2018 IEEE Symposium Series on Computational Intelligence (SSCI) 2018
DOI: 10.1109/ssci.2018.8628632
|View full text |Cite
|
Sign up to set email alerts
|

Structure Optimization of Neural Networks with L<inf>1</inf> Regularization on Gates

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
3
0
1

Year Published

2021
2021
2024
2024

Publication Types

Select...
4
2

Relationship

0
6

Authors

Journals

citations
Cited by 13 publications
(4 citation statements)
references
References 33 publications
0
3
0
1
Order By: Relevance
“…We developed the LiCu-NNIP using the deep potential framework proposed by Zhang et al, [23,24] which is one of the most useful methods to obtain interatomic potentials in recent years. [25,26] The scheme of the LiCu-NNIP model training is shown in Figure 1a. The total energy E of a system is a sum of atomic energies, i.e., E = ΣE i , [23,27] where i is the index of the atom.…”
Section: Resultsmentioning
confidence: 99%
“…We developed the LiCu-NNIP using the deep potential framework proposed by Zhang et al, [23,24] which is one of the most useful methods to obtain interatomic potentials in recent years. [25,26] The scheme of the LiCu-NNIP model training is shown in Figure 1a. The total energy E of a system is a sum of atomic energies, i.e., E = ΣE i , [23,27] where i is the index of the atom.…”
Section: Resultsmentioning
confidence: 99%
“…In the DeePMD method, firstly, a set of symmetry preserving descriptors is constructed by using the embedding net for each atom, whose details can be found in ref. 16. The descriptors are passed to the fitting NN to get the atomic energies.…”
Section: Methodsmentioning
confidence: 99%
“…12 In 2007, Behler and Parrinello 13 proposed the construction of descriptors of systems by using the radial distribution function and angular distribution function with multiple different hyperparameters, which was applied to high-dimensional systems. Based on the development of deep learning, the DeePMD method 14–16 greatly reducing the tedious work of artificially constructing descriptors to a certain extent has been widely used. 17–21…”
Section: Introductionmentioning
confidence: 99%
“…İlk yaklaşım minimal bir mimari ile başlayıp, nöron ekleyerek ağın kapasitesini geliştirip optimal nöron sayısını elde etmektir [15][16][17][18][19][20]. İkinci yaklaşım ise gereğinden fazla nöron bulunan bir mimari ile başlayıp, gereksiz nöronları budayarak optimum nöron sayısını elde etmektir [13,[21][22][23][24][25][26][27][28][29][30]. Bu yaklaşımlara ilave olarak her iki yaklaşımı da kullanan hibrit çalışmalar da mevcuttur [31][32][33][34].…”
Section: 𝑖=1unclassified