EPFL researchers have developed a new modular machine learning model, MultiModN, which can learn from various inputs and make flexible predictions. It addresses the challenge of biased missing information and was presented at the NeurIPS2023 conference.
