arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2006.16627cs.LGcs.NEstat.ML
收藏

在具有随机初始化、固定权重的神经网络内训练高效连接

Training highly effective connectivities within neural networks with randomly initialized, fixed weights

Romanian Institute of Science and Technology(罗马尼亚科学与技术研究院)

Cristian Ivan, Razvan Florian

摘要:

我们提出了一些新的,直接的方法来训练连接图的随机初始化的神经网络没有训练的权重。这些方法不使用定义截止阈值的超参数,因此消除了迭代搜索这种超参数的最佳值的需要。与训练所有权重的情况相比,我们可以实现类似或更高的性能,而计算成本与标准训练技术相似。除了打开和关闭连接之外,我们还介绍了一种通过翻转权重符号来训练网络的新方法。如果我们尝试最小化改变的连接的数量,通过改变不到总数的10%,就已经可以达到标准训练所达到的90%以上的准确率。我们得到了很好的结果,即使与权重的恒定幅度,甚至当权重是从高度不对称的分布。这些结果揭示了神经网络的过度参数化,以及如何将它们减小到有效大小。

英文摘要:

We present some novel, straightforward methods for training the connection graph of a randomly initialized neural network without training the weights. These methods do not use hyperparameters defining cutoff thresholds and therefore remove the need for iteratively searching optimal values of such hyperparameters. We can achieve similar or higher performances than in the case of training all weights, with a similar computational cost as for standard training techniques. Besides switching connections on and off, we introduce a novel way of training a network by flipping the signs of the weights. If we try to minimize the number of changed connections, by changing less than 10% of the total it is already possible to reach more than 90% of the accuracy achieved by standard training. We obtain good results even with weights of constant magnitude or even when weights are drawn from highly asymmetric distributions. These results shed light on the over-parameterization of neural networks and on how they may be reduced to their effective size.

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。

↑