Gravitational Lens Parameters Estimation at Intermediate Redshifts Using Convolutional Neural Networks
Bibliographic record
Abstract
Strong gravitational lensing serves as a powerful astrophysical probe, enabling studies of dark matter, galaxy structure, and cosmological parameters. The number of strong gravitational lensing candidates at the galaxy scale is expected to reach O ~ 5 with ongoing and future wide-field galaxy surveys. Current modeling techniques largely rely on conventional fitting methods, such as least squares or maximum likelihood using Markov Chain Monte Carlo, which despite their effectiveness, are computationally expensive and require manual inspection. This motivates the development of faster yet accurate parameter estimation techniques. In this work, we construct a representative training dataset and develop an efficient Convolutional Neural Network to estimate lens parameters: the Einstein radius, axis ratio, and position angle. We utilize data from Public Data Release 3 of the Hyper Suprime-Cam Subaru Strategic Program, selecting lens galaxies in the range 0.3 ≤ z ≤ 0.9 based on the strong-lens probability distribution. We find that the choice of loss function and regularization strategy is critical. To enhance model generalization, we leverage SpatialDropout, which outperforms standard methods by addressing the spatial correlation inherent in convolutional features. Furthermore, prediction accuracy and convergence speed are strongly affected by the distribution of the training data, highlighting the importance of an appropriate loss function. Our optimized model demonstrates robust performance, achieving a Mean Absolute Error of 0.092 arcsec for the Einstein radius, providing a scalable framework for automated analysis in future wide-field surveys.
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 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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".