TED-LIUM 3: twice as much data and corpus repartition for experiments on speaker adaptation
arXiv:1805.04699 · doi:10.1007/978-3-319-99579-3_21
Abstract
In this paper, we present TED-LIUM release 3 corpus dedicated to speech recognition in English, that multiplies by more than two the available data to train acoustic models in comparison with TED-LIUM 2. We present the recent development on Automatic Speech Recognition (ASR) systems in comparison with the two previous releases of the TED-LIUM Corpus from 2012 and 2014. We demonstrate that, passing from 207 to 452 hours of transcribed speech training data is really more useful for end-to-end ASR systems than for HMM-based state-of-the-art ones, even if the HMM-based ASR system still outperforms end-to-end ASR system when the size of audio training data is 452 hours, with respectively a Word Error Rate (WER) of 6.6% and 13.7%. Last, we propose two repartitions of the TED-LIUM release 3 corpus: the legacy one that is the same as the one existing in release 2, and a new one, calibrated and designed to make experiments on speaker adaptation. Like the two first releases, TED-LIUM 3 corpus will be freely available for the research community.
Submitted to SPECOM 2018, 20th International Conference on Speech and Computer; TED-LIUM 3 corpus available on https://lium.univ-lemans.fr/en/ted-lium3/
References in corpus (1)
Cited by in corpus (26)
- A Comparative Study on Transformer vs RNN in Speech Applications
- Self-Supervised Speech Representation Learning: A Review
- GigaSpeech: An Evolving, Multi-domain ASR Corpus with 10,000 Hours of Transcribed Audio
- BigSSL: Exploring the Frontier of Large-Scale Semi-Supervised Learning for Automatic Speech Recognition
- SpeechStew: Simply Mix All Available Speech Recognition Data to Train One Large Neural Network
- Direct Speech-to-image Translation
- UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data
- VATLM: Visual-Audio-Text Pre-Training with Unified Masked Prediction for Speech Representation Learning
- ATCSpeech: a multilingual pilot-controller speech corpus from real Air Traffic Control environment
- Augmented Datasheets for Speech Datasets and Ethical Decision-Making
- Privacy attacks for automatic speech recognition acoustic models in a federated learning framework
- Beyond Voice Activity Detection: Hybrid Audio Segmentation for Direct Speech Translation
- IMS-Speech: A Speech to Text Tool
- UWB-NTIS Speaker Diarization System for the DIHARD II 2019 Challenge
- A Survey on Speech Large Language Models for Understanding
- Ed-Fed: A generic federated learning framework with resource-aware client selection for edge devices
- Dealing with training and test segmentation mismatch: FBK@IWSLT2021
- Toward Cross-Domain Speech Recognition with End-to-End Models
- FlanEC: Exploring Flan-T5 for Post-ASR Error Correction
- SpeechColab Leaderboard: An Open-Source Platform for Automatic Speech Recognition Evaluation
- Efficient Ensemble for Multimodal Punctuation Restoration using Time-Delay Neural Network
- Towards Maximum Likelihood Training for Transducer-based Streaming Speech Recognition
- Semi-Supervised Speech Confidence Detection using Pseudo-Labelling and Whisper Embeddings
- EasyCall corpus: a dysarthric speech dataset
- Beyond the Utterance: An Empirical Study of Very Long Context Speech Recognition
- Navigating Speech Recording Collections with AI-Generated Illustrations