# Detection of Lexical Stress Errors in Non-Native (L2) English with Data Augmentation and Attention

arXiv: https://arxiv.org/abs/2012.14788 · HTML: https://mlsanity.com/paper/2012.14788/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 Dec 2020.

## Built on by (most cited first, 9 of 9)

- [Improving Mispronunciation Detection with Wav2vec2-based Momentum Pseudo-Labeling for Accentedness and Intelligibility Assessment](https://mlsanity.com/paper/2203.15937/) (2022)
- [Automatic Pronunciation Assessment -- A Review](https://mlsanity.com/paper/2310.13974/) (2023)
- [Weakly-supervised word-level pronunciation error detection in non-native English speech](https://mlsanity.com/paper/2106.03494/) (2021)
- [Computer-assisted Pronunciation Training -- Speech synthesis is almost all you need](https://mlsanity.com/paper/2207.00774/) (2022)
- [Improving Non-native Word-level Pronunciation Scoring with Phone-level Mixup Data Augmentation and Multi-source Information](https://mlsanity.com/paper/2203.01826/) (2022)
- [Preserving Phonemic Distinctions for Ordinal Regression: A Novel Loss Function for Automatic Pronunciation Assessment](https://mlsanity.com/paper/2310.01839/) (2023)
- [How Does a Deep Neural Network Look at Lexical Stress in English Words?](https://mlsanity.com/paper/2508.07229/) (2025)
- [Light-weight Pronunciation Assessment via Discrete Speech Token Surprisal](https://mlsanity.com/paper/2606.19910/) (2026)
- [A Survey of Automated Presentation Coaching: Systems, Methods, and Open Challenges](https://mlsanity.com/paper/2606.27380/) (2026)

## Abstract (arXiv)

This paper describes two novel complementary techniques that improve the detection of lexical stress errors in non-native (L2) English speech: attention-based feature extraction and data augmentation based on Neural Text-To-Speech (TTS). In a classical approach, audio features are usually extracted from fixed regions of speech such as the syllable nucleus. We propose an attention-based deep learning model that automatically derives optimal syllable-level representation from frame-level and phoneme-level audio features. Training this model is challenging because of the limited amount of incorrect stress patterns. To solve this problem, we propose to augment the training set with incorrectly stressed words generated with Neural TTS. Combining both techniques achieves 94.8% precision and 49.2% recall for the detection of incorrectly stressed words in L2 English speech of Slavic and Baltic speakers.