Shinji Watanabe
Johns Hopkins University
543 Papers
2.9K Citations
Shinji Watanabe is an academic researcher from Johns Hopkins University. The author has contributed to research in topics: Computer science & Engineering. The author has an hindex of 53, co-authored 383 publications. Previous affiliations of Shinji Watanabe include Mitsubishi Electric Research Laboratories & Mitsubishi.
Chat about Author
Papers
Improving End-to-End Single-Channel Multi-Talker Speech Recognition
TL;DR: An enhanced end-to-end monaural multi- talker ASR architecture and training strategy to recognize the overlapped speech and demonstrates that the proposed architectures can significantly improve the multi-talker mixed speech recognition.
42
Streaming Transformer Asr With Blockwise Synchronous Beam Search
Emiru Tsunoo,Yosuke Kashiwagi,Shinji Watanabe +2 more
- 19 Jan 2021
TL;DR: In this article, a blockwise synchronous beam search algorithm based on blockwise processing of encoder is proposed to perform streaming E2E Transformer ASR, where encoded feature blocks are synchronously aligned using a block boundary detection technique, where a reliability score of each predicted hypothesis is evaluated based on the end-ofsequence and repeated tokens in the hypothesis.
41
•Posted Content
ESPnet-SLU: Advancing Spoken Language Understanding through ESPnet.
Siddhant Arora,Siddharth Dalmia,Pavel Denisov,Xuankai Chang,Yushi Ueda,Yifan Peng,Yuekai Zhang,Sujay Kumar,Karthik Ganesan,Brian Yan,Ngoc Thang Vu,Alan W. Black,Shinji Watanabe +12 more
TL;DR: The ESPnet-SLU project as mentioned in this paper is a toolkit that can be used to generate reproducible results on different Spoken Language Understanding (SLU) benchmarks, such as ASR, Text to Speech (TTS), and ST.
41
•Posted Content
End-to-End Neural Speaker Diarization with Permutation-Free Objectives
TL;DR: In this paper, a single neural network was proposed to directly output speaker diarization results, and a permutation-free objective function was introduced to minimize the speaker-label permutation problem.
40
Uncertainty propagation through deep neural networks
Ahmed Hussen Abdelaziz,Shinji Watanabe,John R. Hershey,Emmanuel Vincent,Dorothea Kolossa +4 more
- 06 Sep 2015
TL;DR: In this paper, the authors study the propagation of observation uncertainties through the layers of a DNN-based acoustic model and employ approximate propagation methods, including Monte Carlo sampling, the unscented transform, and the piecewise exponential approximation of the activation function, to estimate the distribution of acoustic scores.