ASR Error Correction and Domain Adaptation Using Machine Translation
| Autor: | Florian Metze, Anirudh Mani, Shruti Palaskar, Sandeep Konam, Nimshi Venkat Meripo |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
FOS: Computer and information sciences
Computer Science - Machine Learning Sound (cs.SD) Machine translation Computer science Speech recognition InformationSystems_INFORMATIONSTORAGEANDRETRIEVAL Word error rate Machine Learning (stat.ML) 02 engineering and technology 010501 environmental sciences computer.software_genre ComputingMethodologies_ARTIFICIALINTELLIGENCE 01 natural sciences Computer Science - Sound Machine Learning (cs.LG) Domain (software engineering) Statistics - Machine Learning Audio and Speech Processing (eess.AS) FOS: Electrical engineering electronic engineering information engineering 0202 electrical engineering electronic engineering information engineering 0105 earth and related environmental sciences BLEU Syntax (programming languages) Point (typography) Speaker diarisation Task (computing) 020201 artificial intelligence & image processing Error detection and correction computer Electrical Engineering and Systems Science - Audio and Speech Processing |
| Zdroj: | ICASSP |
| DOI: | 10.1109/icassp40776.2020.9053126 |
| Popis: | Off-the-shelf pre-trained Automatic Speech Recognition (ASR) systems are an increasingly viable service for companies of any size building speech-based products. While these ASR systems are trained on large amounts of data, domain mismatch is still an issue for many such parties that want to use this service as-is leading to not so optimal results for their task. We propose a simple technique to perform domain adaptation for ASR error correction via machine translation. The machine translation model is a strong candidate to learn a mapping from out-of-domain ASR errors to in-domain terms in the corresponding reference files. We use two off-the-shelf ASR systems in this work: Google ASR (commercial) and the ASPIRE model (open-source). We observe 7% absolute improvement in word error rate and 4 point absolute improvement in BLEU score in Google ASR output via our proposed method. We also evaluate ASR error correction via a downstream task of Speaker Diarization that captures speaker style, syntax, structure and semantic improvements we obtain via ASR correction. Comment: Accepted for Oral Presentation at ICASSP 2020 |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|