1,570 Hits in 4.2 sec

A Threshold Neuron Pruning for a Binarized Deep Neural Network on an FPGA

Tomoya FUJII, Shimpei SATO, Hiroki NAKAHARA
2018 IEICE transactions on information and systems  
Tomoya FUJII †a) , Nonmember, Shimpei SATO †b) , and Hiroki NAKAHARA †c) , Members SUMMARY For a pre-trained deep convolutional neural network (CNN) for an embedded system, a high-speed and a low power  ...  In that case, since the weight memory is realized by an on-chip memory on the FPGA, it achieves a high-speed memory access.  ...  Values program (ACCEL) of JST.  ... 
doi:10.1587/transinf.2017rcp0013 fatcat:ce4s7a7enrcaneqqqexyrjvhea

GUINNESS: A GUI Based Binarized Deep Neural Network Framework for Software Programmers

Hiroki NAKAHARA, Haruyoshi YONEKAWA, Tomoya FUJII, Masayuki SHIMODA, Shimpei SATO
2019 IEICE transactions on information and systems  
tool flow for a binarized deep neural network toward FPGA implementation based on the GUI including both the training on the GPU and inference on the FPGA.  ...  Hiroki NAKAHARA †a) , Haruyoshi YONEKAWA †b) , Tomoya FUJII †c) , Masayuki SHIMODA †d) , and Shimpei SATO †e) , Members SUMMARY The GUINNESS (GUI based binarized neural network synthesizer) is an open-source  ...  Introduction Convolutional deep Neural Network (CNN) Convolutional deep Neural Networks (CNN) are the current state-of-the-art for many embedded computer vision tasks, such as a hand-written character  ... 
doi:10.1587/transinf.2018rcp0002 fatcat:55dvdmcw4zf2zeqrmg2tzm6j4e

NullaNet: Training Deep Neural Networks for Reduced-Memory-Access Inference [article]

Mahdi Nazemi, Ghasem Pasandi, Massoud Pedram
2018 arXiv   pre-print
To cope with computational and storage complexity of these models, this paper presents a training method that enables a radically different approach for realization of deep neural networks through Boolean  ...  Deep neural networks have been successfully deployed in a wide variety of applications including computer vision and speech recognition.  ...  ACKNOWLEDGEMENTS Mahdi Nazemi proposed the idea of NullaNet as well as realizations based on technology mapping and input enumeration.  ... 
arXiv:1807.08716v2 fatcat:n2wmi2fugnbsbi7nmeejrpfvoa

Arrhythmia Classifier using Binarized Convolutional Neural Network for Resource-Constrained Devices [article]

Ao Wang, Wenxing Xu, Hanshi Sun, Ninghao Pu, Zijin Liu, Hao Liu
2022 arXiv   pre-print
In recent years, deep learning and convolutional neural networks have been widely used in the classification of cardiac arrhythmias.  ...  In this paper, a binarized convolutional neural network suitable for ECG monitoring is proposed, which is hardware-friendly and more suitable for use in resource-constrained wearable devices.  ...  Design of Baseline Deep Networks Methods of Binarizing Baseline Networks Design basic blocks A.  ... 
arXiv:2205.03661v2 fatcat:hlefq4dnazhjnblrwoly6twy3m

In-Memory Resistive RAM Implementation of Binarized Neural Networks for Medical Applications [article]

Bogdan Penkovsky, Marc Bocquet, Tifenn Hirtzlin, Jacques-Olivier Klein, Etienne Nowak, Elisa Vianello, Jean-Michel Portal, Damien Querlioz
2020 arXiv   pre-print
The deployment of deep neural networks at the edge is however limited by their high memory and energy consumption requirements.  ...  With new memory technology available, emerging Binarized Neural Networks (BNNs) are promising to reduce the energy impact of the forthcoming machine learning hardware generation, enabling machine learning  ...  Fig. 1 . 1 Illustration of a temporal 1D convolution. II. IN-MEMORY IMPLEMENTATION OF BINARIZED NEURAL NETWORKS A.  ... 
arXiv:2006.11595v1 fatcat:sctd4wl3inap3gtg3tyxep5n4i

Hardware-Efficient Stochastic Binary CNN Architectures for Near-Sensor Computing

Vivek Parmar, Bogdan Penkovsky, Damien Querlioz, Manan Suri
2022 Frontiers in Neuroscience  
With recent advances in the field of artificial intelligence (AI) such as binarized neural networks (BNNs), a wide variety of vision applications with energy-optimized implementations have become possible  ...  We then evaluate benefits of implementing such a pipeline using OxRAM-based circuits for stochastic sampling as well as in-memory computing-based binarized multiplication.  ...  Key contributions of this work are as follows: • Stochastic sampling at input layer based on normal distribution is demonstrated for realizing stochastic binarized convolutional neural network (SBCNN)  ... 
doi:10.3389/fnins.2021.781786 pmid:35069101 pmcid:PMC8766965 fatcat:7xpaoytyg5ftpa4ng3jau3dkyi

An Efficient Ensemble Binarized Deep Neural Network on Chip with Perception-Control Integrated

Wei He, Dehang Yang, Haoqi Peng, Songhong Liang, Yingcheng Lin
2021 Sensors  
To reduce the computing requirements and storage occupancy of the neural network model, we proposed the ensemble binarized DroNet (EBDN) model, which implemented the reconstructed DroNet with the binarized  ...  Lightweight UAVs equipped with deep learning models have become a trend, which can be deployed for automatic navigation in a wide range of civilian and military missions.  ...  Binarized Neural Network Complex feature information can be extracted for difficult deep neural network tasks owing to the large-scale DCNN.  ... 
doi:10.3390/s21103407 pmid:34068351 pmcid:PMC8153352 fatcat:2rzm7stc5va5daxqj43e65fwim

