Ryoma Kondo (D3), Takashi Ubukata (D2), Kentaro Matsuura (D1), and Hirofumi Ohzeki (M2), Department of Advanced Interdisciplinary Studies, received Best Performance Award at ITU AI/ML in 5G Challenge Global Round in Japan.
〈Name of award and short explanation about the award〉
ITU AI/ML in 5G Challenge Global Round in Japan "Best Performance Award"
ITU AI/ML in 5G Challenge, conducted by ITU, the International Telecommunications Union, is an international competition to solve 5G network problems with machine learning. In the Japan round the Best Performance Awards were presented to the top three teams that qualified as finalists.
〈About awarded research〉
Autonomous detection of faults within the Internet core network, which is interoperable between internet service providers, is required for high-quality 5G networks. It is difficult to find effective features in a wide range of time-series data and to manage routing path logs in machine learning. We therefore developed a failure detection method based on an effective selection of features using entropy and semantic interpretation of routing information using natural language models.