| Presentation | 2010-02-19 An Easy-To-Use Recurrent Neural Network Architecture for Sequence Recognition Marcus LIWICKI, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | In this presentation the recently introduced Bidirectional Recurrent Neural Networks will be described. This novel type of recurrent neural network has been specifically designed for sequence labelling tasks where the data is hard to segment and contains long-range, bidirectional interdependencies. They allow for a direct recognition of raw pixel data. In experiments on two unconstrained handwriting databases, the new approach achieves word recognition accuracies of 79.7% on online data and 74.1% on offline data, significantly outperforming a state-of-the-art HMM-based system. Promising experimental results on various other datasets from different countries are furthermore presented. Lastly an in-depth discussion of the differences between the network and HMMs is provided, suggesting reasons for the network's superior performance. A toolkit implementing the networks is freely available for public. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | recurrent neural network / pattern recognition / handwriting recognition / sequence recognition |
| Paper # | PRMU2009-219 |
| Date of Issue |
| Conference Information | |
| Committee | PRMU |
|---|---|
| Conference Date | 2010/2/11(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 | Pattern Recognition and Media Understanding (PRMU) |
|---|---|
| Language | ENG |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | An Easy-To-Use Recurrent Neural Network Architecture for Sequence Recognition |
| Sub Title (in English) | |
| Keyword(1) | recurrent neural network |
| Keyword(2) | pattern recognition |
| Keyword(3) | handwriting recognition |
| Keyword(4) | sequence recognition |
| 1st Author's Name | Marcus LIWICKI |
| 1st Author's Affiliation | Kyushu University:Deutsches Forschungszentrum fur Kunstliche Intelligenz() |
| Date | 2010-02-19 |
| Paper # | PRMU2009-219 |
| Volume (vol) | vol.109 |
| Number (no) | 418 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |