arXiv subjects

Theophile Jegou Du Laz

Publications and source records attributed to Theophile Jegou Du Laz.

5 recordsLinked to original sources

arXiv2026

AppleCiDEr. II. SpectraNet: A Spectroscopic Neural Network Classifier for Transients Demonstrated on ZTF Follow-up Data

Time-domain surveys such as the Zwicky Transient Facility have opened a new frontier in the discovery and characterization of transients. While photometric light curves provide broad temporal coverage, spectroscopic observations remain crucial for physical interpretation and source classification. However, existing spectral analysis methods, often reliant on template fitting or parametric models, are limited in their ability to capture the complex and evolving spectra characteristic of such sources, which are sometimes only available at low resolution. In this work, we introduce SpectraNet, a deep convolutional neural network designed to learn robust representations of optical spectra from transients. Our model combines multi-scale convolution kernels and pooling operations to extract features from preprocessed spectra in a hierarchical and interpretable manner. We train and validate SpectraNet on low-resolution time-series spectra obtained from the Spectral Energy Distribution Machine and other instruments, demonstrating better performance in classification compared to other known pipelines.

astro-ph.IM↗

arXiv2026

Catching Disguised Transients with ASTRANet: Anomaly-Aware Spectroscopic Classification and Conformal Calibration

Time-domain surveys discover thousands of transients per year, but the spectroscopic identification of rare and physically peculiar objects remains rate-limited by closed-set classifiers that confidently assign every input to a known class -- including spectra that genuinely belong to no known class. We present the \texttt{ASTRANet} framework, a confidence-aware infrastructure for spectroscopic transient classification built around three coupled modules: a hierarchical spectral classifier that operates directly on observer-frame spectra without requiring host-galaxy redshift or spectral phase as inputs; an anomaly detection layer (\texttt{ASTRANet-Sentinel}) that non-linearly combines $16$ embedding-space anomaly scores spanning four physically motivated families; and a conformal uncertainty quantification layer (\texttt{ASTRANet-CP}). We validate the framework on a held-out evaluation set of $289$ rare and out-of-taxonomy transients spanning $11$ classes deliberately excluded from training, chosen to span the full physical diversity of the rare-anomaly population: AGN-related outliers, GRB-related events, gap transients, novae, and peculiar supernovae. Through five astrophysically distinct failure modes of closed-set classifiers, we show that classifier-internal uncertainty and embedding-based anomaly detection are structurally complementary axes of confidence rather than alternative implementations of the same estimator. We further introduce AD-stratified Mondrian conformal prediction (AD-MCP) within \texttt{ASTRANet-CP}, achieving uniform conditional coverage across anomaly-score strata where vanilla Mondrian under-covers in the operational regime. This establishes the methodological infrastructure for confidence-aware spectroscopic discovery in the Vera C.\ Rubin Observatory era.

astro-ph.IM↗

arXiv2026

The ZTF-ULTRASAT experiment: Characterizing the non-transients in ULTRASAT's high cadence survey

The forthcoming launch of the Ultraviolet Transient Astronomy Satellite (ULTRASAT) will transform our understanding of the transient ultraviolet sky by increasing our ability to identify transients due to its unprecedented 204 deg2 field of view. While rapid (extragalactic) transients are a priority science area for the mission, flaring stars and AGN can often contaminate searches for such objects. To prepare for these challenges, the Zwicky Transient Facility (ZTF)-ULTRASAT experiment observed five fields at high cadence over three nights, in close proximity to ULTRASAT's three northern high-cadence fields. A real-time filter identified seven transient candidates, of which five were persistent variable sources and two were spurious. Periods and amplitudes derived from the ZTF Source Classification Project (SCoPe) showed that three candidates were RR Lyrae stars with short periods and high amplitudes, while the remaining two displayed flaring behavior. We demonstrate that short-timescale, high-amplitude variables can systematically mimic transient alerts in high-cadence UV surveys, and we provide a concrete strategy to this contamination using pre-existing machine learning catalogs.

astro-ph.SR↗

arXiv2025

Applying multimodal learning to Classify transient Detections Early (AppleCiDEr) I: Data set, methods, and infrastructure

Modern time-domain surveys like the Zwicky Transient Facility (ZTF) and the Legacy Survey of Space and Time (LSST) generate hundreds of thousands to millions of alerts, demanding automatic, unified classification of transients and variable stars for efficient follow-up. We present AppleCiDEr (Applying Multimodal Learning to Classify Transient Detections Early), a novel framework that integrates four key data modalities (photometry, image cutouts, metadata, and spectra) to overcome limitations of single-modality classification approaches. Our architecture introduces (i) two transformer encoders for photometry, (ii) a multimodal convolutional neural network (CNN) with domain-specialized metadata towers and Mixture-of-Experts fusion for combining metadata and images, and (iii) a CNN for spectra classification. Training on ~ 30,000 real ZTF alerts, AppleCiDEr achieves high accuracy, allowing early identification and suggesting follow-up for rare transient spectra. The system provides the first unified framework for both transient and variable star classification using real observational data, with seamless integration into brokering pipelines, demonstrating readiness for the LSST era.

astro-ph.IM↗

arXiv2024

The Zwicky Transient Facility Bright Transient Survey. III. $\texttt{BTSbot}$: Automated Identification and Follow-up of Bright Transients with Deep Learning

The Bright Transient Survey (BTS) aims to obtain a classification spectrum for all bright ($m_\mathrm{peak}\,\leq\,18.5\,$mag) extragalactic transients found in the Zwicky Transient Facility (ZTF) public survey. BTS critically relies on visual inspection ("scanning") to select targets for spectroscopic follow-up, which, while effective, has required a significant time investment over the past $\sim5$ yr of ZTF operations. We present $\texttt{BTSbot}$, a multi-modal convolutional neural network, which provides a bright transient score to individual ZTF detections using their image data and 25 extracted features. $\texttt{BTSbot}$ is able to eliminate the need for daily human scanning by automatically identifying and requesting spectroscopic follow-up observations of new bright transient candidates. $\texttt{BTSbot}$ recovers all bright transients in our test split and performs on par with scanners in terms of identification speed (on average, $\sim$1 hour quicker than scanners). We also find that $\texttt{BTSbot}$ is not significantly impacted by any data shift by comparing performance across a concealed test split and a sample of very recent BTS candidates. $\texttt{BTSbot}$ has been integrated into Fritz and $\texttt{Kowalski}$, ZTF's first-party marshal and alert broker, and now sends automatic spectroscopic follow-up requests for the new transients it identifies. During the month of October 2023, $\texttt{BTSbot}$ selected 296 sources in real-time, 93% of which were real extragalactic transients. With $\texttt{BTSbot}$ and other automation tools, the BTS workflow has produced the first fully automatic end-to-end discovery and classification of a transient, representing a significant reduction in the human-time needed to scan. Future development has tremendous potential for creating similar models to identify and request follow-up observations for specific types of transients.

astro-ph.IM↗