StorSeismic Deep Learning Paradigm for Seismic Processing: Attention is All What a Seismic Dataset Needs
Bibliographic record
Abstract
Summary Seismic datasets exhibit distinct characteristics from subsurface properties, survey parameters, and noise conditions. Capturing these details in a pre-trained neural network forms the foundation for stream-lined applications in seismic processing tasks like denoising, demultiple, etc. Our framework, named StorSeismic, takes advantage of shared features within processing tasks stored in the pre-training of a Bidirectional Encoder Representations from Transformers (BERT) model. The self-attention mechanism of BERT utilized in StorSeismic between traces (channels) of the shot gather is instrumental in capturing the geometry and continuity of recorded waves, and offers robustness against overfitting and adversarial attacks. As a result, our network handles seismic gathers, less like images, and more like a sequence of recordings, in which the attention mechanism is solely responsible for the communication between the traces. In other words, the time and offset axes of the shot gather are handled differently and independently, more in line with how we conventionally process seismic data. We share variations to the attention mechanism that may offer better storage of the seismic dataset features at a reduced cost. Considering our analysis objective here, we demonstrate some of the features of StorSeismic on a simple toy example, with more realistic examples presented at the meeting.
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 | 0.001 | 0.001 |
| 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 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
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".