RTRA:持续学习中基于正则化方法的快速训练
RTRA: Rapid Training of Regularization-based Approaches in Continual Learning
AI总结:
针对持续学习中的灾难性遗忘问题,该研究提出基于自然梯度优化的RTRA方法改进EWC正则化方案,在iFood251数据集上验证其可在不损失测试性能的前提下提升训练效率,优于现有最优方法。
摘要:
灾难性遗忘(CF)是持续学习(CL)领域的一项重大挑战。在用于缓解CF的基于正则化的方法中,会通过合适的损失函数,在后续任务中惩罚对重要训练参数的修改。我们提出RTRA,这是对广泛使用的弹性权重巩固(Elastic Weight Consolidation,EWC)正则化方案的改进,采用自然梯度进行损失函数优化。我们的方法在不牺牲测试数据性能的前提下,提升了基于正则化方法的训练效率。我们使用iFood251数据集,将所提出的RTRA方法与EWC进行对比。实验表明,RTRA相比现有最优方法具有明显优势。
英文摘要:
Catastrophic forgetting(CF) is a significant challenge in continual learning (CL). In regularization-based approaches to mitigate CF, modifications to important training parameters are penalized in subsequent tasks using an appropriate loss function. We propose the RTRA, a modification to the widely used Elastic Weight Consolidation (EWC) regularization scheme, using the Natural Gradient for loss function optimization. Our approach improves the training of regularization-based methods without sacrificing test-data performance. We compare the proposed RTRA approach against EWC using the iFood251 dataset. We show that RTRA has a clear edge over the state-of-the-art approaches.
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。