Applying deep learning for identifying bioturbation from core photographs
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
ABSTRACT Advances and availability of deep learning (DL) software have recently allowed the development, testing, and deployment of automated image classification schemes for sedimentary features from core images. The development of these methods is especially relevant for extracting useful geological features from otherwise unused core photographs. This paper demonstrates and tests the use of a DL workflow for the automated extraction of bioturbation data from a core photograph data set. The proposed workflow includes extracting image tiles from core photographs along a grid and referencing each tile with collected sedimentary data. Each labeled image tile is then used as a training and testing input for a machine learning algorithm. This method allows users to quickly generate thousands of labeled training images. To demonstrate and test this workflow, a data set was collected using PyCHNO™, an open-source software specifically designed to collect sedimentary data from core photographs. The resulting data set comprising 13,545 tiles of 128 × 128 pixel resolution is used to train a DL algorithm to automatically predict if a core photograph contains evidence of bioturbation. The trained model was able to predict whether or not an image demonstrated evidence of bioturbation with up to 88% accuracy. The workflow demonstrates one of many possible applications for automatically extracting biogenic or physical sedimentary structure data from core photographs. Models built using this approach can be used to “seed” wells from a given area or interval, which can therefore significantly increase the value of core photograph data sets with relative ease.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 it