UCSD Researchers Propose a General Variational Inference-based Framework (MCD) to Infer the Underlying Causal Models as well as the Mixing Probability of Each Sample