Pre-trained Models for SMP Classification and Segmentation
Bibliographic record
Abstract
This dataset provides access to pre-trained models that were used for SnowMicroPen profile classification and segmentation. The models were trained on a part of the MOSAiC SMP dataset, available on https://doi.pangaea.de/10.1594/PANGAEA.935554. The labeled training data consists mostly of profiles from leg three of the expedition (January - May 2020), some profiles from leg one and two, and no profiles from leg four. Please refer to the snowdragon GitHub repository (https://github.com/liellnima/snowdragon) to access the models' training code and be directed to current publications. The following trained models are available here (alphabetically ordered): Artificial neural networks Bi-directional long short-term memory (blstm.hdf5) Encoder-decoder (enc_dec.hdf5) Long short-term memory (lstm.hdf5) Baseline Majority vote classifier (baseline.model) Semi-supervised models Cluster-then-predict models: Bayesian Gaussian mixture model (gmm.model) Bayesian mixture model (bmm.model) K-means clustering (kmeans.model) Label propagation (label_spreading.model) Self-trained classifier (self_trainer.model) Supervised models Balanced random forest (rf_bal.model) Easy ensemble (easy_ensemble.model) K-nearest neighbors (knn.model) Random forest (rf.model) Support vector machines (svm.model) Loading Instructions: The models with the file-ending ".model" are pickeled Python objects and can be loaded with ``pickle.load(your_model.model)``. The random forest must be loaded with ``joblib.load(rf.model)``. All artificial neural networks are h5py.File objects (tf.keras models) and can be loaded with ``tf.keras.models.load_model(your_ann.model)``.
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 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 | 0.001 | 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 | 0.002 |
| 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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, 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".