Presentation | 2023-03-01 Modification and Comparison of Learning Sudoku with Convolutional Networks Koichiro Ishii, Hiroshi Tamura, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Sudoku is a logic puzzle in which the number between 1 and 9 is to fill in the boxes of the 9x9 grid by using the numbers already filled in as clues. Recently, graph convolutional networks (GCNs) have been developing as a deep learning method for graph structures. In this paper, we attempt to solve a Sudoku puzzle applied to a graph structure by using GCN. Although there have already been studies using convolutional neural networks (CNN), it is difficult to understand the computational process. On the other hand, GCN model can refer to the clues with which you solve a Sudoku puzzle, so we thought that the learning and computation processes can be made closer to the human solving process. We also believe that GCN has the flexibility to adapt to Sudoku and other puzzles having different rules by applying them to graph structures. In this paper, we compared the results with those of conventional neural networks and verified the difference in accuracy of the network model by changing the feature values and internal calculation parameter. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Sudoku / Graph Theory / Machine Learning / Graph Convolutional Network |
Paper # | CAS2022-96,CS2022-73 |
Date of Issue | 2023-02-22 (CAS, CS) |
Conference Information | |
Committee | CAS / CS |
---|---|
Conference Date | 2023/3/1(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | Network processor, Signal processing and circuits for communications, Wireless LAN / PAN, etc. |
Chair | Yoshinobu Maeda(Niigata Univ.) / Daisuke Umehara(Kyoto Inst. of Tech.) |
Vice Chair | Yasutoshi Aibara(OmniVision) / Seiji Kozaki(Mitsubishi Electric) |
Secretary | Yasutoshi Aibara(NIT, Toyama college) / Seiji Kozaki(Renesas Electronics) |
Assistant | Takahide Sato(Univ. of Yamanashi) / Motoi Yamaguchi(TECHNOPRO) / Shinji Shimoda(Sony Semiconductor Solutions) / Shunsuke Koshita(Hachinohe Inst. of Tech.) / Hikaru Kawasaki(NICT) / Yuta Ida(Yamaguchi Univ.) |
Paper Information | |
Registration To | Technical Committee on Circuits and Systems / Technical Committee on Communication Systems |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Modification and Comparison of Learning Sudoku with Convolutional Networks |
Sub Title (in English) | |
Keyword(1) | Sudoku |
Keyword(2) | Graph Theory |
Keyword(3) | Machine Learning |
Keyword(4) | Graph Convolutional Network |
1st Author's Name | Koichiro Ishii |
1st Author's Affiliation | Chuo University(Chuo Univ.) |
2nd Author's Name | Hiroshi Tamura |
2nd Author's Affiliation | Chuo University(Chuo Univ.) |
Date | 2023-03-01 |
Paper # | CAS2022-96,CS2022-73 |
Volume (vol) | vol.122 |
Number (no) | CAS-396,CS-397 |
Page | pp.pp.1-5(CAS), pp.1-5(CS), |
#Pages | 5 |
Date of Issue | 2023-02-22 (CAS, CS) |