Transfer Learning from Pre-trained Language Models Improves End-to-End Speech Summarization
This work addresses the challenge of producing natural summaries from speech for applications like transcription services, but it is incremental as it builds on existing transfer learning methods.
The authors tackled the problem of training data scarcity and unnatural sentence generation in end-to-end speech summarization by integrating a pre-trained language model into the decoder and transferring the baseline encoder, resulting in improved performance over baseline and data-augmented models.
End-to-end speech summarization (E2E SSum) directly summarizes input speech into easy-to-read short sentences with a single model. This approach is promising because it, in contrast to the conventional cascade approach, can utilize full acoustical information and mitigate to the propagation of transcription errors. However, due to the high cost of collecting speech-summary pairs, an E2E SSum model tends to suffer from training data scarcity and output unnatural sentences. To overcome this drawback, we propose for the first time to integrate a pre-trained language model (LM), which is highly capable of generating natural sentences, into the E2E SSum decoder via transfer learning. In addition, to reduce the gap between the independently pre-trained encoder and decoder, we also propose to transfer the baseline E2E SSum encoder instead of the commonly used automatic speech recognition encoder. Experimental results show that the proposed model outperforms baseline and data augmented models.