Niharika Sravan

The follow-up of gravitational-wave events by wide-field surveys is a crucial tool for the discovery of electromagnetic counterparts to gravitational wave sources, such as kilonovae. Machine learning tools can play an important role in aiding search efforts. We have developed a public tool to predict kilonova light curves using simulated low-latency alert data from the International Gravitational Wave Network during observing runs 4 (O4) and 5 (O5). It uses a bidirectional long-short-term memory model to forecast kilonova light curves from binary neutron star and neutron star–black hole mergers in the Zwicky Transient Facility (ZTF) and Rubin Observatory’s Legacy Survey of Space and Time filters. The model achieves a test mean squared error (MSE) of 0.12 for ZTF filters and 0.23 for Rubin filters, calculated by averaging the squared error over all time steps, filters, and light curves in the test set. We evaluate the performance of the model against merger events followed-up by the ZTF partnership during O4c. We also analyze the effect of incorporating constraints on physical features such as ejecta mass. Using ejecta mass, the performance of the model improves to an MSE of 0.1 for ZTF filters and 0.15 for Rubin filters. Our model is publicly available and can help to add important information to help plan follow-up of candidate events discovered by current and next-generation public surveys.

We present the discovery of EP250827b/SN 2025wkm, an X-ray Flash (XRF) discovered by the Einstein Probe (EP), accompanied by a broad-line Type Ic supernova (SN Ic-BL) at$z = 0.1194$ . EP250827b possesses a prompt X-ray luminosity of$\sim 10^{45} \, \rm{erg \, s^{-1}}$ , lasts over 1000 seconds, and has a peak energy$E_{\rm{p}} < 1.5$keV at 90% confidence. SN 2025wkm possesses a double-peaked light curve (LC), though its bolometric luminosity plateaus after its initial peak for$\sim 20$days, giving evidence that a central engine is injecting additional energy into the explosion. Its spectrum transitions from a blue to red continuum with clear blueshifted Fe II and Si II broad absorption features, allowing for a SN Ic-BL classification. We do not detect any transient radio emission and rule out the existence of an on-axis, energetic jet$\gtrsim 10^{50}~$ erg. In the model we invoke, the collapse gives rise to a long-lived magnetar, potentially surrounded by an accretion disk. Magnetically-driven winds from the magnetar and the disk mix together, and break out with a velocity$\sim 0.35c$from an extended circumstellar medium with radius$\sim 10^{13}$cm, generating X-ray breakout emission through free-free processes. The disk outflows and magnetar winds power blackbody emission as they cool, producing the first peak in the SN LC. The spin-down luminosity of the magnetar in combination with the radioactive decay of$^{56}$ Ni produces the late-time SN LC. We end by discussing the landscape of XRF-SNe within the context of EP's recent discoveries.

On August 18, 2025, the LIGO-Virgo-KAGRA collaboration reported gravitational waves from a sub-threshold binary neutron star merger. If astrophysical, this event would have a surprisingly low chirp mass, suggesting that at least one neutron star was below a solar mass. The Zwicky Transient Facility mapped the coarse localization and discovered a transient, ZTF25abjmnps (AT2025ulz), that was spatially and temporally coincident with the gravitational wave trigger. The first week of follow-up suggested properties reminiscent of a GW170817-like kilonova. Subsequent follow-up suggests properties most similar to a young, stripped-envelope, Type IIb supernova. Although we cannot statistically rule out chance coincidence, we undertake due diligence analysis to explore the possible association between ZTF25abjmnps and S250818k. Theoretical models have been proposed wherein sub-solar neutron star(s) may form (and subsequently merge) via accretion disk fragmentation or core fission inside a core-collapse supernova i.e. a ``superkilonova". Here, we qualitatively discuss our multi-wavelength dataset in the context of the superkilonova picture. Future higher significance gravitational wave detections of sub-solar neutron star mergers with extensive electromagnetic follow-up would conclusively resolve this tantalizing multi-messenger association.

Time-domain surveys such as the Zwicky Transient Facility (ZTF) 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 multi-scale pooling 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 (SEDM) and other instruments, demonstrating state-of-the-art performance in classification. Furthermore, in redshift prediction tasks, SpectraNet achieves a root mean squared relative redshift error of 0.02, highlighting its effectiveness in precise regression tasks as well.

We present observations of SN 2023xgo, a transitional Type Ibn/Icn SN, from −5.6 to 63 days relative to r-band peak. Early spectra show C iii λ5696 emission like Type Icn SNe, shifting to Type Ibn features. The He i velocities (1800-10000 km s−1) and pseudo-equivalent widths are among the highest in the Ibn/Icn class. The light curve declines at 0.14mag d−1 until 30 days, matching SNe Ibn/Icn but slower than fast transients. SN 2023xgo is the faintest in our SN Ibn sample (Mr = −17.65 ± 0.04) but shows typical colour and host properties. Semi-analytical modelling of the light curve suggests a compact CSM shell (∼1012 − 1013 cm), mass-loss rate between 10−4 − 10−3 M⊙ yr−1 with CSM and ejecta masses of ∼0.22 and 0.12 M⊙, respectively. Post-maximum light-curve, spectral modelling favours a ∼3 M⊙ helium star progenitor with extended (∼1015 cm), stratified CSM (density exponent of 2.9) and mass-loss rate of 0.1 − 2.7 M⊙ yr−1. These two mass-loss regimes imply a radially varying CSM, shaped by asymmetry or changes in the progenitor's mass loss over time. This mass-loss behavior fits both binary and single-star evolution. Early Icn-like features stem from hot carbon ionization, fading to Ibn-like with cooling. SN 2023xgo thus offers rare insight into the connection between SNe Icn, Ibn, and SNe Ibn with ejecta signatures.

