Gated Recurrent Context: Softmax-free Attention for Online Encoder-Decoder Speech Recognition
This work addresses the latency and tuning challenges in online speech recognition for better user experience, but it is incremental as it builds on existing attention-based encoder-decoder models.
The paper tackled the problem of hyperparameter tuning in online attention for speech recognition by proposing a softmax-free attention method that eliminates the need for additional training hyperparameters, achieving competitive word-error-rates (WERs) with adjustable latency at test time.
Recently, attention-based encoder-decoder (AED) models have shown state-of-the-art performance in automatic speech recognition (ASR). As the original AED models with global attentions are not capable of online inference, various online attention schemes have been developed to reduce ASR latency for better user experience. However, a common limitation of the conventional softmax-based online attention approaches is that they introduce an additional hyperparameter related to the length of the attention window, requiring multiple trials of model training for tuning the hyperparameter. In order to deal with this problem, we propose a novel softmax-free attention method and its modified formulation for online attention, which does not need any additional hyperparameter at the training phase. Through a number of ASR experiments, we demonstrate the tradeoff between the latency and performance of the proposed online attention technique can be controlled by merely adjusting a threshold at the test phase. Furthermore, the proposed methods showed competitive performance to the conventional global and online attentions in terms of word-error-rates (WERs).