Automatic Data Augmentation for Domain Adapted Fine-Tuning of Self-Supervised Speech Representations
Self-Supervised Learning (SSL) has allowed leveraging large amounts of unlabeled speech data to improve the performance of speech recognition models even with small annotated datasets. Despite this, speech SSL representations may fail while facing an acoustic mismatch between the pretraining and target datasets. To address this issue, we propose a novel supervised domain adaptation method, designed for cases exhibiting such a mismatch in acoustic domains. It consists in applying properly calibrated data augmentations on a large clean dataset, bringing it closer to the target domain, and using it as part of an initial fine-tuning stage. Augmentations are automatically selected through the minimization of a conditional-dependence estimator, based on the target dataset. The approach is validated during an oracle experiment with controlled distortions and on two amateur-collected low-resource domains, reaching better performances compared to the baselines in both cases.
Code (0)
등록된 구현이 없습니다.
Tasks
Data AugmentationDomain AdaptationSelf-Supervised Learningspeech-recognitionSpeech RecognitionMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
PAGANDA: An Adaptive Task-Independent Automatic Data Augmentation
Data augmentation is a useful technique to enlarge the size of the training set and prevent overfitting for different machine learning tasks when training data is scarce. However, current data augmentation techniques rel…
BIG-bench Machine LearningData Augmentationimage-classificationImage Classification+1Adaptive Data Augmentation with Deep Parallel Generative Models
Data augmentation(DA) is a useful technique to enlarge the size of the training set and prevent overfitting for different machine learning tasks when training data is scarce. However, current data augmentation techniques…
BIG-bench Machine LearningData Augmentationimage-classificationImage Classification+1TADPOLE: Task ADapted Pre-Training via AnOmaLy DEtection
The paradigm of pre-training followed by finetuning has become a standard procedure for NLP tasks, with a known problem of domain shift between the pre-training and downstream corpus. Previous works have tried to mitigat…
Anomaly DetectionData AugmentationDomain AdaptationAugmentation-Adapted Retriever Improves Generalization of Language Models as Generic Plug-In
Retrieval augmentation can aid language models (LMs) in knowledge-intensive tasks by supplying them with external information. Prior works on retrieval augmentation usually jointly fine-tune the retriever and the LM, mak…
MMLURetrievalZero-shot GeneralizationDomain Adaptation of Document-Level NMT in IWSLT19
We describe our four NMT systems submitted to the IWSLT19 shared task in English→Czech text-to-text translation of TED talks. The goal of this study is to understand the interactions between document-level NMT and domain…
Domain AdaptationNMTSentenceTranslation