Novel software for deep-learning based acoustic data analysis
Bibliographic record
Abstract
In recent years, deep neural networks have been successfully applied to solve a range of detection and classification tasks in underwater acoustics, outperforming existing methods. However, deep learning models are “data hungry” requiring large amounts of accurately labelled acoustic samples to train. Moreover, a certain amount of “fine tuning” is often required to achieve satisfactory performance in a new acoustic environment. Thus, the development of deep learning models depends on the input of expert human analysts, both for building the initial training set and for adjusting the model's performance. However, open-source software to facilitate this collaboration between machine learning developers and acousticians is currently lacking. To address this need, our team is building a web-based application for collaboratively annotating sound samples and validating model predictions. The user interface is designed to be familiar to acousticians, while machine learning developers have access to a dashboard allowing them to efficiently leverage the acousticians' expert knowledge. In this contribution, an overview of the application will be given and its functionalities will bedemonstrated through its application to the HALLO (Humans and ALgorithms Listening for Orcas) project. Future developments will also be described, highlighting complementary applications under development such as a model adaptation tool.
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 | 0.003 | 0.007 |
| Meta-epidemiology (narrow) | 0.003 | 0.002 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.004 | 0.004 |
| Research integrity | 0.002 | 0.004 |
| Insufficient payload (model declined to judge) | 0.051 | 0.026 |
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".