Prof. Jong-Hwan Koh’s Team Won the 2020 AI Grand Challenge for Audio Recognition

The research team led by professor Jong-Hwan Koh at the College of Information and Communication Engineering (IRIS Lab, Members: Tae-Soo Kim, Daniel Noh, Seung-Jin Lee) won the second phase of 2020 AI R&D Grand Challenge for audio recognition. As they won the challenge, they received a Minister prize as well as 700 million won for subsequent research expenses. The challenge was hosted by the Ministry of Science and ICT in December 2020.

The AI R&D Grand Challenge is a challenge where algorithms are developed and competed to solve presented social issue-related challenges, and supports subsequent research funds for teams with outstanding performance. Starting with the first stage of the 2019 challenge, it plans to continue until the fourth stage in 2022.

The team was ranked first with high scores in the “Audio Recognition track”, which uses drones to estimate the gender and direction of origin of the person who requested the rescue, and developed an effective deep learning model that simultaneously performs noise removal and voice classification. After winning third place in the first stage of 2019 and winning the second stage, the team was recognized for its multimedia-based deep learning skills.

Prof. Jong-Hwan Koh said, “The winning of the nation’s most prestigious AI competition for the second consecutive year is meaningful in that AI technology can contribute to solving social issues. We will focus more on research to apply AI technologies, such as development of deep learning model and acceleration system design, in our daily lives.”