It’s time to collaborate: What human linguists can learn from machine linguists
Bibliographic record
Abstract
For decades, one of the primary goals of machine learning has been to emulate human performance by learning from human behavioural data. However, machine learning has now matured to a point where humans are learning from machines—for example, consider playstyles of Go (Silver, D., et al., "Mastering the game of go without human knowledge," Nature 550.7676(2017):354). Drawing on this idea, this project considers what speech scientists might be able to learn from considering how machines analyze the speech signal. One goal in phonetics is to break down the speech signal into units such as phonemes, syllables, and tones. Machines are also able to break down the speech signal in either a supervised or unsupervised manner. In supervised learning, the machine is trained to classify phonemes, tones, etc., using previously known labels. In unsupervised learning, the machine learns a lower-dimensional, latent representation of the speech signal and then identifies meaningful clusters in the latent space. This project takes an unsupervised approach to lexical tone identification in Mandarin. Specifically, an adversarial autoencoder and density-based clustering are used to identify the tones of Mandarin. Results contrast Mandarin lexical tones as determined by linguists and by machine learning.
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 | 0.024 | 0.074 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.012 | 0.053 |
| Scholarly communication | 0.019 | 0.060 |
| Open science | 0.003 | 0.015 |
| Research integrity | 0.009 | 0.011 |
| Insufficient payload (model declined to judge) | 0.011 | 0.005 |
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".