AIPNet: Generative Adversarial Pre-training of Accent-invariant Networks for End-to-end Speech Recognition
This addresses the challenge of building robust, unified speech recognition systems that generalize across accents, which is an incremental improvement over existing methods.
The paper tackled the problem of accent variability in speech recognition by proposing AIPNet, a GAN-based pre-training framework for accent-invariant representation learning, which achieved relative WER reductions of 2.3-4.5% across 9 English accents when transcriptions were available in all accents and 1.6-6.1% when only US accent transcriptions were available.
As one of the major sources in speech variability, accents have posed a grand challenge to the robustness of speech recognition systems. In this paper, our goal is to build a unified end-to-end speech recognition system that generalizes well across accents. For this purpose, we propose a novel pre-training framework AIPNet based on generative adversarial nets (GAN) for accent-invariant representation learning: Accent Invariant Pre-training Networks. We pre-train AIPNet to disentangle accent-invariant and accent-specific characteristics from acoustic features through adversarial training on accented data for which transcriptions are not necessarily available. We further fine-tune AIPNet by connecting the accent-invariant module with an attention-based encoder-decoder model for multi-accent speech recognition. In the experiments, our approach is compared against four baselines including both accent-dependent and accent-independent models. Experimental results on 9 English accents show that the proposed approach outperforms all the baselines by 2.3 \sim 4.5% relative reduction on average WER when transcriptions are available in all accents and by 1.6 \sim 6.1% relative reduction when transcriptions are only available in US accent.