arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1606.05426cs.CV
收藏

DecomposeMe:简化ConvNets以实现端到端学习

DecomposeMe: Simplifying ConvNets for End-to-End Learning

NICTA(国家信息与通信技术研究院);Data61(Data61数据科学研究所)

Jose Alvarez, Lars Petersson

摘要:

深度学习和卷积神经网络(ConvNets)已成功应用于计算机视觉领域的大多数相关任务。然而,这些网络对计算要求很高,不适合内存和时间消耗相关的嵌入式设备。在本文中,我们提出了DecomposeMe,这是一种使用1D卷积学习特征的简单但有效的技术。所提出的架构使简单性和过滤器共享,从而提高学习能力。在ImageNet和Places 2上进行的一组全面的大规模实验证明了我们的方法能够提高性能,同时显着减少所需参数的数量。值得注意的是,在Places 2上,我们获得了7.7%的相对前1分类准确度的改进,与VGG-B相比,该架构需要的参数减少了92%。该网络也被证明可以通过转换现有网络来推广到其他任务。

英文摘要:

Deep learning and convolutional neural networks (ConvNets) have been successfully applied to most relevant tasks in the computer vision community. However, these networks are computationally demanding and not suitable for embedded devices where memory and time consumption are relevant. In this paper, we propose DecomposeMe, a simple but effective technique to learn features using 1D convolutions. The proposed architecture enables both simplicity and filter sharing leading to increased learning capacity. A comprehensive set of large-scale experiments on ImageNet and Places2 demonstrates the ability of our method to improve performance while significantly reducing the number of parameters required. Notably, on Places2, we obtain an improvement in relative top-1 classification accuracy of 7.7\% with an architecture that requires 92% fewer parameters compared to VGG-B. The proposed network is also demonstrated to generalize to other tasks by converting existing networks.

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。

↑