2020/05/31 by Albert Zeyer, André Merboldt, Ralf Schlüter +1 · 1 citation
Computer Science · Engineering · Mathematics · #Acoustics #Algorithm #Artificial intelligence #Artificial neural network #Computer science #Cover (algebra) #Electrical engineering #Engineering #Natural Language Processing Techniques #Physics #Pipeline (software) #Speech Recognition and Synthesis #Topic Modeling #Topology (electrical circuits) #Transducer #cs.LG #cs.NE #eess.AS #stat.ML
paper · pdf · doi:10.21437/interspeech.2020-1855
published at Interspeech 2020
openalex publication_date 2020/10/25 · arxiv created 2020/11/18 · arxiv updated 2020/11/20 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05
The RNN transducer is a promising end-to-end model candidate. We compare the original training criterion with the full marginalization over all alignments, to the commonly used maximum approximation, which simplifies, improves and speeds up our training. We also generalize from the original neural network model and study more powerful models, made possible due to the maximum approximation. We further generalize the output label topology to cover RNN-T, RNA and CTC. We perform several studies among all these aspects, including a study on the effect of external alignments. We find that the transducer model generalizes much better on longer sequences than the attention model. Our final transducer model outperforms our attention model on Switchboard 300h by over 6% relative WER.