Stochastic Computing for Hardware Implementation of Binarized Neural Networks

Tifenn Hirtzlin, Bogdan Penkovsky, Marc Bocquet, Jacques-Olivier Klein, Jean-Michel Portal, Damien Querlioz
2019 IEEE Access  
Binarized neural networks, a recently discovered class of neural networks with minimal memory requirements and no reliance on multiplication, are a fantastic opportunity for the realization of compact  ...  In this paper, we propose a stochastic computing version of binarized neural networks, where the input is also binarized.  ...  Such memories are fast and compact non volatile memories, which can be embedded at the core of CMOS processes, and therefore provide an ideal technology for realizing in-memory neural networks [2] ,  ... 
doi:10.1109/access.2019.2921104 fatcat:xhk6kxuwp5fdxdc3bwbbnm3qay

MB-CNN: Memristive Binary Convolutional Neural Networks for Embedded Mobile Devices

Arjun Pal Chowdhury, Pranav Kulkarni, Mahdi Nazm Bojnordi
2018 Journal of Low Power Electronics and Applications  
Recent work has shown that binarizing a neural network can significantly improve the memory requirements of mobile devices at the cost of minor loss in accuracy.  ...  This paper proposes MB-CNN, a memristive accelerator for binary convolutional neural networks that perform XNOR convolution in-situ novel 2R memristive data blocks to improve power, performance, and memory  ...  Convolutional Neural Networks A convolutional neural network (CNN) is a deep learning architecture that has proven successful in image classification and recognition [4, 11] .  ... 
doi:10.3390/jlpea8040038 fatcat:qwrw67tx4ffuthzy5xi4o4ee2y

Embedded Binarized Neural Networks [article]

Bradley McDanel, Surat Teerapittayanon, H.T. Kung
2017 arXiv   pre-print
We study embedded Binarized Neural Networks (eBNNs) with the aim of allowing current binarized neural networks (BNNs) in the literature to perform feedforward inference efficiently on small embedded devices  ...  To accomplish this, eBNN reorders the computation of inference while preserving the original BNN structure, and uses just a single floating-point temporary for the entire neural network.  ...  Deep Neural Network Layers A DNN is a machine learning method composed of multiple layers.  ... 
arXiv:1709.02260v1 fatcat:2rqiac5dlvfubeqhkivbqxrp6a

Margin-Aware Binarized Weight Networks for Image Classification [chapter]

Ting-Bing Xu, Peipei Yang, Xu-Yao Zhang, Cheng-Lin Liu
2017 Lecture Notes in Computer Science  
Deep neural networks (DNNs) have achieved remarkable successes in many vision tasks.  ...  For compressing and accelerating deep neural networks, many techniques have been proposed recently.  ...  This work has been supported by the National Natural Science Foundation of China (Grant No. 61633021).  ... 
doi:10.1007/978-3-319-71607-7_52 fatcat:6cyxr2knofdoffdhga35ntjagy

BitFlow-Net: Towards Fully Binarized Convolutional Neural Networks

Lijun wu, Peiqing Jiang, Zhicong Chen, Xu Lin, Yunfeng Lai, Peijie Lin, Shuying Cheng
2019 IEEE Access  
Ren, ‘‘A 34-FPS 698-GOP/s/W binarized deep neural network-based neering, Fuzhou University, China.  ...  ABSTRACT Binarization can greatly compress and accelerate deep convolutional neural networks (CNNs) for real-time industrial applications.  ... 
doi:10.1109/access.2019.2945488 fatcat:sq7hcsiconamdborwecu35bsmq

HadaNets: Flexible Quantization Strategies for Neural Networks

Yash Akhauri
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)  
On-board processing elements on UAVs are currently inadequate for training and inference of Deep Neural Networks. This is largely due to the energy consumption of memory accesses in such a network.  ...  Unlike wider reduced precision neural network models, we preserve the train-time parameter count, thus out-performing XNOR-Nets without a traintime memory penalty.  ...  State-of-the-art Convolutional Neural Networks typically require large amounts of memory and computation.  ... 
doi:10.1109/cvprw.2019.00078 dblp:conf/cvpr/Akhauri19 fatcat:e4ztyufc2bekdhzx46boz2bag4

Digital Biologically Plausible Implementation of Binarized Neural Networks with Differential Hafnium Oxide Resistive Memory Arrays [article]

Tifenn Hirtzlin, Marc Bocquet, Bogdan Penkovsky, Jacques-Olivier Klein, Etienne Nowak, Elisa Vianello, Jean-Michel Portal, Damien Querlioz
2019 arXiv   pre-print
In parallel, the recently proposed concept of Binarized Neural Network, where multiplications are replaced by exclusive NOR (XNOR) logic gates, offers a way to implement artificial intelligence using very  ...  operations of the neural network directly within the sense amplifiers.  ...  Then, we design and simulate a fully binarized neural network based on this memory array.  ... 
arXiv:1908.04066v2 fatcat:uxdfotmafrbdtkpcsj4k5j2qaa

Power Efficient Object Detector with an Event-Driven Camera for Moving Object Surveillance on an FPGA

Masayuki SHIMODA, Shimpei SATO, Hiroki NAKAHARA
2019 IEICE transactions on information and systems  
convolutional neural network, FPGA  ...  Since the event-driven camera output consists of binary precision frames, an all binarized convolutional neural network (ABCNN) can be available, which means that it allows all convolutional layers to  ...  Realization of all binarization leads to area reduction since a single binarized convolutional circuit is shared among all convolutional layers.  ... 
doi:10.1587/transinf.2018rcp0005 fatcat:s4kifuh74vgdrdvxclnhanldf4
« Previous Showing results 1 — 15 out of 1,570 results