Auto-encoder variants for solving handwritten digits classification problem
Auto-encoders (AEs) have been proposed for solving many problems in the domain of machine learning and deep learning since the last few decades. Due to their satisfactory performance, their multiple variations have also recently appeared. First, we introduce the conventional AE model and its differe...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
The Korean Institute of Intelligent Systems
2020
|
| Subjects: | |
| Online Access: | http://eprints.uthm.edu.my/6246/1/AJ%202020%20%28254%29.pdf |
| _version_ | 1825710155236376576 |
|---|---|
| author | Aamir, Muhammad Mohd Nawi, Nazri Mahdin, Hairulnizam Naseem, Rashid Zulqarnain, Muhammad |
| author_facet | Aamir, Muhammad Mohd Nawi, Nazri Mahdin, Hairulnizam Naseem, Rashid Zulqarnain, Muhammad |
| author_sort | Aamir, Muhammad |
| collection | UTHM |
| description | Auto-encoders (AEs) have been proposed for solving many problems in the domain of machine learning and deep learning since the last few decades. Due to their satisfactory performance, their multiple variations have also recently appeared. First, we introduce the conventional AE model and its different variant for learning abstract features from data by using a contrastive divergence algorithm. Second, we present the major differences among the following three popular AE variants: sparse AE (SAE), denoising AE (DAE), and contractive AE (CAE). Third, the main contribution of this study is performing the comparative study of the aforementioned three AE variants on the basis of their mathematical modeling and experiments. All the variants of the standard AE are evaluated on the basis of the MNIST benchmark handwritten digit dataset for classification problem. The observed output reveals the benefit of using the AE model and its variants. From the experiments, it is concluded that CAE achieved better classification accuracy than those of SAE and DAE. |
| first_indexed | 2024-03-05T21:53:19Z |
| format | Article |
| id | uthm.eprints-6246 |
| institution | Universiti Tun Hussein Onn Malaysia |
| language | English |
| last_indexed | 2024-03-05T21:53:19Z |
| publishDate | 2020 |
| publisher | The Korean Institute of Intelligent Systems |
| record_format | dspace |
| spelling | uthm.eprints-62462022-01-27T06:47:00Z http://eprints.uthm.edu.my/6246/ Auto-encoder variants for solving handwritten digits classification problem Aamir, Muhammad Mohd Nawi, Nazri Mahdin, Hairulnizam Naseem, Rashid Zulqarnain, Muhammad TJ210.2-211.47 Mechanical devices and figures. Automata. Ingenious mechanisms. Robots (General) Auto-encoders (AEs) have been proposed for solving many problems in the domain of machine learning and deep learning since the last few decades. Due to their satisfactory performance, their multiple variations have also recently appeared. First, we introduce the conventional AE model and its different variant for learning abstract features from data by using a contrastive divergence algorithm. Second, we present the major differences among the following three popular AE variants: sparse AE (SAE), denoising AE (DAE), and contractive AE (CAE). Third, the main contribution of this study is performing the comparative study of the aforementioned three AE variants on the basis of their mathematical modeling and experiments. All the variants of the standard AE are evaluated on the basis of the MNIST benchmark handwritten digit dataset for classification problem. The observed output reveals the benefit of using the AE model and its variants. From the experiments, it is concluded that CAE achieved better classification accuracy than those of SAE and DAE. The Korean Institute of Intelligent Systems 2020 Article PeerReviewed text en http://eprints.uthm.edu.my/6246/1/AJ%202020%20%28254%29.pdf Aamir, Muhammad and Mohd Nawi, Nazri and Mahdin, Hairulnizam and Naseem, Rashid and Zulqarnain, Muhammad (2020) Auto-encoder variants for solving handwritten digits classification problem. International Journal of Fuzzy Logic and Intelligent Systems, 20 (1). pp. 8-16. ISSN 1598-2645 http://doi.org/10.5391/IJFIS.2020.20.1.8 |
| spellingShingle | TJ210.2-211.47 Mechanical devices and figures. Automata. Ingenious mechanisms. Robots (General) Aamir, Muhammad Mohd Nawi, Nazri Mahdin, Hairulnizam Naseem, Rashid Zulqarnain, Muhammad Auto-encoder variants for solving handwritten digits classification problem |
| title | Auto-encoder variants for solving handwritten digits classification problem |
| title_full | Auto-encoder variants for solving handwritten digits classification problem |
| title_fullStr | Auto-encoder variants for solving handwritten digits classification problem |
| title_full_unstemmed | Auto-encoder variants for solving handwritten digits classification problem |
| title_short | Auto-encoder variants for solving handwritten digits classification problem |
| title_sort | auto encoder variants for solving handwritten digits classification problem |
| topic | TJ210.2-211.47 Mechanical devices and figures. Automata. Ingenious mechanisms. Robots (General) |
| url | http://eprints.uthm.edu.my/6246/1/AJ%202020%20%28254%29.pdf |
| work_keys_str_mv | AT aamirmuhammad autoencodervariantsforsolvinghandwrittendigitsclassificationproblem AT mohdnawinazri autoencodervariantsforsolvinghandwrittendigitsclassificationproblem AT mahdinhairulnizam autoencodervariantsforsolvinghandwrittendigitsclassificationproblem AT naseemrashid autoencodervariantsforsolvinghandwrittendigitsclassificationproblem AT zulqarnainmuhammad autoencodervariantsforsolvinghandwrittendigitsclassificationproblem |