Filters








74 Hits in 5.8 sec

Stabilizing DARTS with Amended Gradient Estimation on Architectural Parameters [article]

Kaifeng Bi, Changping Hu, Lingxi Xie, Xin Chen, Longhui Wei, Qi Tian
2020 arXiv   pre-print
Our approach bridges the gap from two aspects, namely, amending the estimation on the architectural gradients, and unifying the hyper-parameter settings in the search and re-training stages.  ...  Then, we point out that the gap is due to the inaccurate estimation of the architectural gradients, based on which we propose an amended estimation method.  ...  Appendix: Stabilizing DARTS with Amended Gradient Estimation on Architectural Parameters A.  ... 
arXiv:1910.11831v5 fatcat:mb324n7if5faram342rr7xxidy

ZARTS: On Zero-order Optimization for Neural Architecture Search [article]

Xiaoxing Wang, Wenxuan Guo, Junchi Yan, Jianlin Su, Xiaokang Yang
2022 arXiv   pre-print
Also, we search on the search space of DARTS to compare with peer methods, and our discovered architecture achieves 97.54% accuracy on CIFAR-10 and 75.7% top-1 accuracy on ImageNet, which are state-of-the-art  ...  It introduces trainable architecture parameters to represent the importance of candidate operations and proposes first/second-order approximation to estimate their gradients, making it possible to solve  ...  To verify the stability of our method, we search on S1-S4 proposed by R-DARTS and conduct convergence analysis following Amended-DARTS. Performance on S1-S4.  ... 
arXiv:2110.04743v2 fatcat:nc4bfigemjahbpmaa5ws5qmmqe

iDARTS: Improving DARTS by Node Normalization and Decorrelation Discretization [article]

Huiqun Wang, Ruijie Yang, Di Huang, Yunhong Wang
2021 arXiv   pre-print
parameters.  ...  We then propose an improved version of DARTS, namely iDARTS, to deal with the two problems. In the training phase, it introduces node normalization to maintain the norm balance.  ...  [24] amends the approximation in updating architecture parameters by gradient descent. [25] manually intervenes the architecture based on a set of pre-defined rules.  ... 
arXiv:2108.11014v1 fatcat:vgvqjpigvzdobg6fxkghew5sa4

Marthe: Scheduling the Learning Rate Via Online Hypergradients

Michele Donini, Luca Franceschi, Orchid Majumder, Massimiliano Pontil, Paolo Frasconi
2020 Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence  
We describe the structure of the gradient of a validation error w.r.t. the learning rate schedule -- the hypergradient.  ...  Based on this, we introduce MARTHE, a novel online algorithm guided by cheap approximations of the hypergradient that uses past information from the optimization trajectory to simulate future behaviour  ...  DARTS+ introduced early stopping into NAS, Fair-DARTS [Chu et al., 2019] replaced the softmax function with sigmoid function, Amended-DARTS added the hyperparameter to th gradient of DARTS.  ... 
doi:10.24963/ijcai.2020/289 dblp:conf/ijcai/WangDLZX20 fatcat:5bzkmryscjavnml545blmxncbq

BARS: Joint Search of Cell Topology and Layout for Accurate and Efficient Binary ARchitectures [article]

Tianchen Zhao, Xuefei Ning, Xiangsheng Shi, Songyi Yang, Shuang Liang, Peng Lei, Jianfei Chen, Huazhong Yang, Yu Wang
2021 arXiv   pre-print
On ImageNet, with similar resource consumption, BARS-discovered architecture achieves a 6% accuracy gain than hand-crafted binary ResNet-18 architectures and outperforms other binary architectures while  ...  On CIFAR-10, BARS achieves 1.5% higher accuracy with 2/3 binary operations and 1/10 floating-point operations comparing with existing BNN NAS studies.  ...  Other aspects for improving BNN performance focuses on improving the loss function like [36] , or amend the gradient estimation as [19, 24] .  ... 
arXiv:2011.10804v3 fatcat:67m4p5vdofev5ldb7cf56g3kqu

Advances in Electron Microscopy with Deep Learning

Jeffrey Ede
2020 Zenodo  
and automatic data clustering by t-distributed stochastic neighbour embedding; adaptive learning rate clipping to stabilize learning; generative adversarial networks for compressed sensing with spiral  ...  This doctoral thesis covers some of my advances in electron microscopy with deep learning.  ...  In addition, part of the text in section 1.2 is adapted from our earlier work with permission 201 under a Creative Commons Attribution 4.0 73 license.  ... 
doi:10.5281/zenodo.4598227 fatcat:hm2ksetmsvf37adjjefmmbakvq

Advances in Electron Microscopy with Deep Learning

Jeffrey Ede
2020 Zenodo  
and automatic data clustering by t-distributed stochastic neighbour embedding; adaptive learning rate clipping to stabilize learning; generative adversarial networks for compressed sensing with spiral  ...  This doctoral thesis covers some of my advances in electron microscopy with deep learning.  ...  In addition, part of the text in section 1.2 is adapted from our earlier work with permission 201 under a Creative Commons Attribution 4.0 73 license.  ... 
doi:10.5281/zenodo.4591029 fatcat:zn2hvfyupvdwlnvsscdgswayci

