arXiv subjects

Jorge Martinez-Palomera

Publications and source records attributed to Jorge Martinez-Palomera.

8 recordsLinked to original sources

arXiv2026

Trajectory-Agnostic Asteroid Detection in TESS with Deep Learning

We present a novel method for extracting moving objects from TESS data using machine learning. Our approach uses two stacked 3D U-Nets with skip connections, which we call a W-Net, to filter background and identify pixels containing moving objects in TESS image time-series data. By augmenting the training data through rotation of the image cubes, our method is robust to differences in speed and direction of asteroids, requiring no assumptions for either parameter range which are typically required in "shift-and-stack" type algorithms. We also developed a novel method for learned data scaling that we call Adaptive Normalization, which allows the neural network to learn the ideal range and scaling distribution required for optimal data processing. We built a code for creating TESS training data with asteroid masks that served as the foundation of our effort (tess-asteroid-ml), which we publicly released for the benefit of the community. Our method is not limited to TESS, but applicable for implementation in other similar time-domain surveys, making it of particular interest for use with data from upcoming missions such as the Nancy Grace Roman Space Telescope and NEOSurveyor.

astro-ph.EP↗

arXiv2026

Technique-agnostic exoplanet demography for the Roman era -- I. Testing a demography retrieval framework using simulated Kepler-like transit datasets

The Nancy Grace Roman Space Telescope (Roman) will unveil for the first time the full architecture of planetary systems across Galactic distances through the discovery of up to 200,000 cool and hot exoplanets using microlensing and transit detection methods. Roman's huge exoplanet haul, and Galactic reach, will require new methods to leverage the full exoplanet demographic content of the combined microlensing and transit samples, given the different sensitivity bias of the techniques to planet and host properties and Galactic location. We present a framework for technique-agnostic exoplanet demography (TAED) that can allow large, multi-technique exoplanet samples distributed over Galactic distance scales to be combined for demographic studies. Our TAED forward modelling and retrieval framework uses parameterised model exoplanet demographic distributions to embed planetary systems within a stellar population synthesis model of the Galaxy, enabling internally consistent forecasts to be made for all detection methods that are based on spatio-kinematic system properties. In this paper, as a first test of the TAED framework, we apply it to simulated transit datasets based on the Kepler Data Release 25 to assess parameter recovery accuracy and method scalability for a single large homogeneous dataset. We find that optimisation using differential evolution provides a computationally scalable framework that gives a good balance between computational efficiency and accuracy of parameter recovery.

astro-ph.EP↗

arXiv2026

TSSC comet-centered data products from TESS 3I/ATLAS observations

3I/ATLAS is the third known interstellar object to pass through our Solar System. NASA's Transiting Exoplanet Survey Satellite (TESS) made dedicated observations of 3I/ATLAS between 15 -- 22 January 2026 (Sector 1751), capturing high-cadence observations at 200s and 20s cadence. We present two High Level Science Products (HLSPs): (1) comet-centered image time series, corrected for background scattered light and stars; and (2) aperture light curves extracted from the corrected images. We created these data products using the official TESS products and they are publicly available at the Mikulski Archive for Space Telescopes (MAST). TESS's high-precision, near-continuous photometry will provide unique insights into the comet's activity following its closest approach to the Sun. The TESS Science Support Center (TSSC) has created these data products to facilitate scientific analyses by the TESS and Solar System communities.

astro-ph.EP↗

arXiv2025

Extrapolation of Periodic Functions Using Binary Encoding of Continuous Numerical Values

We report the discovery that binary encoding allows neural networks to extrapolate periodic functions beyond their training bounds. We introduce Normalized Base-2 Encoding (NB2E) as a method for encoding continuous numerical values and demonstrate that, using this input encoding, vanilla multi-layer perceptrons (MLP) successfully extrapolate diverse periodic signals without prior knowledge of their functional form. Internal activation analysis reveals that NB2E induces bit-phase representations, enabling MLPs to learn and extrapolate signal structure independently of position.

cs.LG↗

arXiv2025

Pre-discovery TESS Observations of Interstellar Object 3I/ATLAS

3I/ATLAS, also known as C2025 N1 (ATLAS), is the third known macroscopic interstellar object to pass through our Solar System. We report serendipitous Transiting Exoplanet Survey Satellite (TESS) observations of 3I/ATLAS taken between 2025-05-07 and 2025-06-02, 55 days prior to the discovery date (2025-07-01). We retrieve the TESS pixel data, perform a robust background correction and use a data-driven approach to compute the objects position on the TESS detectors. We find a consistent offset between the targets observed and predicted positions which is dominated by uncertainty in the TESS World Coordinate System (WCS) rather than ephemeris errors. 3I/ATLAS is too faint to be detected in the individual 200 second TESS integrations, so we stack images to improve detectability. We perform aperture and Pixel Response Function (PRF) photometry on the stacked images to create two light curves. Each light curve consists of 15 measurements with $\text{SNR}>3$, collected across two different TESS cameras during the 26 days that the object was observed. The PRF light curve, which is more robust against image noise, in the TESS bandpass shows a gradual increase in brightness from $T_{\text{mag}}=20.9\pm0.29$ to $19.57\pm0.15$. This is expected as 3I/ATLAS approaches the inner Solar System. Its absolute magnitude decreases from $H_{V}=14.3\pm0.4$ to $13.7\pm0.3$ and shows signs of faint activity consistent with other observations. This paper highlights the power of using TESS for Solar System science by increasing the number of pre-discovery observations, in an otherwise sparsely populated region of the light curve, the long-term behavior of 3I/ATLAS can be investigated.

