用CTC模块替代原模型中的积分放电模块，提升语音识别速度与抗噪能力。
Paraformer-v2: An improved non-autoregressive transformer for noise-robust speech recognition
- 用CTC模块提取词元嵌入，替代原有的连续积分放电机制
- 在多个数据集上性能超越原版，英语数据集错误率降低超14%
- 更适合嘈杂环境下的快速高精度语音识别任务
基于注意力的编码器-解码器模型（如Transformer及其变体）以自回归方式生成输出序列。尽管性能优越，但自回归模型计算效率低下，其生成需与输出长度等量迭代。本文提出Paraformer-v2，一种改进的非自回归语音识别模型，具备更快、更准、更强抗噪能力。Paraformer-v2采用CTC模块提取词元嵌入，替代原Paraformer中的连续积分放电模块。大量实验表明，Paraformer-v2在多个数据集上优于原版，尤其在英语数据集上词错误率（WER）改善超过14%，且在噪声环境下表现更稳健。
原文摘要 · Abstract (English)
Attention-based encoder-decoder, e.g. transformer and its variants, generates the output sequence in an autoregressive (AR) manner. Despite its superior performance, AR model is computationally inefficient as its generation requires as many iterations as the output length. In this paper, we propose Paraformer-v2, an improved version of Paraformer, for fast, accurate, and noise-robust non-autoregressive speech recognition. In Paraformer-v2, we use a CTC module to extract the token embeddings, as the alternative to the continuous integrate-and-fire module in Paraformer. Extensive experiments demonstrate that Paraformer-v2 outperforms Paraformer on multiple datasets, especially on the English datasets (over 14% improvement on WER), and is more robust in noisy environments.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。