DeLLA: Deep-Learning-based speech recognition with limited training material
At a glance
- Project leader : Prof. Dr. Hans-Peter Hutter
- Project team : Andreas Ahlenstorf, Matthias Büchi, Dr. Thilo Stadelmann
- Project budget : CHF 210'000
- Project status : completed
- Funding partner : CTI
- Project partner : SlowSoft GmbH
- Contact person : Hans-Peter Hutter
Description
Speech recognition systems baed on Deep Neural Networks (DNN) currently brake all records and is being applied already in different products. These systems normally are trained with thousands of hours of training material for applications and languages where these amounts of data are available. In this feasibility study a new training approach for DNNs is developed and evaluated that need significantly less training data.