Incorporating Geo-Diverse Knowledge into Prompting for Increased Geographical Robustness in Object Recognition

arXiv:2401.01482v2 Announce Type: replace-cross
Abstract: Existing object recognition models have been shown to lack robustness in diverse geographical scenarios due to domain shifts in design and context. Class representations need to be adapted to more accurately reflect an object concept under these shifts. In the absence of training data from target geographies, we hypothesize that geographically diverse descriptive knowledge of categories can enhance robustness. For this purpose, we explore the feasibility of probing a large language model for geography-based object …

abstract arxiv class concept context cs.ai cs.cv cs.lg design diverse domain geo knowledge object prompting recognition robustness training type