Optimization and Evaluation of Spoken English CAF Based on Artificial Intelligence and Corpus
Bibliographic record
Abstract
English is the most widely used language in the world, and the pronunciation of its spoken language is equally important. The traditional methods are not high in complexity, accuracy and fluency (CAF) for spoken English recognition. Therefore, it is very important to use AI and corpus to optimize and evaluate spoken English CAF. This paper aims to study the optimization and evaluation of spoken English CAF using AI and corpus, and proposes to use the Hidden Markov (HMM) model and convolutional neural network (CNN) model in the field of AI to optimize and evaluate spoken English CAF. By selecting a variety of English voices from the BNC corpus for model training and testing, and selecting the complexity, accuracy, fluency and harmonic average of the CNN model recognition as evaluation indicators, the HMM model's recognition spectrogram is added up and analyzed. In the experimental test, it was found that when the number of frames is 210, the indicators of the CNN model have been greatly improved, so the number of frames selected for the test in this paper is 210. The results show that the A value obtained by the HMM model test is about 85%, the CNN model is 67%, and the traditional SVM model is only 35%. The HMM model is tested with a C value of about 60%, the CNN model is 65%, and the traditional model is only 45%. The F-value obtained from the test of the HMM model is about 83%, the CNN model is 67%, and the traditional model is 46%. In contrast, the HMM model has higher recognition accuracy for spoken English, and the recognition results are more fluent. However, the CNN model can recognize spoken English with higher complexity, and both the CNN model and the HMM model can improve the CAF optimization effect of spoken English.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".