检索结果-内蒙古大学图书馆

2021 IEEE International Conference on Integrated Circuits, Technologies and Applications, ICTA 2021

作者： Liu, Feng Zhang, Bowen Chen, Gang Gong, Guoliang Lu, Huaxiang Li, Wenchang Chinese Academy of Sciences Institute of Semiconductors Beijing100083 China University of Chinese Academy of Sciences Beijing100089 China Center for Excellence in Brain Science and Intelligence Technology CAS Shanghai200031 China Semiconductor Neural Network Intelligent Perception and Computing Technology Beijing Key Lab Beijing100083 China

ISBN: (纸本)9781665417471

This paper proposes a configurable Sigmoid and Tanh activation function circuit design by second-order approximation and deviation compensation. The proposed circuit is synthesized in TSMC 180nm technology within a 7218 um2 area at 500MHz, it is also deployed on Xilinx ZC706 FPGA by 158 LUTs and 46 FFs at 333MHz. It achieves almost 4.68× better precision and 20.2% less area than previous works. The experimental results show that our method has almost no accuracy loss on LSTM Chinese speech recognition tasks and the YOLOv3 image recognition post-processing task. © 2021 IEEE.

关键词： Chemical activation

来源：评论

学校读者我要写书评

暂无评论

PGGAN: Improve Password Cover Rate Using the Controller

PGGAN: Improve Password Cover Rate Using the Controller

引用

2021 International Conference on Computer network Security and Software Engineering, CNSSE 2021

作者： Guo, Xiaozhou Liu, Yi Tan, Kaijun Jin, Min Lu, Huaxiang Institute of Semiconductors Chinese Academy of Sciences Beijing100083 China University of Chinese Academy of Sciences Beijing100089 China Cas Center for Excellence in Brain Science and Intelligence Technology Chinese Academy of Sciences Beijing200031 China Semiconductor Neural Network Intelligent Perception and Computing Technology Beijing Key Lab Beijing100083 China

Password generation model based on generative adversarial network usually has the problem of high duplicate rate, which further leads to low cover rate. In this regard, we propose PGGAN model. It sets up an additional controller network which is similar to the discriminator in the aspect of structure and function. The discriminator and the controller respectively learn the measure between the distribution of generated password with the real password distribution and the uniform distribution, and then use two measures to teach generator meanwhile. By changing the activation function and loss function of the controller, different measure functions can be selected. The experimental results show that compared with GAN, our PGGAN performs better both in cover rate and duplicate rate. Moreover, Wasserstein distance usually has a better effect to the other measure in model. Specifically, PGGAN with Wasserstein distance can increase the cover rate by 3.57% and reduce the duplicate rate by 30.85% on rockyou dataset. © Published under licence by IOP Publishing Ltd.

关键词： Controllers

来源：评论

学校读者我要写书评

暂无评论

Sdgan: Improve Speech Enhancement Quality by Information Filter 6

Sdgan: Improve Speech Enhancement Quality by Information Fil...

引用

2021 6th International Symposium on Advances in Electrical, Electronics and Computer Engineering, ISAEECE 2021

作者： Guo, Xiaozhou Liu, Yi Mao, Wenyu Li, Jixing Li, Wenchang Gong, Guoliang Lu, Huaxiang Institute of Semiconductors Chinese Academy of Sciences Beijing100083 China University of Chinese Academy of Sciences Beijing100089 China CAS Center for Excellence in Brain Science and Intelligence Technology Chinese Academy of Sciences Beijing200031 China Semiconductor Neural Network Intelligent Perception and Computing Technology Beijing Key Lab Beijing100083 China

The speech denoising model based on adversarial generative network has achieved better results than the traditional machine learning model. In this paper, for the short cut connection in the generator, we discuss its influence on the information transfer between encoder and decoder, and propose SDGAN at target. SDGAN sets linear and convolution filters in the short cut connection which adaptively learn the optimal information processing. The information filter still enables the generator to solve the gradient vanishing problem, and it can also avoid information redundancy and improve expression ability. In addition, SDGAN replaces the L1 regularization term in loss function with the L2 regularization term, which not only makes the output speech of the generator closer to the clean speech, but also avoids sparsity. In the experiments, SDGAN significantly performs better than other traditional GAN in five performance metrics (such as PESQ), and the effect of convolution filter is better than that of linear filter. © Published under licence by IOP Publishing Ltd.

