Skipformer: A Skip-and-Recover Strategy for Efficient Speech Recognition
This addresses computational bottlenecks in speech recognition for applications requiring efficient processing, though it is an incremental improvement on existing methods.
The paper tackles the computational inefficiency of Conformer-based speech recognition models by proposing Skipformer, a 'Skip-and-Recover' architecture that dynamically reduces input sequence length by 31 times on Aishell-1 and 22 times on Librispeech, achieving better accuracy and faster inference than baselines.
Conformer-based attention models have become the de facto backbone model for Automatic Speech Recognition tasks. A blank symbol is usually introduced to align the input and output sequences for CTC or RNN-T models. Unfortunately, the long input length overloads computational budget and memory consumption quadratically by attention mechanism. In this work, we propose a "Skip-and-Recover" Conformer architecture, named Skipformer, to squeeze sequence input length dynamically and inhomogeneously. Skipformer uses an intermediate CTC output as criteria to split frames into three groups: crucial, skipping and ignoring. The crucial group feeds into next conformer blocks and its output joint with skipping group by original temporal order as the final encoder output. Experiments show that our model reduces the input sequence length by 31 times on Aishell-1 and 22 times on Librispeech corpus. Meanwhile, the model can achieve better recognition accuracy and faster inference speed than recent baseline models. Our code is open-sourced and available online.