Leveraging small datasets for ethical and responsible AI music making
Bibliographic record
Abstract
The impact of Artificial Intelligence is felt on every stage of contemporary musicking and is shaping our interaction with sound. Deep learning Generative AI (GenAI) systems for high-quality music generation rely on extremely large musical datasets for training. As a result, AI models tend to be trained on dominant mainstream musical genres, such as Western classical music, where large datasets are more readily available. In addition, the reliance on extremely powerful computing resources for deep learning creates barriers to use and negatively impacts our environment. This paper reports on contemporary concerns and interests of musicians, researchers, and music industry stakeholders in the responsible use of GenAI models for music and audio. Through analysis of focus group discussions and exemplar case studies of the use of GenAI in music making at a hybrid workshop of 148 participants, we offer insights into current discourses about the use of GenAI beyond dominant musical styles and suggest ways forward to increase creative agency in music making beyond the mainstream. Our findings highlight the value of small datasets of music for GenAI, the suitability of AI models for working with small datasets of music, and pose questions around what constitutes a ‘small’ dataset of music.
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 | 0.127 | 0.245 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.009 | 0.019 |
| Scholarly communication | 0.018 | 0.027 |
| Open science | 0.005 | 0.023 |
| Research integrity | 0.004 | 0.007 |
| Insufficient payload (model declined to judge) | 0.007 | 0.002 |
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".