Domain-Aware Continual Zero-Shot Learning

arXiv:2112.12989v3 Announce Type: replace-cross
Abstract: Modern visual systems have a wide range of potential applications in vision tasks for natural science research, such as aiding in species discovery, monitoring animals in the wild, and so on. However, real-world vision tasks may experience changes in environmental conditions, leading to shifts in how captured images are presented. To address this issue, we introduce Domain-Aware Continual Zero-Shot Learning (DACZSL), a task to recognize images of unseen categories in continuously changing domains. Accordingly, we …

abstract animals applications arxiv continual cs.cv cs.lg discovery domain environmental experience however images modern monitoring natural research science species systems tasks type vision visual world zero-shot