Picture for Thai-Son Nguyen

Thai-Son Nguyen

Multi-stage Large Language Model Correction for Speech Recognition

Add code
Oct 17, 2023
Viaarxiv icon

Super-Human Performance in Online Low-latency Recognition of Conversational Speech

Add code
Oct 22, 2020
Figure 1 for Super-Human Performance in Online Low-latency Recognition of Conversational Speech
Figure 2 for Super-Human Performance in Online Low-latency Recognition of Conversational Speech
Figure 3 for Super-Human Performance in Online Low-latency Recognition of Conversational Speech
Figure 4 for Super-Human Performance in Online Low-latency Recognition of Conversational Speech
Viaarxiv icon

ELITR Non-Native Speech Translation at IWSLT 2020

Add code
Jun 05, 2020
Figure 1 for ELITR Non-Native Speech Translation at IWSLT 2020
Figure 2 for ELITR Non-Native Speech Translation at IWSLT 2020
Figure 3 for ELITR Non-Native Speech Translation at IWSLT 2020
Figure 4 for ELITR Non-Native Speech Translation at IWSLT 2020
Viaarxiv icon

Relative Positional Encoding for Speech Recognition and Direct Translation

Add code
May 20, 2020
Figure 1 for Relative Positional Encoding for Speech Recognition and Direct Translation
Figure 2 for Relative Positional Encoding for Speech Recognition and Direct Translation
Figure 3 for Relative Positional Encoding for Speech Recognition and Direct Translation
Figure 4 for Relative Positional Encoding for Speech Recognition and Direct Translation
Viaarxiv icon

High Performance Sequence-to-Sequence Model for Streaming Speech Recognition

Add code
Mar 22, 2020
Figure 1 for High Performance Sequence-to-Sequence Model for Streaming Speech Recognition
Figure 2 for High Performance Sequence-to-Sequence Model for Streaming Speech Recognition
Figure 3 for High Performance Sequence-to-Sequence Model for Streaming Speech Recognition
Figure 4 for High Performance Sequence-to-Sequence Model for Streaming Speech Recognition
Viaarxiv icon

Toward Cross-Domain Speech Recognition with End-to-End Models

Add code
Mar 09, 2020
Figure 1 for Toward Cross-Domain Speech Recognition with End-to-End Models
Figure 2 for Toward Cross-Domain Speech Recognition with End-to-End Models
Figure 3 for Toward Cross-Domain Speech Recognition with End-to-End Models
Figure 4 for Toward Cross-Domain Speech Recognition with End-to-End Models
Viaarxiv icon

Improving sequence-to-sequence speech recognition training with on-the-fly data augmentation

Add code
Oct 29, 2019
Figure 1 for Improving sequence-to-sequence speech recognition training with on-the-fly data augmentation
Figure 2 for Improving sequence-to-sequence speech recognition training with on-the-fly data augmentation
Figure 3 for Improving sequence-to-sequence speech recognition training with on-the-fly data augmentation
Figure 4 for Improving sequence-to-sequence speech recognition training with on-the-fly data augmentation
Viaarxiv icon

Very Deep Self-Attention Networks for End-to-End Speech Recognition

Add code
May 03, 2019
Figure 1 for Very Deep Self-Attention Networks for End-to-End Speech Recognition
Figure 2 for Very Deep Self-Attention Networks for End-to-End Speech Recognition
Figure 3 for Very Deep Self-Attention Networks for End-to-End Speech Recognition
Figure 4 for Very Deep Self-Attention Networks for End-to-End Speech Recognition
Viaarxiv icon

Learning Shared Encoding Representation for End-to-End Speech Recognition Models

Add code
Mar 31, 2019
Figure 1 for Learning Shared Encoding Representation for End-to-End Speech Recognition Models
Figure 2 for Learning Shared Encoding Representation for End-to-End Speech Recognition Models
Figure 3 for Learning Shared Encoding Representation for End-to-End Speech Recognition Models
Figure 4 for Learning Shared Encoding Representation for End-to-End Speech Recognition Models
Viaarxiv icon

Using multi-task learning to improve the performance of acoustic-to-word and conventional hybrid models

Add code
Feb 02, 2019
Figure 1 for Using multi-task learning to improve the performance of acoustic-to-word and conventional hybrid models
Figure 2 for Using multi-task learning to improve the performance of acoustic-to-word and conventional hybrid models
Figure 3 for Using multi-task learning to improve the performance of acoustic-to-word and conventional hybrid models
Figure 4 for Using multi-task learning to improve the performance of acoustic-to-word and conventional hybrid models
Viaarxiv icon