关键词： Speech enhancement

来源：评论

学校读者我要写书评

暂无评论

General nonlinear function neural network fitting algorithm based on CNN 3

General nonlinear function neural network fitting algorithm ...

引用

3rd International Conference on Computer Vision, Image and Deep Learning and International Conference on Computer Engineering and Applications, CVIDL and ICCEA 2022

作者： Xu, Xintao Jiang, Zhelong Chen, Gang Li, Zhigang Gong, Guoliang Lu, Huaxiang University of Science and Technology of China School of Microelectronics Hefei China Institute of Semiconductors Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Materials and Optoelectronics Research Center Beijing China Semiconductor Neural Network Intelligent Perception and Computing Technology Beijing Key Laboratory Beijing China

ISBN: (纸本)9781665459112

This paper proposes a generic neural network fitting algorithm based on CNN for nonlinear functions that overcomes the challenges of a large number of nonlinear functions in terms of hardware deployment and computing circuit generality in diverse neural network models. The model takes advantage of the principle that functions have varying degrees of difficulty fitting in different spaces, mapping the input to high-dimensional space with 1x1 convolution, and utilizing CNN to extract features of nonlinear functions with its strong feature extraction ability in high-dimensional space. Furthermore, MaxPool and ReLU improve the ability of nonlinear fitting. When fitting Tanh, Sigmoid, and ELU activation functions with 16bit accuracy, the proposed algorithm has an average error of less than 0.0006, with a parameter size of 5.793 k. © 2022 IEEE.

关键词： Functions

来源：评论

学校读者我要写书评

暂无评论

Blind image quality assessment based on the multiscale and dual-domains features fusion

Blind image quality assessment based on the multiscale and d...

引用

作者： Lu, Yaxuan Li, Weijun Ning, Xin Dong, Xiaoli Zhang, Liping Sun, Linjun Cheng, Chuantong Institute of Semiconductors Chinese Academy of Sciences Beijing China Beijing Key Laboratory of Semiconductor Neural Network Intelligent Sensing and Computing Technology Beijing China School of Microelectronics University of Chinese Academy of Sciences Beijing China Cognitive Computing Technology Joint Laboratory Wave Group Beijing China

Image quality assessment is to simulate subjective human visual perception and realize image quality inference automatically. Although deep neural networks have achieved great success, the majority of them do not fully consider perception characteristics. Therefore, according to the human visual scale characteristics, we proposed an image quality assessment algorithm based on multiscale and dual domains fusion. Firstly, the original image and its phase congruency respectively input into two branches, feature pyramid and channel attention mechanism are adopted to extract multiscale features. After that, bilinear pool is used to aggregate the spatial and frequency domain characteristics of the corresponding scales, and allows arbitrary scale input to ensure that the features are extracted from the inherent quality images. Finally, the single quality score is obtained through learned weights of each scale. Comparative experiments between our approach and state-of-the-art are conducted on five public databases, the results demonstrate that the proposed algorithm is not only robust to different types and across database, but also sensitive to scale. © 2021 John Wiley & Sons, Ltd.

关键词： Image quality

来源：评论

学校读者我要写书评

暂无评论

Blind separation of noncooperative paired carrier multiple access signals based on improved quantum-inspired evolutionary algorithm and receding horizon optimization

Blind separation of noncooperative paired carrier multiple a...

