Filters








10,153 Hits in 4.9 sec

Soft Realization: a Bio-inspired Implementation Paradigm [article]

Hamid Reza Mahdiani (Computer Science and Engineering Department, Shahid Beheshti University, Tehran, IRAN-Institute for Cognitive and Brain Science, Shahid Beheshti University, Tehran, Iran), Mahdi Nazm Bojnordi, Sied Mehdi Fakhraie (Electrical and Computer Engineering Department, University of Tehran, Tehran, IRAN)
2018 arXiv   pre-print
The proposed paradigm mitigates major weaknesses of hard realization by (1) alleviating incompatibilities with today's soft and bio-inspired algorithms such as artificial neural networks, fuzzy systems  ...  Moreover, the incoming nanotechnologies face increasing reliability issues that prevent them from being efficiently exploited in hard realization of applications.  ...  Acknowledgements We would like to thank profeesor Caro Lucas (RIP) for his helpful discussions and advice, as well as his kind and great support during establishmnet of the Soft realization idea foundations  ... 
arXiv:1812.08430v1 fatcat:voc6sqpsjfa5hn2tztw5vq3ef4

Dynamic Neural Networks: A Survey [article]

Yizeng Han, Gao Huang, Shiji Song, Le Yang, Honghui Wang, Yulin Wang
2021 arXiv   pre-print
Dynamic neural network is an emerging research topic in deep learning.  ...  adaptive inference along the temporal dimension for sequential data such as videos and texts.  ...  For example, in hierarchical multi-scale recurrent neural network (HM-RNN) [162] , when the low-level (character-level) model detects that the input satisfies certain conditions, it will "flush" (reset  ... 
arXiv:2102.04906v4 fatcat:zelspxwv6nel7kv2yu6ynakyuu

Neutrosophy for physiological data compression: in particular by neural nets using deep learning

Philippe Schweizer
2020 Zenodo  
Deep learning neural networks have recently exceeded the compression rates of algorithmic techniques for text.  ...  The pace of data production continues to grow, leading to increased needs for efficient storage and transmission.  ...  Finally for the videos a temporal redundancy is also exploited, an image of a film varying only little from the previous one.  ... 
doi:10.5281/zenodo.3988027 fatcat:ju6xbcnt5ngdbefhlq3h2ojfga

Self-Compression in Bayesian Neural Networks [article]

Giuseppina Carannante, Dimah Dera, Ghulam Rasool, Nidhal C. Bouaynaya
2021 arXiv   pre-print
We show that Bayesian neural networks automatically discover redundancy in model parameters, thus enabling self-compression, which is linked to the propagation of uncertainty through the layers of the  ...  Current methods compress the networks by reducing the precision of the parameters or by eliminating redundant ones.  ...  We exploit the uncertainty information inherent in Bayesian neural networks and propose a method for self-compression of convolutional neural networks.  ... 
arXiv:2111.05950v1 fatcat:u3fl6pooonghrj2e6haoolibju

1D-FALCON: Accelerating Deep Convolutional Neural Network Inference by Co-optimization of Models and Underlying Arithmetic Implementation [chapter]

Partha Maji, Robert Mullins
2017 Lecture Notes in Computer Science  
Deep convolutional neural networks (CNNs), which are at the heart of many new emerging applications, achieve remarkable performance in audio and visual recognition tasks, at the expense of high computational  ...  We also introduce a new class of fast 1-D convolutions for CNNs using the Toom-Cook algorithm.  ...  [14] proposed a Sparse Convolutional Neural Networks (SCNN) model that exploits both inter-channel and intra-channel redundancy to maximize sparsity in a model.  ... 
doi:10.1007/978-3-319-68612-7_3 fatcat:p65rpuqlzreffca4rwtrvmaeeu

NestedNet: Learning Nested Sparse Structures in Deep Neural Networks

Eunwoo Kim, Chanho Ahn, Songhwai Oh
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
In this work, we propose a novel deep learning framework, called a nested sparse network, which exploits an n-in-1-type nested structure in a neural network.  ...  While many recent works focus on reducing the redundancy by eliminating unneeded weight parameters, it is not possible to apply a single deep network for multiple devices with different resources.  ...  In this work, we aim to exploit a nested structure in a deep neural architecture which realizes an n-in-1 versatile network to conduct multiple tasks within a single neural network (see Figure 1 ).  ... 
doi:10.1109/cvpr.2018.00904 dblp:conf/cvpr/KimAO18 fatcat:mrnzlehq2zcfvgwnab2gvwf6om

Building Fast and Compact Convolutional Neural Networks for Offline Handwritten Chinese Character Recognition [article]

Xuefeng Xiao, Lianwen Jin, Yafeng Yang, Weixin Yang, Jun Sun, Tianhai Chang
2017 arXiv   pre-print
We design a nine-layer CNN for HCCR consisting of 3,755 classes, and devise an algorithm that can reduce the networks computational cost by nine times and compress the network to 1/18 of the original size  ...  Such networks intuitively appear to incur high computational cost, and require the storage of a large number of parameters, which renders them unfeasible for deployment in portable devices.  ...  [12] exploited the cross-channel or filter redundancy to formulate a low-rank basis for filters, and proposed filter and data reconstruction techniques for optimization. Zhang et al.  ... 
arXiv:1702.07975v1 fatcat:paxuegfmg5azdd445xd7esstru

