Theoretical Understanding of the Information Flow on Continual Learning Performance
Theoretical Understanding of the Information Flow on Continual Learning Performance
复制标题
DOI：
10.48550/arxiv.2204.12010
复制
发表时间：
2022-04
期刊：
影响因子：
--
通讯作者：
Joshua Andle;S. Y. Sekeh
中科院分区：
文献类型：
--
作者：
Joshua Andle;S. Y. Sekeh
持续学习（CL）是一种设置，其中代理必须顺序地从传入的数据流中学习。CL性能评估模型在保留先前知识的同时，随着时间的推移不断学习和解决新问题的能力。尽管有许多以前的解决方案，以绕过灾难性遗忘（CF）的先前看到的任务在学习过程中，他们中的大多数仍然遭受严重的遗忘，昂贵的内存成本，或缺乏理论上的理解，神经网络的行为，而学习新的任务。虽然CL性能下降的问题，在不同的训练制度下已被广泛的经验研究，从理论的角度还没有足够的关注。在本文中，我们建立了一个概率框架来分析信息流通过网络层的任务序列及其对学习性能的影响。我们的目标是优化层之间的信息保存，同时学习新的任务，以管理在整个层中传递的特定于任务的知识，同时保持先前任务的模型性能。特别是，我们研究了CL性能与网络中信息流的关系，以回答“如何利用层间信息流的知识来缓解CF？”".我们的分析提供了新的见解，在增量任务学习过程中的层内的信息适应。通过我们的实验，我们提供了经验证据，并实际上突出了跨多个任务的性能改善。
Continual learning (CL) is a setting in which an agent has to learn from an incoming stream of data sequentially. CL performance evaluates the model's ability to continually learn and solve new problems with incremental available information over time while retaining previous knowledge. Despite the numerous previous solutions to bypass the catastrophic forgetting (CF) of previously seen tasks during the learning process, most of them still suffer significant forgetting, expensive memory cost, or lack of theoretical understanding of neural networks' conduct while learning new tasks. While the issue that CL performance degrades under different training regimes has been extensively studied empirically, insufficient attention has been paid from a theoretical angle. In this paper, we establish a probabilistic framework to analyze information flow through layers in networks for task sequences and its impact on learning performance. Our objective is to optimize the information preservation between layers while learning new tasks to manage task-specific knowledge passing throughout the layers while maintaining model performance on previous tasks. In particular, we study CL performance's relationship with information flow in the network to answer the question"How can knowledge of information flow between layers be used to alleviate CF?". Our analysis provides novel insights of information adaptation within the layers during the incremental task learning process. Through our experiments, we provide empirical evidence and practically highlight the performance improvement across multiple tasks.