Machines Are Learning Chest Auscultation. Will They Also Become Our Teachers?
Bibliographic record
Abstract
Great strides in the development of machine learning techniques are bringing applications of artificial intelligence to ever more areas of clinical medicine. Their potential in the evaluation of visual images and in speech recognition is well established. Recently, the capabilities of machine hearing have been also applied to chest auscultation (ie, the automated analysis, characterization, and classification of heart and lung sounds). Comparing strengths and limitations of human vs machine hearing can help to put these developments in perspective. Humans have multisensory perception (ie, they receive visual and tactile information while auscultating). Humans also surpass machines in the ability to focus attention on listening for specific sounds in noisy environments. Together with information on a patient's history and presumed medical diagnosis, and with frequent repetition, chest auscultation remains a trainable and valuable human skill. Advantages of machine hearing of chest sounds with digital stethoscopes include not only objective acoustic analysis but also storage of data that allows comparisons over time, presentation in audiovisual format, and wireless communication. Machines can support patient management by relating acoustic analyses to clinical diagnoses, serving as decision support for further investigations, and by monitoring of patients over time. The potential of machines to become teachers of chest auscultation is only now coming into focus. In the near future, assessment of chest sounds will largely remain in the domain of traditional acoustic stethoscopes. However, machines may well be used for training students in different health care professions and nonmedical caregivers, provided that humans remain part of the process.
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 | 0.013 | 0.058 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.002 | 0.008 |
| Scholarly communication | 0.008 | 0.027 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.009 | 0.009 |
| Insufficient payload (model declined to judge) | 0.014 | 0.011 |
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".