Dgist Develops An Artificial Intelligence Model For Effectively Removing Biases In A Dataset

The research team of Professor Sang-hyun Park at the Department of Robotics and Mechatronics Engineering at the Daegu Gyeongbuk Institute of Science and Technology (DGIST) has developed a new image translation model that can effectively reduce biases in data. This model can remove data biases without the need for information on factors that may cause them, and is expected to facilitate innovations in the fields of self-driving, content creation, and medicine. The model can generate a dataset applying texture debiasing and perform the learning process based on the generated dataset, which can help address challenges related to existing deep learning models.