Acoustic Modeling Using a Shallow CNN-HTSVM Architecture
This work addresses speech recognition for scenarios with limited data, though it appears incremental as it combines existing techniques like CNN and SVM in a hierarchical structure.
The paper tackled the challenge of high-accuracy speech recognition with limited data by proposing a Shallow-CNN-HTSVM architecture, which outperformed traditional GMM-HMM models and a MLP classifier, achieving robust recognition rates even with small datasets.
High-accuracy speech recognition is especially challenging when large datasets are not available. It is possible to bridge this gap with careful and knowledge-driven parsing combined with the biologically inspired CNN and the learning guarantees of the Vapnik Chervonenkis (VC) theory. This work presents a Shallow-CNN-HTSVM (Hierarchical Tree Support Vector Machine classifier) architecture which uses a predefined knowledge-based set of rules with statistical machine learning techniques. Here we show that gross errors present even in state-of-the-art systems can be avoided and that an accurate acoustic model can be built in a hierarchical fashion. The CNN-HTSVM acoustic model outperforms traditional GMM-HMM models and the HTSVM structure outperforms a MLP multi-class classifier. More importantly we isolate the performance of the acoustic model and provide results on both the frame and phoneme level considering the true robustness of the model. We show that even with a small amount of data accurate and robust recognition rates can be obtained.