Efficient Deep Learning in Network Compression and Acceleration [chapter]

Shiming Ge
2018 Digital Systems  
In this chapter, I will present a comprehensive survey of several advanced approaches for efficient deep learning in network compression and acceleration.  ...  While deep learning delivers state-of-the-art accuracy on many artificial intelligence tasks, it comes at the cost of high computational complexity due to large parameters.  ...  Network densifying Another direct category for obtaining network compression and acceleration is to design more efficient but low-cost network architecture.  ... 
doi:10.5772/intechopen.79562 fatcat:ya65wwhk5neppgxrut5phd42dy

Wavelet-based Enhanced Medical Image Super-Resolution

Farah Deeba, She Kun, Fayaz Ali Dharejo, Yuanchun Zhou
2020 IEEE Access  
To solve this problem, we propose a wavelet-based mini-grid network medical image super-resolution (WMSR) method, which is similar to the three-layer hidden-layer-based super-resolution convolutional neural  ...  In order to ensure the reproducibility of the image, a method of adding a sub-pixel layer is proposed to realize the hidden layer, and replacing the small mini-grid-network on the hidden layer is of considerable  ...  Additionally, mini-grid-network is added for exploiting the speed of the network due to its good quick performance on the network [29] .  ... 
doi:10.1109/access.2020.2974278 fatcat:ovgfi6ygnjeerd4522v4qndezi

Towards Modality Transferable Visual Information Representation with Optimal Model Compression [article]

Rongqun Lin, Linwei Zhu, Shiqi Wang, Sam Kwong
2020 arXiv   pre-print
Although numerous approaches were developed for improving the image and video coding performance by removing the redundancies within visual signals, much less work has been dedicated to the transformation  ...  of the visual signals to another well-established modality for better representation capability.  ...  neural network for better signal representation capability.  ... 
arXiv:2008.05642v1 fatcat:cam5jnjcorelfaiy2p3s6hbgki

Spiking Neural Networks Hardware Implementations and Challenges

Maxence Bouvier, Alexandre Valentian, Thomas Mesquida, Francois Rummens, Marina Reyboz, Elisa Vianello, Edith Beigne
2019 ACM Journal on Emerging Technologies in Computing Systems  
They are expected to improve the computational performance and efficiency of neural networks, but are best suited for hardware able to support their temporal dynamics.  ...  In this survey, we present the state of the art of hardware implementations of spiking neural networks and the current trends in algorithm elaboration from model selection to training mechanisms.  ...  enable very large-scale neural simulation at low cost, both in time and power.  ... 
doi:10.1145/3304103 fatcat:p3frra3osnhybj4hkor4y5cyqm

Approximated Prediction Strategy for Reducing Power Consumption of Convolutional Neural Network Processor

Takayuki Ujiie, Masayuki Hiromoto, Takashi Sato
2016 2016 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)  
Convolutional neural network (CNN) is becoming popular because of its great ability for accurate image recognition.  ...  The LCP exploits redundancy of operations in CNN and only executes essential convolutions by an approximated prediction technique.  ...  Computational cost of App-Conv must be low to realize power reduction through LCP.  ... 
doi:10.1109/cvprw.2016.113 dblp:conf/cvpr/UjiieHS16 fatcat:lt4h4cx2jzfxdjpsqo3qcj26jq

On Moving Target Techniques for Network Defense Security

Shouq Mohsen Alnemari, Sabah M Alzahrani
2021 International journal of recent technology and engineering  
The traditional technologies, tools and procedures of any network cannot be protected from attackers due to the unchanged services and configurations of the networks.  ...  Adversarial neural networks are neural network entries that lead to false classification results.  ...  A defender designs space is a group of deep neural networks that have been made for the same mission, but it not influenced by the aforesaid attack.  ... 
doi:10.35940/ijrte.e5111.019521 fatcat:stts5i22abfvzpcf3cawvjvpmy

Encoding Complete Body Models Enables Task Dependent Optimal Behavior

Oliver Herbort, Martin V. Butz
2007 Neural Networks (IJCNN), International Joint Conference on  
This paper shows that our approach accounts for two forms of effective human behavior based on exploiting kinematic redundancy.  ...  We propose a neural network architecture (SURE REACH) that acquires complete body models through unsupervised learning.  ...  Acknowledgments The authors are grateful for the fruitful discussions with their colleagues in Würzburg. This work was supported by the European commission contract no. FP6-511931.  ... 
doi:10.1109/ijcnn.2007.4371203 dblp:conf/ijcnn/HerbortB07 fatcat:dsiflmathrgtbjabav6s77b2ve

Design and Realization of an Efficient Large-Area Event-Driven E-Skin

Florian Bergner, Emmanuel Dean-Leon, Gordon Cheng
2020 Sensors  
Overall, this work develops a systematic approach towards realizing a flexible event-driven information handling system on standard computer systems for large-scale e-skin with detailed descriptions on  ...  While our previous works focused on the implementation and the experimental validation of the approach, this work now provides the consolidated foundations for realizing, designing, and understanding large-area  ...  Sebastian Stenner for their technical assistance, and Mr. J. Rogelio Guadarrama-Olvera for his support with the H1 robot.  ... 
doi:10.3390/s20071965 pmid:32244511 pmcid:PMC7180917 fatcat:acaii3peyzemlmsozcwxmx4pga
« Previous Showing results 1 — 15 out of 10,153 results