ProxyBNN: Learning Binarized Neural Networks via Proxy Matrices | |
He, Xiangyu1,3; Mo, Zitao1,3; Cheng, Ke1,3; Xu, Weixiang1,3; Hu, Qinghao1,3; Wang, Peisong1,3; Liu, Qingshan2; Cheng, Jian1,3 | |
2020 | |
会议日期 | 2020 |
会议地点 | Online |
英文摘要 | Training Binarized Neural Networks (BNNs) is challenging due to the discreteness. In order to efficiently optimize BNNs through backward propagations, real-valued auxiliary variables are commonly used to accumulate gradient updates. Those auxiliary variables are then directly quantized to binary weights in the forward pass, which brings about large quantization errors. In this paper, by introducing an appropriate proxy matrix, we reduce the weights quantization error while circumventing explicit binary regularizations on the full-precision auxiliary variables. Specifically, we regard pre-binarization weights as a linear combination of the basis vectors. The matrix composed of basis vectors is referred to as the proxy matrix, and auxiliary variables serve as the coefficients of this linear combination. We are the first to empirically identify and study the effectiveness of learning both basis and coefficients to construct the pre-binarization weights. This new proxy learning contributes to new leading performances on benchmark datasets. |
内容类型 | 会议论文 |
源URL | [http://ir.ia.ac.cn/handle/173211/40619] |
专题 | 自动化研究所_模式识别国家重点实验室_图像与视频分析团队 |
通讯作者 | Cheng, Jian |
作者单位 | 1.Institute of Automation, Chinese Academy of Sciences 2.Nanjing University of Information Science and Technology 3.University of Chinese Academy of Sciences |
推荐引用方式 GB/T 7714 | He, Xiangyu,Mo, Zitao,Cheng, Ke,et al. ProxyBNN: Learning Binarized Neural Networks via Proxy Matrices[C]. 见:. Online. 2020. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论