TADIL: Task-Agnostic Domain-Incremental Learning through Task-ID Inference using Transformer Nearest-Centroid Embeddings
Published in International Conference on Pattern Recognition (ICPR), 2024
Recommended citation: Bravo-Rocca, G., et al. (2024). "TADIL: Task-Agnostic Domain-Incremental Learning through Task-ID Inference using Transformer Nearest-Centroid Embeddings." ICPR 2024. Kolkata, India. https://link.springer.com/chapter/10.1007/978-3-031-78110-0_22
This paper presents TADIL, a task-agnostic domain-incremental learning approach that enables 6% performance improvement across diverse domains without requiring task labels. Selected for oral presentation at ICPR 2024 in Kolkata, India. Preprint available on arXiv.
Recommended citation: Bravo-Rocca, G., et al. (2024). “TADIL: Task-Agnostic Domain-Incremental Learning through Task-ID Inference using Transformer Nearest-Centroid Embeddings.” ICPR 2024. Kolkata, India.