引用

作者： Deng, Qi Zhang, Shanshan Chen, Gang Lu, Huaxiang Lab of Artificial Networks Institute of Semiconductors CAS Beijing China University of Chinese Academy of Sciences Beijing China CAS Center for Excellence in Brain Science and Intelligence Technology Chinese Academy of Sciences Beijing China Beijing Key Laboratory of Semiconductor Neural Network Intelligent Sensing and Computing Technology Beijing China

The single-channel blind source separation of paired carrier multiple access (PCMA) signal is a key technology in satellite communications. Due to the high-order complexity of existing separation methods and the uncertainty of channel parameters, the processing of noncooperative PCMA signals with long memory remains a great challenge. In this article, the blind separation was solved as the joint channel estimation and sequence detection, and a novel blind separation algorithm based on the improved quantum-inspired evolutionary algorithm (IQEA) and receding horizon optimization was proposed. Considering the practical scenarios, a mixed PCMA signal model with noninteger period sampling was presented in advance. Combined with the proposed signal model, the IQEA was applied to recover the uplink signals from the mixed PCMA signals by its iterative estimation. Meanwhile, to reduce the computational complexity caused by long memory of PCMA signals, the receding horizon optimization was introduced to process the whole signal sequence as a dynamic programming problem. The simulation results verified that the proposed RH-IQEA outperforms existing state-of-the-art separation algorithms in the blind separation of noncooperative PCMA signals, with higher accuracy, lower-order complexity and stronger robustness to parameter estimation errors. © 2021 John Wiley & Sons, Ltd.

关键词： Iterative methods

来源：评论

学校读者我要写书评

暂无评论

Image Quality Assessment based on Dual Domains Fusion

Image Quality Assessment based on Dual Domains Fusion

引用

2020 International Conference on High Performance Big Data and intelligent Systems, HPBD and IS 2020

作者： Lu, Yaxuan Li, Weijun Ning, Xin Dong, Xiaoli Zhang, Yakun Sun, Linjun Institute of Semiconductors Chinese Academy of Sciences Beijing100083 China Beijing Key Laboratory of Semiconductor Neural Network Intelligent Sensing and Computing Technology Beijing100083 China School of Microelectronics University of Chinese Academy of Sciences Beijing100049 China Cognitive Computing Technology Joint Laboratory Wave Group Beijing100083 China

ISBN: (纸本)9781728165110

Image quality assessment is an important and necessary task in the field of image processing. It can simulate human visual perception accurately and effectively to ensure the credibility of information. Although the existing IQA algorithm based on CNN has achieved excellent success, the generalization and robustness of the algorithm are limited due to the loss of image information during feature extraction. The research shows that the phase and amplitude of image frequency domain will change with the quality, so we proposes an image quality assessment algorithm based on dual domains fusion (DualD-IQA). Using the frequency domain and spatial domain as multiple inputs of convolutional neural network, we can complement each other to represent image quality related information. Moreover, the input of any scale can be accepted by adding bilinear pooling, so as to ensure the reliability and robustness of the quality evaluation results. Experimental results show that the algorithm in this paper achieves higher consistency and accuracy in two commonly used public databases, and has higher robustness in different distortion types and cross databases. © 2020 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Novel activation function with pixelwise modeling capacity for lightweight neural network design

Novel activation function with pixelwise modeling capacity f...

引用

作者： Liu, Yi Guo, Xiaozhou Tan, Kaijun Gong, Guoliang Lu, Huaxiang Lab of Artificial Networks Institute of Semiconductors Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China Center for Excellence in Brain Science and Intelligence Technology Chinese Academy of Sciences Shanghai China Beijing Key Laboratory of Semiconductor Neural Network Intelligent Sensing and Computing Technology Beijing China

The development of lightweight networks makes neural networks more efficient to be widely applied to various tasks. Considering the deployment of hardware like edge devices and mobile phones, we prioritize lightweight networks. However, their accuracy has always lagged far behind SOTA networks. In this article, we present a simple yet effective activation function, called WReLU, to improve the performance of lightweight networks significantly by adding a residual spatial condition. Moreover, we use a strategy to switch activation functions after determining which convolutional layer to use. We perform experiments on ImageNet 2012 classification dataset in CPU, GPU, and edge devices. Experiments demonstrate that WReLU improves the accuracy of classification significantly. Meanwhile, our strategy balances the effect of additional parameters and multiply accumulate. Our method improves the accuracy of SqueezeNet and SqueezeNext by more than 5% without increasing extensive parameters and computation. For the lightweight network with a large number of parameters, such as MobileNet and ShuffleNet, there is also a significant improvement. Additionally, the inference speed of most lightweight networks using our WReLU strategy is almost the same as the baseline model on different platforms. Our approach not only ensures the practicability of the lightweight network but also improves its performance. © 2021 John Wiley & Sons Ltd.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

