持续学习：用于视频表示的无遗忘中奖子网络
Continual Learning: Forget-free Winning Subnetworks for Video Representations
AI总结：
受彩票假说启发，本文提出将中奖子网络（WSN）及傅里叶子神经算子（FSO）应用于多种持续学习场景，通过稀疏重用预存权重和紧凑编码视频，有效防止过拟合并显著提升各卷积层的任务性能。
摘要：
受彩票假说（LTH）的启发，该假说强调了在更大规模稠密网络中存在高效子网络，在适当的稀疏条件下，基于任务性能表现优异的中奖子网络（WSN）被考虑用于各种持续学习任务。它利用稠密网络中预存的权重，在任务增量学习（TIL）和任务无关增量学习（TaIL）场景中实现高效学习。在小样本类增量学习（FSCIL）中，设计了一种名为Soft子网络（SoftNet）的WSN变体，以防止在数据样本稀缺时发生过拟合。此外，在视频增量学习（VIL）中考虑了WSN权重的稀疏重用。WSN中还考虑了傅里叶子神经算子（FSO）的使用。它能够实现视频的紧凑编码，并识别在不同带宽下可重用的子网络。我们将FSO集成到用于持续学习的不同架构框架中，包括VIL、TIL和FSCIL。我们全面的实验证明了FSO的有效性，显著提升了各种卷积表示层面的任务性能。具体而言，FSO增强了TIL和FSCIL中的高层性能，以及VIL中的低层性能。
英文摘要：
Inspired by the Lottery Ticket Hypothesis (LTH), which highlights the existence of efficient subnetworks within larger, dense networks, a high-performing Winning Subnetwork (WSN) in terms of task performance under appropriate sparsity conditions is considered for various continual learning tasks. It leverages pre-existing weights from dense networks to achieve efficient learning in Task Incremental Learning (TIL) and Task-agnostic Incremental Learning (TaIL) scenarios. In Few-Shot Class Incremental Learning (FSCIL), a variation of WSN referred to as the Soft subnetwork (SoftNet) is designed to prevent overfitting when the data samples are scarce. Furthermore, the sparse reuse of WSN weights is considered for Video Incremental Learning (VIL). The use of Fourier Subneural Operator (FSO) within WSN is considered. It enables compact encoding of videos and identifies reusable subnetworks across varying bandwidths. We have integrated FSO into different architectural frameworks for continual learning, including VIL, TIL, and FSCIL. Our comprehensive experiments demonstrate FSO's effectiveness, significantly improving task performance at various convolutional representational levels. Specifically, FSO enhances higher-layer performance in TIL and FSCIL and lower-layer performance in VIL.
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。
