation at ICCV, the most prestigious computer vision conference, recognizing their technological prowess in the vision field. In addition, papers were presented or are scheduled to be presented at many other conferences such as CHI, NAACL, ICML, and KDD.

More than 40% of the papers accepted at the conferences have been applied in various ways to Naver's actual services. The speaker recognition technology research results presented at ICASSP were used to enhance speaker-specific speech recognition in 'Clova Note.' The image recognition model 'RexNet' presented at CVPR and the deep learning optimization technique 'AdamP' presented at ICLR are core technologies of 'X-eye 2.0,' which filters obscene content. In addition, AI research results have been integrated into various services and solutions such as 'Clova Dubbing,' 'Clova Forecast,' and 'Clova Care Call.'

Naver collaborated with leading domestic universities such as Yonsei University, Korea University, Gwangju Institute of Science and Technology (GIST), Inha University, and Ulsan National Institute of Science and Technology (UNIST) for AI research. It established joint research centers with about 100 researchers each with Seoul National University and KAIST to closely cooperate on next-generation AI research.

Overseas, Naver established joint research centers with Vietnam's HUST and PTIT, and recently signed a Memorandum of Understanding (MOU) with the University of T?bingen in Germany for research on 'Trustworthy AI,' continuously expanding its global AI research and development (R&D) ecosystem. Among the 51 papers accepted this year, 21 are the results of joint research through industry-academia cooperation.

Naver plans to continue active collaboration with academia while accelerating AI research centered on 'hyper-scale AI' technology. Naver is commercializing the hyper-scale AI 'HyperCLOVA' across services such as search query correction and shopping review summarization. The 'no-code AI' tool HyperCLOVA Studio has also begun internal beta testing.

Ha Jung-woo, head of Naver AI Lab, said, "As a leading AI company, we will strive to secure global AI competitiveness through in-depth mid- to long-term advanced research and to further strengthen the domestic AI technology ecosystem."