A hardware-friendly low-bit post-training quantization algorithm for lightweight networks

TechRxiv

引用

TechRxiv 2023年

作者： Li, Jixing Zhao, Zhiyuan Chen, Gang Jin, Ming Lu, Huaxiang Institute of Semiconductors Chinese Academy of Sciences Beijing100083 China University of Chinese Academy of Sciences Beijing100089 China Beijing Key Lab of Semiconductor Neural Network Intelligent Perception and Computing Technology Beijing100083 China

Lightweight convolutional neural networks (LCNNs) are commonly quantized and deployed on edge devices to fulfill the requirements of low-power, high-performance tasks. Utilizing uniform, symmetric, per-tensor quantization with power-of-two scale factors can enhance hardware efficiency, albeit at the cost of significant accuracy degradation. In this work, we present a hardware-friendly post-training quantization (HFPTQ) framework to address this problem. HFPTQ employs a synergistic combination of established techniques, including Cross-Layer Equalization (CLE), Absorbing High Bias (AHB), and Adaptive Rounding. Furthermore, HFPTQ introduces quantization-aware equalization by optimizing the equalization scale and absorption vector to minimize block reconstruction error and network prediction difference. Additionally, HFPTQ introduces an adaptive bias correction mechanism. While optimizing weight rounding, bias term is co-optimized to rectify the mean shift in activations induced by quantization. Experimental results, based on the ImageNet dataset, demonstrate that HFPTQ significantly enhances the low-bit hardware-friendly quantization performance of lightweight networks. © 2023, CC BY.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

SSCGAN: Speech style conversion based on GAN

SSCGAN: Speech style conversion based on GAN

引用

2022 International Conference on Algorithms, Microchips and network Applications

作者： Li, Jixing Guo, Xiaozhou Shen, Ronxuan Lu, Huaxiang Wang, Xinggang Cao, Zhanzhong Zhang, Chi Mao, Wenyu Institute of Semiconductors Chinese Academy of Sciences Beijing100083 China University of Chinese Academy of Sciences Beijing100089 China Materials and Optoelectronics Research Center University of Chinese Academy of Sciences Beijing100083 China College of Microelectronics University of Chinese Academy of Sciences Beijing China Center for Excellence in Brain Science and Intelligence Technology Chinese Academy of Sciences Shanghai200031 China Semiconductor Neural Network Intelligent Perception and Computing Technology Beijing Key Laboratory Beijing100083 China Nanjing Research Institute of Information Technology Nanjing210009 China

ISBN: (纸本)9781510653290

Speech conversion has significant applications in medical, robotics, and other industries. With the rise of deep learning, CycleGAN is widely used in speech conversion technology. However, the existing CycleGAN-based methods do not consider the speech signal's temporal and spatial features. In addition, the training of CycleGAN is difficult to converge due to the gradient disappearance problem of the generator. We propose SSCGAN, whose generator is a U-shaped encoder-decoder network that extracts the temporal and spatial features by using 1DCNN and 2DCNN in parallel. A feature fusion module based on multi-scale mixed convolution is embedded between encoder and decoder to achieve high-level fusion of spatial features and temporal features. To make the network training more stable and easier to converge, SSCGAN uses Wasserstein distance instead of the original Jensen-Shannon divergence to calculate the distance of the probability distribution, which can alleviate the gradient extinction problem for generators. In addition, SSCGAN utilizes the PatchGAN structure in the discriminator, which considers the samples' local details by dividing them into different patches. It can improve the discriminative ability of SSCGAN. The experiment results in the nonparallel corpus database VCC 2018 show that SSCGAN is superior to existing methods such as CycleGAN-VC, StarGan-VC. In inter-gender speech conversion, the MSD of SSCGAN is decreased by 0.162 on average compared to other methods, and in intra-gender speech conversion, the MSD is decreased by 0.118 on average. In subjective evaluation, participants also think SSCGAN is the best. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：