%0 Journal Article %1 martinez2018dnnbased %A Martinez, Angel Mario Castro %A Gerlach, Lukas %A Payá Vayá, Guillermo %A Hermansky, Hynek %A Ooster, Jasper %A Meyer, Bernd T. %D 2018 %J Speech Communication %K myown %P 44 - 56 %R https://doi.org/10.1016/j.specom.2018.11.006 %T DNN-based performance measures for predicting error rates in automatic speech recognition and optimizing hearing aid parameters %U http://www.sciencedirect.com/science/article/pii/S0167639317303850 %V 106 %X In several applications of machine listening, predicting how well an automatic speech recognition system will perform before the actual decoding enables the system to adapt to unseen acoustic characteristics dynamically. Feedback about speech quality, for instance, could allow modern hearing aids to select a speech source in complex acoustic scenes with the aim of enhancing the speech intelligibility of a target speaker. In this study, we look at different performance measures to estimate the word error rates of simulated behind-the-ear hearing aid signals and detect the azimuth angle of the target source in 180-degree spatial scenes. These measures derive from phoneme posterior probabilities produced by a deep neural network acoustic model. However, the more complex the model is, the more computationally expensive it becomes to obtain these measures; therefore, we assess how the model size affects prediction performance. Our findings suggest measures derived from smaller nets are suitable to predict error rates of more complex models reliably enough to be implemented in hearing aid hardware.

@article{martinez2018dnnbased, abstract = {In several applications of machine listening, predicting how well an automatic speech recognition system will perform before the actual decoding enables the system to adapt to unseen acoustic characteristics dynamically. Feedback about speech quality, for instance, could allow modern hearing aids to select a speech source in complex acoustic scenes with the aim of enhancing the speech intelligibility of a target speaker. In this study, we look at different performance measures to estimate the word error rates of simulated behind-the-ear hearing aid signals and detect the azimuth angle of the target source in 180-degree spatial scenes. These measures derive from phoneme posterior probabilities produced by a deep neural network acoustic model. However, the more complex the model is, the more computationally expensive it becomes to obtain these measures; therefore, we assess how the model size affects prediction performance. Our findings suggest measures derived from smaller nets are suitable to predict error rates of more complex models reliably enough to be implemented in hearing aid hardware.}, added-at = {2022-02-23T09:36:58.000+0100}, author = {Martinez, Angel Mario Castro and Gerlach, Lukas and Pay{\'a} Vay{\'a}, Guillermo and Hermansky, Hynek and Ooster, Jasper and Meyer, Bernd T.}, biburl = {https://www.bibsonomy.org/bibtex/22c70d1cb65b822e43546a9b51fdcff32/guipava}, description = {DNN-based performance measures for predicting error rates in automatic speech recognition and optimizing hearing aid parameters - ScienceDirect}, doi = {https://doi.org/10.1016/j.specom.2018.11.006}, interhash = {01c57009810c976edf7588c9a3553102}, intrahash = {2c70d1cb65b822e43546a9b51fdcff32}, issn = {0167-6393}, journal = {Speech Communication}, keywords = {myown}, pages = {44 - 56}, timestamp = {2022-02-23T09:36:58.000+0100}, title = {DNN-based performance measures for predicting error rates in automatic speech recognition and optimizing hearing aid parameters}, url = {http://www.sciencedirect.com/science/article/pii/S0167639317303850}, volume = 106, year = 2018 }