I will be presenting our paper “Calibrating MLLM-as-a-judge via Multimodal Bayesian Prompt Ensembles” at ICCV 2025.