AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

A Toolbox for Construction and Analysis of Speech Datasets

Evelina Bakhturina, Vitaly Lavrukhin, Boris Ginsburg

NeurIPS 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add RuLS data to MCV-Ru for fine-tuning the Russian ASR modelQuartzNet-15x5 (Russian) · MCV Russian devgreedy WER (%): 12.63 -> 9.65Improved
3-fold cross-validation re-alignment with re-tuned ASR model (ablation)CTC-Segmentation on RuLS · RuLS booksaverage book WER (%): 45.8 -> 18.83Improved
filter MCV English utterances by CER threshold of 25%MCV English 7.0 validated · MCV Englishduration (hours): 1 970 -> 1 938Mixed
filter MLS English train by CER threshold of 10%MLS English train · MLS Englishduration (hours): 44659 -> 42967Mixed
text normalization of MCV English (remove punctuation, drop foreign/special characters)MCV English 7.0 validated · MCV Englishalphabet size: 149 -> 28Improved
re-segment LibriSpeech dev-clean with CTC-Segmentation instead of original segmentationLibriSpeech dev-clean · LibriSpeech dev-cleanWER (%): 3.78 -> 3.82No clear effect
re-segment LibriSpeech dev-clean with CTC-Segmentation instead of original segmentationLibriSpeech dev-clean · LibriSpeech dev-cleanCER (%): 1.14 -> 1.16No clear effect
split sentences in the middle to shorten utterances (ablation)RuLS text pre-processing · RuLS segmentationsegmentation qualityHurt
run aligner multiple times with window sizes 8000, 10000, 12000 and keep matched segments (ablation)NeMo CTC-Segmentation · RuLS segmentationalignment errorsImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy