Unsupervised ground-roll attenuation via implicit neural representations
Bibliographic record
Abstract
ABSTRACT Coherent noise attenuation in land seismic data is particularly challenging, especially when dealing with ground roll. Unlike incoherent noise, ground roll overlaps with reflections in time-space and frequency-wavenumber domains, making it difficult to separate them without distorting the signal. Traditional attenuation methods often struggle with this overlap, leading to a trade-off between preserving the reflections and effectively reducing noise. Recent advances in deep learning offer promising alternatives, but many rely on supervised learning, which requires a substantial amount of paired training data, which is often unavailable in real-world scenarios. Unsupervised approaches, although avoiding the need for labeled data, frequently face issues such as convergence instability and extensive parameter tuning. We develop an unsupervised deep-learning framework for separating reflections from ground roll to address these challenges. Our method leverages the inherent low-frequency bias of implicit neural representations, which emphasizes self-similarity features during training. The network initially learns to represent smoother, flattened events in seismic data before focusing on features with deeper dips and incoherent noise. To enhance the network’s ability to capture the self-similarity of reflections, we apply a normal moveout (NMO) correction to flatten the reflections before using the network to extract these features from the NMO-corrected data. We further incorporate a horizontal derivative regularization term into the loss function. This term penalizes horizontal variations, ensuring a more stable convergence and reducing the burden of parameter tuning, thereby eliminating the need for early stopping. Our approach is validated with synthetic and real land data examples and compared against traditional f-k filtering methods. The results demonstrate its power in effectively attenuating noise while preserving the integrity of seismic reflections.
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 | 0.000 | 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 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".