Byfunctioning overaregion-to-regiongraph,bothmodules can be individualized and adapted to coordinate with visual patterns in each image, guided by specific knowledge forms.

VIENNA performs robust on unseen environments, through learning task-shared, general representations.

Methods for social gaze prediction in the literature adopt one of two approaches.

Semantic cues and statistical regularities in real-world environment layouts can improve efficiency for navigation in novel environments.

Self-supervised representation learning approaches haverecently surpassed their supervised learning counterparts on downstream tasks like object detection andimage classification.