astro-ph.EP↗

arXiv2023

Kepler Bonus: Light Curves of Kepler Background Sources

NASA's \textit{Kepler} primary mission observed about 116 $deg^2$ in the sky for 3.5 consecutive years to discover Earth-like exoplanets. This mission recorded pixel cutouts, known as Target Pixel Files (TPFs), of over $200,000$ targets selected to maximize the scientific yield. The Kepler pipeline performed aperture photometry for these primary targets to create light curves. However, hundreds of thousands of background sources were recorded in the TPFs and have never been systematically analyzed. This work uses the Linearized Field Deblending (LFD) method, a Point Spread Function (PSF) photometry algorithm, to extract light curves. We use Gaia DR3 as input catalog to extract $606,900$ light curves from long-cadence TPFs. $406,548$ are new light curves of background sources, while the rest are Kepler's targets. These light curves have comparable quality as those computed by the Kepler pipeline, with CDPP values $<100$ ppm for sources $G<16$. The light curve files are available as high-level science products at MAST. Files include PSF and aperture photometry, and extraction metrics. Additionally, we improve the background and PSF modeling in the LFD method. The LFD method is implemented in the \texttt{Python} library \texttt{psfmachine}. We demonstrate the advantages of this new dataset with two examples; deblending of contaminated false positive Kepler Object of Interest identifying the origin of the transit signal; and the changes in estimated transit depth of planets using PSF photometry which improves dilution when compared to aperture photometry. This new nearly unbiased catalog enables further studies in planet search, occurrence rates, and other time-domain studies.

astro-ph.EP↗

arXiv2021

Kepler Bonus: Aperture Photometry Light Curves of EXBA Sources

NASA's Kepler mission observed background regions across its field of view for more than three consecutive years using custom designed super apertures (EXBA masks). Since these apertures were designed to capture a region of the sky rather than single targets, the Kepler Science Data Processing pipeline produced Target Pixel Files, but did not produce light curves for the sources within these background regions. In this work we produce light curves for $9,327$ sources observed in the EXBA masks. These light curves are generated using aperture photometry estimated from the instrument's Pixel Response Function (PRF) profile computed from Kepler's full-frame images. The PRF models enable the creation of apertures that follow the characteristic shapes of the PSF in the image and the computation of flux completeness and contamination metrics. The light curves are available at MAST as a High Level Science Product (kbonus-apexba). Alongside this dataset, we present kepler-apertures, a Python library to compute PRF models and use them to perform aperture photometry on Kepler-like data. Using light curves from the EXBA masks we found an exoplanet candidate around Gaia EDR3 2077240046296834304 consistent with a large planet companion with a $0.81 R_J$ radius. Additionally, we report a catalog of 69 eclipsing binaries. We encourage the community to exploit this new dataset to perform in depth time domain analysis, such as eclipsing binaries demographic and others.

astro-ph.EP↗

arXiv2020

deepSIP: Linking Type Ia Supernova Spectra to Photometric Quantities with Deep Learning

We present {\tt deepSIP} (deep learning of Supernova Ia Parameters), a software package for measuring the phase and -- for the first time using deep learning -- the light-curve shape of a Type Ia supernova (SN~Ia) from an optical spectrum. At its core, {\tt deepSIP} consists of three convolutional neural networks trained on a substantial fraction of all publicly-available low-redshift SN~Ia optical spectra, onto which we have carefully coupled photometrically-derived quantities. We describe the accumulation of our spectroscopic and photometric datasets, the cuts taken to ensure quality, and our standardised technique for fitting light curves. These considerations yield a compilation of 2754 spectra with photometrically characterised phases and light-curve shapes. Though such a sample is significant in the SN community, it is small by deep-learning standards where networks routinely have millions or even billions of free parameters. We therefore introduce a data-augmentation strategy that meaningfully increases the size of the subset we allocate for training while prioritising model robustness and telescope agnosticism. We demonstrate the effectiveness of our models by deploying them on a sample unseen during training and hyperparameter selection, finding that Model~I identifies spectra that have a phase between $-10$ and 18\,d and light-curve shape, parameterised by $Δm_{15}$, between 0.85 and 1.55\,mag with an accuracy of 94.6\%. For those spectra that do fall within the aforementioned region in phase--$Δm_{15}$ space, Model~II predicts phases with a root-mean-square error (RMSE) of 1.00\,d and Model~III predicts $Δm_{15}$ values with an RMSE of 0.068\,mag.

astro-ph.IM↗