End 2 End ASR

End-to-end (E2E) automatic speech recognition (ASR) aims to directly transcribe speech to text without intermediate steps like phoneme recognition, improving efficiency and accuracy. Current research focuses on enhancing E2E ASR's robustness to domain-specific vocabulary (e.g., named entities, technical jargon) through techniques like incorporating contextual information (e.g., descriptions, dialog acts) and leveraging pre-trained language models. These advancements are significant because they improve the accuracy and reliability of speech-to-text systems across diverse applications, from voice assistants to medical transcription.

Papers

July 25, 2024

Improving Domain-Specific ASR with LLM-Generated Contextual Descriptions
Jiwon Suh, Injae Na, Woohwan Jung
Automatic Speech Recognition Speech Recognition Context Information End 2 End ASR

March 26, 2024

DANCER: Entity Description Augmented Named Entity Corrector for Automatic Speech Recognition
Yi-Cheng Wang, Hsin-Wei Wang, Bi-Cheng Yan, Chi-Han Lin, Berlin Chen
Automatic Speech Recognition Named Entity Recognition High Quality Dance Edit Distance Entity Level Factual End 2 End ASR

May 29, 2023

Retraining-free Customized ASR for Enharmonic Words Based on a Named-Entity-Aware Model and Phoneme Similarity Estimation
Yui Sudo, Kazuya Hata, Kazuhiro Nakadai
Speech Recognition End 2 End ASR

May 21, 2023

CASA-ASR: Context-Aware Speaker-Attributed ASR
Mohan Shi, Zhihao Du, Qian Chen, Fan Yu, Yangze Li, Shiliang Zhang, Jie Zhang, Li-Rong Dai
ASR Model End 2 End ASR Speech Recognition Error Speaker Attributed Automatic Speech Recognition Contextual Asr

March 31, 2023

Dialog act guided contextual adapter for personalized speech recognition
Feng-Ju Chang, Thejaswi Muniyappa, Kanthashree Mysore Sathyendra, Kai Wei, Grant P. Strimel, Ross McGowan
Speech Recognition Multi Turn Dialogue Google Speech Command Contextual Adapter End 2 End ASR

January 17, 2023

Two Stage Contextual Word Filtering for Context bias in Unified Streaming and Non-streaming Transducer
Zhanheng Yang, Sining Sun, Xiong Wang, Yike Zhang, Long Ma, Lei Xie
Contextual Information Light Stage End 2 End ASR Non Streaming Context Bias Contextual Asr

November 2, 2022

BECTRA: Transducer-based End-to-End ASR with BERT-Enhanced Encoder
Yosuke Higuchi, Tetsuji Ogawa, Tetsunori Kobayashi, Shinji Watanabe
End to End End 2 End ASR

October 21, 2022

Audio-to-Intent Using Acoustic-Textual Subword Representations from End-to-End ASR
Pranay Dighe, Prateeth Nayak, Oggi Rudovic, Erik Marchi, Xiaochuan Niu, Ahmed Tewfik
End to End Acoustic Word Embeddings Automatic Speech Recognition Hypothesis End 2 End ASR Zero Shot Intent Classification

June 29, 2022

Contextual Density Ratio for Language Model Biasing of Sequence to Sequence ASR Systems
Jesús Andrés-Ferrer, Dario Albesano, Puming Zhan, Paul Vozila
Language Model Sequence of Sequence End 2 End Density Ratio End 2 End ASR

May 26, 2022

Contextual Adapters for Personalized Speech Recognition in Neural Transducers
Kanthashree Mysore Sathyendra, Thejaswi Muniyappa, Feng-Ju Chang, Jing Liu, Jinru Su, Grant P. Strimel, Athanasios Mouchtaris, Siegfried Kunzmann
Speech Recognition End to End ASR Model Neural Transducer Rare Word Contextual Adapter End 2 End ASR Context Neural Network

April 21, 2022

Layer-wise Fast Adaptation for End-to-End Multi-Accent Speech Recognition
Xun Gong, Yizhou Lu, Zhikai Zhou, Yanmin Qian
Speech Recognition End to End Accented Speech End 2 End ASR Accent Adaptation Adaptation Layer

February 12, 2022

USTED: Improving ASR with a Unified Speech and Text Encoder-Decoder
Bolaji Yusuf, Ankur Gandhe, Alex Sokolov
Automatic Speech Recognition End to End ASR Model End 2 End ASR