Summary of Reading Miscue Detection in Primary School Through Automatic Speech Recognition, by Lingyun Gao et al.
Reading Miscue Detection in Primary School through Automatic Speech Recognition
by Lingyun Gao, Cristian Tejedor-Garcia, Helmer Strik, Catia Cucchiarini
First submitted to arxiv on: 11 Jun 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS); Signal Processing (eess.SP)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This study explores the application of Automatic Speech Recognition (ASR) models for detecting reading miscues in Dutch native children’s speech. The researchers investigate the efficiency of state-of-the-art (SOTA) pre-trained ASR models, specifically Hubert Large and Whisper, in recognizing phonemes and words. They found that Hubert Large achieved a SOTA phoneme-level child speech recognition rate of 23.1%, while Whisper demonstrated a SOTA word-level performance with a Word Error Rate (WER) of 9.8%. The study suggests that Wav2Vec2 Large and Whisper are the top-performing ASR models for reading miscue detection, with Wav2Vec2 Large exhibiting high recall (0.83) and Whisper showing high precision (0.52) and an F1 score of 0.52. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study looks at how well computers can recognize the sounds and words spoken by children learning to read in Dutch. The researchers compared different computer models to see which one was best at detecting when a child makes a mistake while reading aloud. They found that some models were much better than others, with two models standing out as being very good at recognizing what kids are saying. This could help teachers quickly and easily grade children’s reading exercises, and also provide feedback to kids so they can practice more effectively. |
Keywords
» Artificial intelligence » F1 score » Precision » Recall
