Domain-Aware Continual Zero-Shot Learning
This addresses the challenge of domain shifts in real-world vision tasks for applications like species discovery and wildlife monitoring, representing an incremental improvement with specific gains.
The paper tackles the problem of recognizing images of unseen categories in continuously changing domains, introducing Domain-Aware Continual Zero-Shot Learning (DACZSL) and a Domain-Invariant Network (DIN) that outperforms existing baselines by over 5% in harmonic accuracy and over 1% in backward transfer.
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 propose a Domain-Invariant Network (DIN) to learn factorized features for shifting domains and improved textual representation for unseen classes. DIN continually learns a global shared network for domain-invariant and task-invariant features, and per-task private networks for task-specific features. Furthermore, we enhance the dual network with class-wise learnable prompts to improve class-level text representation, thereby improving zero-shot prediction of future unseen classes. To evaluate DACZSL, we introduce two benchmarks, DomainNet-CZSL and iWildCam-CZSL. Our results show that DIN significantly outperforms existing baselines by over 5% in harmonic accuracy and over 1% in backward transfer and achieves a new SoTA.