Advances in Electron Microscopy with Deep Learning

Jeffrey Ede
2020 Zenodo  
and automatic data clustering by t-distributed stochastic neighbour embedding; adaptive learning rate clipping to stabilize learning; generative adversarial networks for compressed sensing with spiral  ...  This doctoral thesis covers some of my advances in electron microscopy with deep learning.  ...  In addition, part of the text in section 1.2 is adapted from our earlier work with permission 201 under a Creative Commons Attribution 4.0 73 license.  ... 
doi:10.5281/zenodo.4399748 fatcat:63ggmnviczg6vlnqugbnrexsgy

Advances in Electron Microscopy with Deep Learning

Jeffrey Ede
2020 Zenodo  
and automatic data clustering by t-distributed stochastic neighbour embedding; adaptive learning rate clipping to stabilize learning; generative adversarial networks for compressed sensing with spiral  ...  This doctoral thesis covers some of my advances in electron microscopy with deep learning.  ...  In addition, part of the text in section 1.2 is adapted from our earlier work with permission 201 under a Creative Commons Attribution 4.0 73 license.  ... 
doi:10.5281/zenodo.4413249 fatcat:35qbhenysfhvza2roihx52afuy

Advances in Electron Microscopy with Deep Learning

Jeffrey Ede
2020 Zenodo  
and automatic data clustering by t-distributed stochastic neighbour embedding; adaptive learning rate clipping to stabilize learning; generative adversarial networks for compressed sensing with spiral  ...  This doctoral thesis covers some of my advances in electron microscopy with deep learning.  ...  In addition, part of the text in section 1.2 is adapted from our earlier work with permission 201 under a Creative Commons Attribution 4.0 73 license.  ... 
doi:10.5281/zenodo.4429792 fatcat:qs6yuapx4vdbdmwna7ix7nnwty

Advances in Electron Microscopy with Deep Learning

Jeffrey Ede
2020 Zenodo  
and automatic data clustering by t-distributed stochastic neighbour embedding; adaptive learning rate clipping to stabilize learning; generative adversarial networks for compressed sensing with spiral  ...  This doctoral thesis covers some of my advances in electron microscopy with deep learning.  ...  In addition, part of the text in section 1.2 is adapted from our earlier work with permission 201 under a Creative Commons Attribution 4.0 73 license.  ... 
doi:10.5281/zenodo.4415407 fatcat:6dejwzzpmfegnfuktrld6zgpiq

Understanding Root Biology for Enhancing Cotton Production [chapter]

Jayant H. Meshram, Sunil S. Mahajan, Dipak Nagrale, Nandini Gokte-Narkhedkar, Harish Kumbhalkar
2021 Plant Roots  
It is widely accepted that breeding efforts on aboveground traits are not sufficient to the necessary yield advantage.  ...  Cotton plants with efficient root system capture water and nutrients from soil having these features of longer tap root.  ...  With the help of modern phenotypic tools to understand root system, studies on adaptive root system architecture can be one of the breeding strategies to incorporate into modern cultivar with taking advantage  ... 
doi:10.5772/intechopen.95547 fatcat:yasazpvl7fhkxhnjxt2yiv3xka

Faulting and fault sealing in production simulation models: Brent Province, northern North Sea

S. J. Jolley, H. Dijk, J. H. Lamens, Q. J. Fisher, T. Manzocchi, H. Eikmans, Y. Huang
2007 Petroleum Geoscience  
We have systematically modelled transmissibility multipliers from the upscaled cellular structure and property grids of geometrically robust models -with reference to data on clay content and permeability  ...  experience that these fault juxtapositions impact the 'plumbing' of the faulted layering system in the reservoirs and the models that are built to mimic them -and are, in fact, a first-order sensitivity on  ...  The authors thank predecessors and office co-habitants who have worked with them on these fields, for their part in supporting this study, (in particular Simon Price, Hans de Keijzer, Tim Stevens, and  ... 
doi:10.1144/1354-079306-733 fatcat:ncvfineutnd7rd6pa5z75b3jqu

Local Model Feature Transformations [article]

CScott Brown
2020 arXiv   pre-print
We summarize this generic technique of using local models as a feature extraction step with the term "local model feature transformations."  ...  We hypothesize that, using a sufficiently complex local model family, various properties of the individual local models, such as their learned parameters, can be used as features for further learning.  ...  This suggests that the algorithm is capable of extracting meaningful parameters, but has a problem with stability of the optimal GPR parameters over time.  ... 
arXiv:2004.06149v1 fatcat:mgosv4hjabbj7esu7ezdnr3sl4

Abstracts (Continue in Part XXXIII)

1996 Proceedings of the International Society for Magnetic Resonance in Medicine  
As the gradient reversal method does not depend on phase, it is not affected by eddy currents.  ...  The gradient reversal method is less sensitive to the experiment type and works well with images obtained from half Fourier spin echo EPI experiments.  ...  filter that regularises the estimation of the image gradient at each pixel.  ... 
doi:10.1002/mrmp.22419960133 fatcat:qeadomlnojcy7fzo4lmbpjl7oq
« Previous Showing results 1 — 15 out of 74 results