[论文解读] Learning Invariant Representation for Continual Learning

Ghada Sokar, Decebal Constantin Mocanu+1 more|University of Twente Research Information|Jan 15, 2021

Domain Adaptation and Few-Shot Learning参考文献 28被引用 4

一句话总结

该论文提出IRCL，一种用于持续学习的伪回放方法，通过条件变分自编码器（cVAE）解耦类不变表征，并将此类表征与任务特定表征联合用于分类。通过生成高质量、类条件的伪样本，并在推理过程中利用不变表征，IRCL在分拆MNIST和Fashion MNIST上的准确率和负向回溯能力上优于正则化方法及当前最先进伪回放方法3.5–5.5%，尤其在回放样本量较小时表现更优。

ABSTRACT

Continual learning aims to provide intelligent agents that are capable of learning continually a sequence of tasks, building on previously learned knowledge. A key challenge in this learning paradigm is catastrophically forgetting previously learned tasks when the agent faces a new one. Current rehearsal-based methods show their success in mitigating the catastrophic forgetting problem by replaying samples from previous tasks during learning a new one. However, these methods are infeasible when the data of previous tasks is not accessible. In this work, we propose a new pseudo-rehearsal-based method, named learning Invariant Representation for Continual Learning (IRCL), in which class-invariant representation is disentangled from a conditional generative model and jointly used with class-specific representation to learn the sequential tasks. Disentangling the shared invariant representation helps to learn continually a sequence of tasks, while being more robust to forgetting and having better knowledge transfer. We focus on class incremental learning where there is no knowledge about task identity during inference. We empirically evaluate our proposed method on two well-known benchmarks for continual learning: split MNIST and split Fashion MNIST. The experimental results show that our proposed method outperforms regularization-based methods by a big margin and is better than the state-of-the-art pseudo-rehearsal-based method. Finally, we analyze the role of the shared invariant representation in mitigating the forgetting problem especially when the number of replayed samples for each previous task is small.

研究动机与目标

• 为解决在无法获取先前任务真实数据时持续学习中的灾难性遗忘问题。
• 开发一种不依赖存储真实过往样本的伪回放方法。
• 通过使用条件生成模型从数据中解耦类不变表征，提升对遗忘的鲁棒性。
• 实现在推理时无需任务标识的类别增量学习。
• 增强知识迁移并减少负面回溯，尤其在回放样本有限时。

提出的方法

• 使用条件变分自编码器（cVAE）从输入数据中解耦出共享的、类不变表征（z）。
• 将不变表征z与任务特定的分类头联合使用以实现识别。
• 通过cVAE生成的、以类标签为条件的样本进行伪回放，替代真实数据回放。
• 模型通过cVAE的重建损失和分类头的分类损失进行端到端训练。
• 将解耦的不变表征输入分类器，以稳定学习并减少遗忘。
• 在推理时无任务标识的类别增量学习设置下评估该方法。

实验结果

研究问题

• RQ1 通过条件生成模型学习到的解耦不变表征是否能减少持续学习中的灾难性遗忘？
• RQ2 当真实数据不可用时，cVAE生成的伪样本质量如何影响性能？
• RQ3 向分类器提供共享不变表征是否能提升泛化能力并减少遗忘，尤其在回放样本较少时？
• RQ4 IRCL在准确率和回溯能力方面与基于正则化和当前最先进伪回放方法相比表现如何？
• RQ5 条件生成是否能实现每类可控的回放样本生成，以平衡学习干扰？

主要发现

• IRCL在分拆MNIST和分拆Fashion MNIST基准上显著优于基于正则化的方法。
• 在分拆MNIST上，IRCL比当前最先进伪回放方法准确率高出3.5%，回溯能力提升5.5%。
• 在分拆Fashion MNIST上，IRCL相比SOTA伪回放方法准确率提升4%，回溯能力提升3%。
• 在分类器中引入共享不变表征（z）能持续提升准确率并减少负面回溯，尤其在回放样本量较小时效果更显著。
• 条件生成产生更高质量的伪样本，有助于提升性能并减少遗忘。
• 消融研究证实，不变表征更不易被遗忘，在维持旧任务性能方面起着关键作用。