对比多种语音识别模型组合方法，提升识别准确率。
A Comparative Analysis on ASR System Combination for Attention, CTC, Factored Hybrid, and Transducer Models
- 两阶段重评分：先合并候选序列，再用对数线性组合打分
- 在Librispeech 960小时数据上验证，不同架构模型组合效果更优
- 统一解码流程，确保跨模型比较公平可靠
语音识别系统组合方法包括句级或词级结构化融合，以及束搜索中模型得分的合并。本文对比了主流ASR架构（注意力、CTC、因子混合、转换器）的组合效果。方法通过整合两个模型候选的联合假设列表，利用序列级得分的对数线性组合选出最优假设，发挥不同模型在搜索空间中的互补优势。为避免因解码方式差异带来的性能波动，采用两阶段方法保证所有组合结果的比较一致性。实验基于Librispeech 960小时任务，评估了不同架构、标签拓扑与单元配置的模型对组合表现。
原文摘要 · Abstract (English)
Combination approaches for speech recognition (ASR) systems cover structured sentence-level or word-based merging techniques as well as combination of model scores during beam search. In this work, we compare model combination across popular ASR architectures. Our method leverages the complementary strengths of different models in exploring diverse portions of the search space. We rescore a joint hypothesis list of two model candidates. We then identify the best hypothesis through log-linear combination of these sequence-level scores. While model combination during first-pass recognition may yield improved performance, it introduces variability due to differing decoding methods, making direct comparison more challenging. Our two-pass method ensures consistent comparisons across all system combination results presented in this study. We evaluate model pair candidates with varying architectures and label topologies and units. Experimental results are provided for the Librispeech 960h task.
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。