| Presentation | 2011-03-11 A study on Effect of Feeding Method of Training Data in Improved online quasi-Newton Training Algorithm Toshikazu ABE, Yoshihiko SAKASHITA, Hiroshi NINOMIYA, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Various techniques based on the gradient descent method have been studied as training algorithms for neural networks. Neural network training poses data-driven optimization problems in which the objective function involves the summation of loss terms over a set of data to be modeled. For a given set of training data, the gradient-based algorithms operate in one of two modes: stochastic (online) or batch. Recently, the robust training algorithm based on quasi-Newton method has been introduced improving the feeding-technique of training data. The algorithm combines the "stochastic (online)" mode with the "batch" one. In this paper the improved feeding-technique of training data is applied to the other gradient-based training algorithms. Moreover, the convergence properties of the improved feeding-technique of training data to the effect on randomsize feeding method are studied through the computer simulations. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | neural networks / learning training algorithm / online training method / batch training method |
| Paper # | NLP2010-192 |
| Date of Issue |
| Conference Information | |
| Committee | NLP |
|---|---|
| Conference Date | 2011/3/3(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Nonlinear Problems (NLP) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | A study on Effect of Feeding Method of Training Data in Improved online quasi-Newton Training Algorithm |
| Sub Title (in English) | |
| Keyword(1) | neural networks |
| Keyword(2) | learning training algorithm |
| Keyword(3) | online training method |
| Keyword(4) | batch training method |
| 1st Author's Name | Toshikazu ABE |
| 1st Author's Affiliation | Graduate school of Electrical and Information Engineering, Shonan Institute of Technology() |
| 2nd Author's Name | Yoshihiko SAKASHITA |
| 2nd Author's Affiliation | Department of Information Science, Faculty of Engineering, Shonan Institute of Technology |
| 3rd Author's Name | Hiroshi NINOMIYA |
| 3rd Author's Affiliation | Department of Information Science, Faculty of Engineering, Shonan Institute of Technology |
| Date | 2011-03-11 |
| Paper # | NLP2010-192 |
| Volume (vol) | vol.110 |
| Number (no) | 465 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |