arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2404.00901cs.CV
收藏

略微偏移新类以记住旧类用于视频类增量学习

Slightly Shift New Classes to Remember Old Classes for Video Class-Incremental Learning

University of Electronic Science and Technology of China(电子科技大学)

Jian Jiao, Yu Dai, Hefei Mei, Heqian Qiu, Chuanyang Gong, Shiyuan Tang, Xinpeng Hao, Hongliang Li

AI总结:

针对视频类增量学习中的灾难性遗忘问题,本文提出SNRO方法,通过稀疏采样构建记忆集和提前终止训练来偏移新类特征,在相同内存消耗下提升旧类记忆能力。

摘要:

近期的视频类增量学习通常过度追求新见类别的准确率,并依赖记忆集来缓解旧类别的灾难性遗忘。然而,有限的存储空间只允许存储少量有代表性的视频。因此,我们提出SNRO,它通过略微偏移新类别的特征来记住旧类别。具体而言,SNRO包含Examples Sparse(ES)和Early Break(EB)。ES以较低的采样率进行抽取以构建记忆集,并在未来使用插值来对齐这些稀疏帧。通过这种方式,SNRO在相同内存消耗下存储更多样本,并迫使模型关注更难被遗忘的低语义特征。EB在小训练轮数时终止训练,防止模型过度延伸到当前任务的高语义空间。在UCF101、HMDB51和UESTC-MMEA-CL数据集上的实验表明,SNRO在相同内存消耗下比其他方法表现更好。

英文摘要:

Recent video class-incremental learning usually excessively pursues the accuracy of the newly seen classes and relies on memory sets to mitigate catastrophic forgetting of the old classes. However, limited storage only allows storing a few representative videos. So we propose SNRO, which slightly shifts the features of new classes to remember old classes. Specifically, SNRO contains Examples Sparse(ES) and Early Break(EB). ES decimates at a lower sample rate to build memory sets and uses interpolation to align those sparse frames in the future. By this, SNRO stores more examples under the same memory consumption and forces the model to focus on low-semantic features which are harder to be forgotten. EB terminates the training at a small epoch, preventing the model from overstretching into the high-semantic space of the current task. Experiments on UCF101, HMDB51, and UESTC-MMEA-CL datasets show that SNRO performs better than other approaches while consuming the same memory consumption.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