This repository contains the dataset and code for baseline models for ADA-X datathon
The simulated dataset used is modeled after the Euclid satellite galaxy survey. When generating a large, realistic, simulated spectroscopic dataset, we need to ensure that it is representative of the expected quality of the Euclid data. The first requirement is to have a realistic distribution of galaxies in several photometric observational parameters. We want the simulated data to follow representative redshift, color, magnitude, and spectral type distributions. These quantities depend on each other in intricate ways, and correctly capturing the correlations is important if we want to have a realistic assessment of the accuracy of our proposed method.
We define a master catalog for the analyses with the COSMOSSNAP simulation pipeline, which calibrates property distributions with real data from the COSMOS survey. The generated COSMOS mock Catalog (CMC) is based on the 30-band COSMOS photometric redshift catalogue with magnitudes, colors, shapes and photometric redshifts for
For the specific purposes of this analysis, we require realistic SEDs and emission line strengths. Euclid will observe approximately 50 million spectra in the wavelength range
The ADAX_datathon_classification.ipynb is jupyter notebook provides an easy-to-follow example for analyzing the data as a classification problem.
Specifically, the code demonstrates the design and training of a 1-dimensional Convolutional Neural Network.
*Input: spectroscopic profile (SED)
*Output: label indicating estimated redshift (rounded)
Use the "open on Colab" to execute the code online.
R. Stivaktakis, G. Tsagkatakis, B. Moares, F. Abdalla, J. L. Starck, and P. Tsakalides, "Convolutional Neural Networks for Spectroscopic Redshift Estimation on Euclid Data," IEEE Transactions on Big Data, vol. 6, no. 3, pp. 460-476, September 2020, doi: 10.1109/TBDATA.2019.2934475. arxiv