arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1306.0543cs.LGcs.NEstat.ML
收藏

预测深度学习中的参数

Predicting Parameters in Deep Learning

University of Oxford(牛津大学);University of British Columbia(不列颠哥伦比亚大学);Université de Montréal(蒙特利尔大学);Facebook Inc.(脸书公司)

Misha Denil, Babak Shakibi, Laurent Dinh, Marc'Aurelio Ranzato, Nando de Freitas

摘要:

我们证明了在几个深度学习模型的参数化中存在显著的冗余。对于每个特征,仅给定几个权重值,就可以准确地预测剩余的值。此外,我们表明,不仅可以预测的参数值,但他们中的许多人根本不需要学习。我们训练几种不同的架构,只学习一小部分权重,并预测其余的权重。在最好的情况下,我们能够预测超过95%的网络权重,而不会降低准确性。

英文摘要:

We demonstrate that there is significant redundancy in the parameterization of several deep learning models. Given only a few weight values for each feature it is possible to accurately predict the remaining values. Moreover, we show that not only can the parameter values be predicted, but many of them need not be learned at all. We train several different architectures by learning only a small number of weights and predicting the rest. In the best case we are able to predict more than 95% of the weights of a network without any drop in accuracy.

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。

↑