However,most existing studies concentrate onthetext-to-image synthesis [11-14], which generates images from text descriptions without the original image.

People are relying on AI agents to assist them with various tasks.

This involves introducing specific prompts such as "in

As a crucial step for analyzing human actions, hand-object interaction understanding is researchworthy in a broad range of applications related to virtual or augmented reality. Current works largely focus on hand-object pose estimation (HOPE) [16, 19, 21], which aims to capture the pose configuration of the given hand-object image.

A.21 SocietalImpact Our proposed SALKG approach for learning from KG explanations can be applied to any KGaugmented model and can be adapted from any off-the-shelf saliency method. This enables KGaugmented models to improve generalization ability and learn more efficiently from data, thus yielding better performance while requiring less labeled data.