How to address novelty concerns in top ai conference? [D]

A computer vision researcher is seeking advice on how to address recurring "novelty" concerns when submitting papers to top-tier AI conferences like NeurIPS, ICLR, and CVPR. They have submitted several papers over the past few years and consistently face this feedback. The researcher is asking for tips or experiences from others to help overcome this challenge.

This post highlights a recurring challenge for researchers submitting to top-tier AI conferences, unlike discussions focusing on acceptance rates or specific technical feedback.