Lattice-Free MMI Adaptation Of Self-Supervised Pretrained Acoustic Models

12/28/2020

∙

In this work, we propose lattice-free MMI (LFMMI) for supervised adaptation of self-supervised pretrained acoustic model. We pretrain a Transformer model on thousand hours of untranscribed Librispeech data followed by supervised adaptation with LFMMI on three different datasets. Our results show that fine-tuning with LFMMI, we consistently obtain relative WER improvements of 10 and 35.3 Switchboard (300h), and 4.3 compared to the baseline trained only with supervised data.

READ FULL TEXT

Lattice-Free MMI Adaptation Of Self-Supervised Pretrained Acoustic Models

Comparing CTC and LFMMI for out-of-domain adaptation of wav2vec 2.0 acoustic model

Injecting Text and Cross-lingual Supervision in Few-shot Learning from Self-Supervised Models

Improving Hybrid CTC/Attention End-to-end Speech Recognition with Pretrained Acoustic and Language Model

Improved acoustic-to-articulatory inversion using representations from pretrained self-supervised learning models

What My Motion tells me about Your Pose: Self-Supervised Fine-Tuning of Observed Vehicle Orientation Angle

Don't stop the training: continuously-updating self-supervised algorithms best account for auditory responses in the cortex

GPT Self-Supervision for a Better Data Annotator

Lattice-Free MMI Adaptation Of Self-Supervised Pretrained Acoustic Models

Related Research

Comparing CTC and LFMMI for out-of-domain adaptation of wav2vec 2.0 acoustic model

Injecting Text and Cross-lingual Supervision in Few-shot Learning from Self-Supervised Models

Improving Hybrid CTC/Attention End-to-end Speech Recognition with Pretrained Acoustic and Language Model

Improved acoustic-to-articulatory inversion using representations from pretrained self-supervised learning models

What My Motion tells me about Your Pose: Self-Supervised Fine-Tuning of Observed Vehicle Orientation Angle

Don't stop the training: continuously-updating self-supervised algorithms best account for auditory responses in the cortex

GPT Self-Supervision for a Better Data Annotator