The paper ‘ReQAT: Achieving full-precision inference accuracy through 4-bit floating point quantization recognition learning’, in which Jang-hwan Lee, a researcher in the Department of Convergence Electronic Engineering at Hanyang University, participated as the first author, and Professor Jeongwook Choi of the Department of Convergence Electronic Engineering and AI semiconductor company Rebellions, was officially presented at the International Conference on Machine Learning (ICML 2026). The paper was selected as an 'oral presentation' paper, which is considered to be the most innovative and influential among approximately 15,000 submitted papers, and is evaluated as a noteworthy achievement by AI societies around the world. The paper ReQAT: Achieving full-precision inference accuracy through 4-bit floating point quantization recognition learning submitted by Hanyang University and the Rebellion research team was selected as an ICML oral presentation paper. Typically, oral presentations only receive the top 1% of all submitted papers / Source = ReQAT submitted by IT Dong-A Hanyang University and the Rebellion research team: The paper achieving full-precision inference accuracy through 4-bit floating point quantization recognition learning is ICML