Learning in deep neural networks and brains with similarity-weighted interleaved learning.
Learning in deep neural networks and brains with similarity-weighted interleaved learning.
复制标题
DOI：
10.1073/pnas.2115229119
复制
发表时间：
2022-07-05
影响因子：
11.1
通讯作者：
McNaughton, Bruce L.
中科院分区：
文献类型：
--
作者：
Saxena, Rajat;Shobe, Justin L.;McNaughton, Bruce L.
与人类不同，人工神经网络在学习新事物时会迅速忘记先前学习的信息，必须通过交错新旧项来重新训练;然而，交错所有旧项是耗时的，并且可能是不必要的。仅交错与新项目具有实质相似性的旧项目可能就足够了。我们发现，使用新旧项目的相似加权交织进行训练，可以让深度网络快速学习新项目而不会忘记，同时使用更少的数据。我们假设如何相似加权交织可能会在大脑中使用持久的兴奋性痕迹最近活跃的神经元和吸引动力学。这些发现可能会推动神经科学和机器学习。了解大脑如何在一生中学习仍然是一个长期的挑战。在人工神经网络（ANN）中，过快地合并新信息会导致灾难性的干扰，即，突然失去以前获得的知识。互补学习系统理论（CLST）认为，新的记忆可以通过将新的记忆与现有的知识交织在一起而逐渐整合到新皮层中。然而，这种方法被认为需要在每次学习新知识时交织所有现有知识，这是不可信的，因为它很耗时，需要大量数据。我们表明，深度，非线性人工神经网络可以学习新的信息，交织只有一个子集的旧项目，共享大量的代表性相似的新信息。通过使用这种相似加权交错学习（SWIL），ANN可以以相似的准确度和最小的干扰快速学习新信息，同时使用每个时期呈现的旧项目数量少得多（快速和数据高效）。SWIL被证明可以与各种标准分类数据集（Fashion-MNIST，CIFAR 10和CIFAR 100），深度神经网络架构和顺序学习框架一起工作。我们发现，学习新项目的数据效率和加速与网络中存储的非重叠类的数量大致成比例地增加，这意味着人类大脑中可能存在巨大的加速，人类大脑编码了大量的单独类别。最后，我们提出了一个理论模型，如何SWIL可能会在大脑中实施。
Unlike humans, artificial neural networks rapidly forget previously learned information when learning something new and must be retrained by interleaving the new and old items; however, interleaving all old items is time-consuming and might be unnecessary. It might be sufficient to interleave only old items having substantial similarity to new ones. We show that training with similarity-weighted interleaving of old items with new ones allows deep networks to learn new items rapidly without forgetting, while using substantially less data. We hypothesize how similarity-weighted interleaving might be implemented in the brain using persistent excitability traces on recently active neurons and attractor dynamics. These findings may advance both neuroscience and machine learning. Understanding how the brain learns throughout a lifetime remains a long-standing challenge. In artificial neural networks (ANNs), incorporating novel information too rapidly results in catastrophic interference, i.e., abrupt loss of previously acquired knowledge. Complementary Learning Systems Theory (CLST) suggests that new memories can be gradually integrated into the neocortex by interleaving new memories with existing knowledge. This approach, however, has been assumed to require interleaving all existing knowledge every time something new is learned, which is implausible because it is time-consuming and requires a large amount of data. We show that deep, nonlinear ANNs can learn new information by interleaving only a subset of old items that share substantial representational similarity with the new information. By using such similarity-weighted interleaved learning (SWIL), ANNs can learn new information rapidly with a similar accuracy level and minimal interference, while using a much smaller number of old items presented per epoch (fast and data-efficient). SWIL is shown to work with various standard classification datasets (Fashion-MNIST, CIFAR10, and CIFAR100), deep neural network architectures, and in sequential learning frameworks. We show that data efficiency and speedup in learning new items are increased roughly proportionally to the number of nonoverlapping classes stored in the network, which implies an enormous possible speedup in human brains, which encode a high number of separate categories. Finally, we propose a theoretical model of how SWIL might be implemented in the brain.
登录
查看更多内容
影响因子：
5.4
作者：
Gepperth, Alexander;Karaoguz, Cem
通讯作者：
Karaoguz, Cem
影响因子：
14.4
作者：
McNaughton, Bruce L.
通讯作者：
McNaughton, Bruce L.
DOI：
10.1073/pnas.1820226116
发表时间：
2019-06-04
影响因子：
11.1
作者：
Saxe, Andrew M.;McClelland, James L.;Ganguli, Surya
通讯作者：
Ganguli, Surya
影响因子：
56.9
作者：
Tse, Dorothy;Langston, Rosamund F.;Morris, Richard G. M.
通讯作者：
Morris, Richard G. M.
影响因子：
2
作者：
DEJONGE, MC;BLACK, J;DISTERHOFT, JF
通讯作者：
DISTERHOFT, JF