让模型随时学习新类别，且能灵活预测任意标签组合。
Anytime Continual Learning for Open Vocabulary Classification
- 动态融合微调模型与固定开放词汇模型的预测结果。
- 支持在任意时间接收少量样本后持续优化，无需重新训练。
- 适合需要实时更新、开放场景的图像分类任务。
我们提出一种针对开放词汇图像分类的任意时间持续学习（AnytimeCL）方法。该问题旨在摆脱批量训练和固定模型的限制，要求系统能在任意时刻对任意标签集合进行预测，并在接收到一个或多个新样本时高效更新和改进。尽管目标极具挑战性，我们的方法仍显著优于现有技术。我们提出一种动态加权机制，融合部分微调模型与固定开放词汇模型的预测结果，使系统在仅获得部分标签训练样本时仍可持续提升性能。此外，我们引入注意力加权的PCA特征压缩方法，大幅降低存储与计算开销，同时几乎不影响模型精度。实验验证了方法在学习与推理灵活性方面的优势。代码已开源：https://github.com/jessemelpolio/AnytimeCL。
原文摘要 · Abstract (English)
We propose an approach for anytime continual learning (AnytimeCL) for open vocabulary image classification. The AnytimeCL problem aims to break away from batch training and rigid models by requiring that a system can predict any set of labels at any time and efficiently update and improve when receiving one or more training samples at any time. Despite the challenging goal, we achieve substantial improvements over recent methods. We propose a dynamic weighting between predictions of a partially fine-tuned model and a fixed open vocabulary model that enables continual improvement when training samples are available for a subset of a task's labels. We also propose an attention-weighted PCA compression of training features that reduces storage and computation with little impact to model accuracy. Our methods are validated with experiments that test flexibility of learning and inference. Code is available at https://github.com/jessemelpolio/AnytimeCL.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。