Analyzing Hidden Representations in End-to-End Automatic Speech Recognition Systems
This work provides insights into interpretability for researchers in speech recognition, but it is incremental as it focuses on analyzing existing models rather than introducing new methods.
The authors analyzed the hidden representations learned by an end-to-end automatic speech recognition system based on convolutional and recurrent layers with CTC loss, evaluating how different layers predict phone labels to understand model design aspects.
Neural models have become ubiquitous in automatic speech recognition systems. While neural networks are typically used as acoustic models in more complex systems, recent studies have explored end-to-end speech recognition systems based on neural networks, which can be trained to directly predict text from input acoustic features. Although such systems are conceptually elegant and simpler than traditional systems, it is less obvious how to interpret the trained models. In this work, we analyze the speech representations learned by a deep end-to-end model that is based on convolutional and recurrent layers, and trained with a connectionist temporal classification (CTC) loss. We use a pre-trained model to generate frame-level features which are given to a classifier that is trained on frame classification into phones. We evaluate representations from different layers of the deep model and compare their quality for predicting phone labels. Our experiments shed light on important aspects of the end-to-end model such as layer depth, model complexity, and other design choices.