Naver Cloud recognized for hyperscale AI research achievements

[Courtesy of Naver]

South Korea’s Naver Cloud Co. is making waves in the field of artificial intelligence (AI) with its groundbreaking HyperCLOVA X, as it continues to receive widespread recognition for its impressive research accomplishments in various domains, including multimodal analysis, lightweight AI modeling and ethical considerations.

According to Naver Cloud on Tuesday, the company achieved a milestone in the first half of this year by having 61 research papers accepted at prestigious AI conferences worldwide. This feat demonstrates the company’s commitment to pushing the boundaries of AI research, matching the rapid pace it set last year with the publication of 100 papers.

Notably, Naver’s efforts in the hyperscale AI field have gone beyond mere paper acceptance, with several research studies earning recognition as “best papers” and receiving prestigious accolades. One notable study proposed innovative evaluation metrics for assessing the performance of image generation models, securing its place as a top-tier research endeavor at the ICLR 2023, the premier machine learning conference. This research by Naver Cloud introduces groundbreaking metrics to measure the diversity and creativity of images produced by hyperscale AI systems.

Naver Cloud’s research on lightweight hyperscale AI also garnered attention at ICML 2023, a renowned global machine learning conference. The company’s cutting-edge proposal for novel quantization techniques, aimed at reducing the operational costs of hyperscale language models, has significant implications. These advancements will not only optimize Naver’s in-house development of HyperCLOVA X but also contribute to the overall efficiency of hyperscale AI.

Naver Cloud has emerged as a leading proponent in the discussion of AI ethics. At ACL 2023, a distinguished international conference on natural language processing, Naver Cloud presented a research study that introduces learning datasets and protocols to mitigate AI’s potential biased statements in contentious social issues. The study, which evolved from a year-long operation of an “AI Ethics Forum” involving experts from diverse fields including social sciences and law, has been acknowledged as a significant contribution to the field.

Additionally, a groundbreaking interview-based study validating the societal value of Naver’s AI well-being call service, “Clova Care Call,” designed specifically for elderly individuals, has received the “Best Paper” award at CHI 2023, a preeminent conference on human-computer interaction.

Yun Sang-doo, research director of Naver Cloud AI Lab, expressed his enthusiasm for the company’s achievements, saying that “throughout the first half of this year, we have presented meaningful research results across a range of conferences, spanning machine learning, speech recognition, computer vision, and more.”

He also emphasized Naver Cloud’s commitment to collaboration with esteemed research institutions such as Seoul National University, KAIST, and New York University to further enhance the capabilities of hyperscale AI technology. By partnering with renowned institutions like Tübingen University and the University of Washington, Naver Cloud aims to establish a solid foundation for building even more reliable and trustworthy hyperscale AI systems.