| Presentation | 2018-11-05 [Poster Presentation] Parameter Density Inheritance Using Kernel Density Estimation for Efficient CNN Learning Keisuke Horiuchi, Keisuke Kameyama, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | CNNs have shown state-of-the-art performances on a large variety of classification tasks. However, because of its complexity, they may require an enormous amount of time for training and still not converge. One of the key ideas to address these problems is to initialize the network parameters with appropriate values before training. Several known methods pro- posed the initialization using an existing trained network, but they suppose both the trained network and the new network have the same structure. In this paper, we propose an initialization method that utilizes existing trained network and can be applied to a network with a different structure based on parameter density inheritance strategy. In the experiments, we verified the effectiveness of the proposed initialization method. With the proposed method, initialization can be applied for arbitrary layers of CNN, but we initialize only the first layer which requires the longest time in training with backpropagation. Through the experiments, the effectiveness of the proposed initialization method in improving the efficiency of CNN learning was verified. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | CNN / Transfer Learning / Kernel Density Estimation |
| Paper # | IBISML2018-45 |
| Date of Issue | 2018-10-29 (IBISML) |
| Conference Information | |
| Committee | IBISML |
|---|---|
| Conference Date | 2018/11/5(3days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Hokkaido Citizens Activites Center (Kaderu 2.7) |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Information-Based Induction Science Workshop (IBIS2018) |
| Chair | Hisashi Kashima(Kyoto Univ.) |
| Vice Chair | Masashi Sugiyama(Univ. of Tokyo) / Koji Tsuda(Univ. of Tokyo) |
| Secretary | Masashi Sugiyama(Nagoya Inst. of Tech.) / Koji Tsuda(AIST) |
| Assistant | Tomoharu Iwata(NTT) / Shigeyuki Oba(Kyoto Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Infomation-Based Induction Sciences and Machine Learning |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | [Poster Presentation] Parameter Density Inheritance Using Kernel Density Estimation for Efficient CNN Learning |
| Sub Title (in English) | |
| Keyword(1) | CNN |
| Keyword(2) | Transfer Learning |
| Keyword(3) | Kernel Density Estimation |
| 1st Author's Name | Keisuke Horiuchi |
| 1st Author's Affiliation | Tsukuba University(Tsukuba Univ.) |
| 2nd Author's Name | Keisuke Kameyama |
| 2nd Author's Affiliation | Tsukuba University(Tsukuba Univ.) |
| Date | 2018-11-05 |
| Paper # | IBISML2018-45 |
| Volume (vol) | vol.118 |
| Number (no) | IBISML-284 |
| Page | pp.pp.9-15(IBISML), |
| #Pages | 7 |
| Date of Issue | 2018-10-29 (IBISML) |