arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1908.07724cs.CLcs.LG
收藏

受限循环神经网络

Restricted Recurrent Neural Networks

Duke University(杜克大学);University of Minnesota(明尼苏达大学)

Enmao Diao, Jie Ding, Vahid Tarokh

摘要:

递归神经网络(RNN)及其变体,如长短期记忆(LSTM)和门控递归单元(GRU),已成为许多研究领域(包括自然语言处理和语音数据分析)中学习连续性在线数据的标准构建块。在本文中,我们提出了一种新的方法来显着减少RNN中的参数数量,同时保持与经典RNN相当甚至更好的性能。新的建议,被称为限制递归神经网络(RRNN),限制对应于输入数据和隐藏状态在每个时间步的权重矩阵共享一个大比例的参数。新架构可以被视为其经典对应物的压缩,但它不需要预训练或复杂的参数微调,这两者都是大多数现有压缩技术中的主要问题。在自然语言建模上的实验表明,在压缩率为50%左右的情况下,限制递归结构与经典递归结构的结果基本相当。特别是,受限LSTM可以用更少的参数来超越经典RNN。

英文摘要:

Recurrent Neural Network (RNN) and its variations such as Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU), have become standard building blocks for learning online data of sequential nature in many research areas, including natural language processing and speech data analysis. In this paper, we present a new methodology to significantly reduce the number of parameters in RNNs while maintaining performance that is comparable or even better than classical RNNs. The new proposal, referred to as Restricted Recurrent Neural Network (RRNN), restricts the weight matrices corresponding to the input data and hidden states at each time step to share a large proportion of parameters. The new architecture can be regarded as a compression of its classical counterpart, but it does not require pre-training or sophisticated parameter fine-tuning, both of which are major issues in most existing compression techniques. Experiments on natural language modeling show that compared with its classical counterpart, the restricted recurrent architecture generally produces comparable results at about 50\% compression rate. In particular, the Restricted LSTM can outperform classical RNN with even less number of parameters.

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。

↑