The acceptance rate of CAI-202 1 is half of the normal standard, and the acceptance rate of other conferences such as NeurIPS and ICML is above 20%. I don't think IJCAI's paper is that bad.