DesCLIP: Robust Continual Learning via General Attribute Descriptions for VLM-Based Visual Recognition

@article{he2026desclip,
  title = {DesCLIP: Robust Continual Learning via General Attribute Descriptions for VLM-Based Visual Recognition},
  author = {Chiyuan He and Zihuan Qiu and Fanman Meng and Linfeng Xu and Qingbo Wu and Hui Li},
  year = {2026},
  journal = {IEEE Transactions on Multimedia},
  eprint = {2502.00618},
  archivePrefix = {arXiv},
  doi = {10.1109/TMM.2026.3664953},
  url = {https://arxiv.org/abs/2502.00618},
}