2026-05-12T03:46:24Z https://ipsj.ixsq.nii.ac.jp/oai

oai:ipsj.ixsq.nii.ac.jp:00192700 2025-01-20T00:02:25Z 1164:5159:9402:9617

Feature Transfer Learning for Wav2Text Sequence-to-Sequence ASR Feature Transfer Learning for Wav2Text Sequence-to-Sequence ASR Andros, Tjandra Sakriani, Sakti Satoshi, Nakamura Andros, Tjandra Sakriani, Sakti Satoshi, Nakamura セッション1 音声認識 In this paper, we construct the first end-to-end attention-based encoder-decoder model to process directly from raw speech waveform to the text transcription. We called the model as ”Attention-basedWav2Text”. To assist the training process of the end-to-end model, we propose to utilize a feature transfer learning. Experimental results also reveal that the proposed Attention-based Wav2Text model directly with raw waveform could achieve a better result in comparison with the attentional encoder-decoder model trained on standard front-end filterbank features. In this paper, we construct the first end-to-end attention-based encoder-decoder model to process directly from raw speech waveform to the text transcription. We called the model as ”Attention-basedWav2Text”. To assist the training process of the end-to-end model, we propose to utilize a feature transfer learning. Experimental results also reveal that the proposed Attention-based Wav2Text model directly with raw waveform could achieve a better result in comparison with the attentional encoder-decoder model trained on standard front-end filterbank features. technical report 情報処理学会 2018-12-03 application/pdf 研究報告音声言語情報処理（SLP） 3 2018-SLP-125 1 2 2188-8663 AN10442647 https://ipsj.ixsq.nii.ac.jp/record/192700/files/IPSJ-SLP18125003.pdf eng