提出新方法，让语音识别在低延迟下更准地输出中间结果。
X2Streaming-ASR: wait when uncertain, emit when ready for streaming ASR
- 分离决策时机与输出内容，动态决定何时提交识别结果。
- 平均提交延迟仅27-84毫秒，远低于基线的409-585毫秒。
- 适合实时语音助手、全双工对话等对延迟敏感场景。
实时语音识别系统需在低提交延迟下提供准确的中间转录文本。现有方法通常采用固定分块、前瞻或目标延迟，或在声学边界附近鼓励输出，但未直接优化单次遍历中每个输出位置应使用多少上下文。本文提出X2Streaming-ASR，将流式识别分解为‘何时提交’和‘提交什么’两个阶段。三阶段训练先建立流式识别能力，再用自动探测轨迹预热提交策略，最后通过字符级、段落级相对奖励优化精度与延迟。在AISHELL-1/2/3和WenetSpeech数据集上，该模型平均字符级提交延迟为27-84毫秒（相对于强制对齐字符端点），显著优于基线的409-585毫秒。在AISHELL-1和AISHELL-3上达到最优流式词错误率，且延迟大幅降低。
原文摘要 · Abstract (English)
Streaming automatic speech recognition (ASR) for real-time voice agents and full-duplex dialogue must provide accurate partial transcripts with low commit latency. Existing systems commonly use a fixed chunk size, look-ahead, or target delay, or encourage emissions near estimated acoustic boundaries. These approaches do not directly optimize how much additional context to use at each output position under a single-pass, hard-commit constraint. We propose X2Streaming-ASR, which decomposes streaming recognition into when to commit and what to commit. Its three-stage training procedure first establishes streaming recognition ability, then warm-starts the commit policy with automatically probed trajectories, and finally refines the policy using character-level, segment-assigned group-relative rewards for recognition accuracy and latency. Across AISHELL-1/2/3 and WenetSpeech, X2Streaming-ASR achieves a mean character-level commit latency of 27-84 ms relative to forced-aligned character endpoints, compared with 409-585 ms for the evaluated streaming baselines. It achieves the best streaming CER among the evaluated systems on AISHELL-1 and AISHELL-3 with substantially lower latency.
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。