Interactive speech recognition systems must generate words quickly while also producing accurate results. Two-pass models excel at these requirements by employing a first-pass decoder that quickly emits words, and a second-pass decoder that requires more context but is more accurate. Previous work has established that a deliberation network can be an effective second-pass model. The model attends to two kinds of inputs at once: encoded audio frames and the hypothesis text from the first-pass model. In this work, we explore using transformer layers instead of long-short term memory (LSTM) layers for deliberation rescoring. In transformer layers, we generalize the "encoder-decoder" attention to attend to both encoded audio and first-pass text hypotheses. The output context vectors are then combined by a merger layer. Compared to LSTM-based deliberation, our best transformer deliberation achieves 7% relative word error rate improvements along with a 38% reduction in computation. We also compare against non-deliberation transformer rescoring, and find a 9% relative improvement.
翻译:互动语音识别系统必须快速生成单词, 同时产生准确的结果 。 两通模式通过使用第一通解码器, 快速发布单词, 和第二通解码器, 需要更多上下文, 但更准确 。 先前的工作已经确定审议网络可以是一个有效的第二通模式 。 该模式会同时进行两种输入: 编码音频框架和第一通模式的假设文本 。 在这项工作中, 我们探索使用变压器层, 而不是长期短期内存层, 进行评分再校。 在变压层中, 我们将“ encoder- decoder” 的注意力普遍化, 以同时关注加密音频和第一通文本假体 。 然后, 输出环境矢由合并层组合在一起。 与基于 LSTM 的评断相比, 我们最好的变压器评在计算中实现了7% 相对单词错误率的改进, 并减少了 38% 。 我们还将非解放变压器再比较, 并发现9% 相对改进 。