Bibliographic record
Abstract
Photometric redshifts (photo-z's) are vital for the success of current and forthcoming cosmological galaxy surveys. This work focuses on three different approaches to enhance photo-z's. Firstly, we study the extent to which galaxy morphology improves photo-z's. Using artificial neural networks, we compare the performances of several morphological parameters and find that galaxy size and surface brightness bring about the most improvement to photo-z's in bright samples. When multiple morphological parameters are used, the improvement in scatter reaches as high as 12% for the Main Galaxy Sample (MGS) of the Sloan Digital Sky Survey (SDSS). We also find that the improvement becomes significant under suboptimal conditions: when surveys have limited numbers of bands, low quality photometry, or an imperfect star-galaxy separator. Next we study aspects of photo-z probability density functions (PDFs) and the resulting redshift distributions of galaxy samples in the context of the Canada-France-Hawaii Telescope Stripe-82 (CS82) Survey. We discover that, while galaxy morphology brings marginal improvement to both, we are able to produce accurate redshift distributions using a single photometric band and multiple galaxy morphological parameters, and apply this to the CS82 survey. As part of the photo-z Working Group of the Large Synoptic Survey Telescope Dark Energy Science Collaboration (LSST-DESC), we use several metrics to assess the performances of two state-of-the-art photo-z codes, ANNz2 and Delight, and concluded that the photo-z's produced by both are close to the standard for the current photo-z requirements of LSST. Finally, we explore the performances of multiple photo-z codes on narrowband surveys. Using simulated and real data from the 40-narrowband Physics of the Accelerating Universe (PAU) Survey, we find that the hybrid spectral template-machine learning code Delight outperforms monolithic machine learning as well as template codes. Using the large suite of spectral templates and well-calibrated additional broadband fluxes, we are able to produce competitive photo-z's close to the nominal PAU requirement at 40% quality cut. We believe these method would be useful for the next generation of photometric surveys, like Euclid and LSST.
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 | 0.008 | 0.037 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.003 | 0.007 |
| Open science | 0.001 | 0.004 |
| 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".