Among the various classes of fast optical transients (FOTs), kilonovae (KNe), which can emerge as a result of neutron star mergers, are extremely challenging to observe because of not only the rapid timescale on which they fade (on the order of days), but also due to the relative scarcity of their occurrence. This scarcity is compounded by the large number of other FOTs that may initially resemble the characteristic rise of a KNe. While these objects can be ruled out as candidate KNe by taking spectroscopy, a method of confidently ruling out transients based on photometric analysis alone would be incredibly valuable. We describe the compilation of various “imposter” transients, including a plurality of IIb SNe, and investigate a number of comparative metrics by which one might be able to remove transients from consideration without the use of spectroscopy. We provide a list of these objects and their classifications as well as a glossary of the transient types included in the sample.

We present optical, radio, and X-ray observations of EP250108a/SN 2025kg, a broad-line Type Ic supernova (SN Ic-BL) accompanying an Einstein Probe (EP) fast X-ray transient at z = 0.176. EP250108a/SN 2025kg possesses a double-peaked optical light curve, and its spectrum transitions from a blue underlying continuum to a typical SN Ic-BL spectrum over time. We fit a radioactive decay model to the second peak of the optical light curve and find SN parameters that are consistent with the SN Ic-BL population, while its X-ray and radio properties are consistent with those of low-luminosity GRB (LLGRB) 060218/SN 2006aj. We explore three scenarios to understand the system's multiwavelength emission: (a) SN ejecta interacting with an extended circumstellar medium (CSM), (b) the shocked cocoon of a collapsar-driven jet choked in its stellar envelope, and (c) the shocked cocoon of a collapsar-driven jet choked in an extended CSM. Models (b) and (c) can explain the optical light curve and are also consistent with the radio and X-ray observations. We favor model (c) because it can self-consistently explain both the X-ray prompt emission and first optical peak, but we do not rule out model (b). From the properties of the first peak in model (c), we find evidence that EP250108a/SN 2025kg interacts with an extended CSM and infer an envelope mass M-e similar to 0.1 M-circle dot and radius R-e similar to 4 x 10(13) cm. EP250108a/SN 2025kg's multiwavelength properties make it a close analog to LLGRB 060218/SN 2006aj and highlight the power of early follow-up observations in mapping the environments of massive stars prior to core collapse

QZO.csv
The QZO catalog, which includes 4,849,574 objects and columns as described below, excluding the duplicate objects flag. The classifications are based on XGB models trained on ZTF g-band median magnitude and light curves classification with transformer model, as well as WISE W[1-4] magnitudes and colors. The photo-zs are based on ZTF g-band magnitude and WISE magnitudes and colors. We remove duplicated ZTF light curves by removing objects which within the full ZTF catalog have at least one neighbour within 1 arcsec with more ZTF observation epochs. The final number of quasars was achieved with magnitude, number of observation epochs, and minimum quasar classification probability cuts, such that g < n_obs / 80 + 20.375, where n_obs is the number of ZTF observational epochs per light curve, and p_(QSO) > 0.9, where p_(QSO) is XGB classification probability for the QSO class. The photo-zs are available for 35% of these objects, depending on the availability of WISE observations.
ZTF_all_QSO.csv
This file provides all the columns for 78,078,450 objects classified as QSOs by at least one of the two XGB models with and without the WISE features. There are no cuts applied, and there are no duplicates removed. 26% of objects are marked with the duplicates flag.
train.csv
The train data predictions. This file contains 2,588,221 records, with ZTF ID and duplicates flag missing. Selecting the longest ZTF light curve for each non duplicated SDSS object removed ZTF duplicates.
Catalog columns
ID                                                ZTF identifier
ra                                                right ascension
dec                                             declination
n_obs                                         number of ZTF observation epochs
is_duplicate                               flag indicating duplicated light curves
mag_median                             ZTF g-band median magnitude
p_[galaxy, QSO, star]               classification probabilities
p_WISE_[galaxy, QSO, star]    classifications with added WISE data
redshift                                      redshift estimate
ANN_clf.[data-00000-of-00001, index]
ANN model for classification of ZTF g-band light curves. ANN model is trained on ZTF g-band data with at least 20 observation epochs per light curve. It does not require scaling of input light curves, which is done separately for each light curve as part of the transformer model. An example on how to load and use the ANN can be found in the script “run_inference.py” in the GitHub repository.
XGB_clf__ZTF_[PS, WISE, GAIA, PS_WISE, PS_GAIA, WISE_GAIA, PS_WISE_GAIA].pickle
XGB_z__ZTF_WISE.pickle
XGB classification and redshift models for different combinations of input surveys. XGB classification models are trained on all ZTF data with available ANN classification, learning to classify missing features. The XGB redshift model does not include ANN classification as features. An example on how to load and use XGB models can be found in the script “run_inference_XGB.py” in the GitHub repository.
Features order
ZTF         g_mag_median, p_ANN_galaxy, p_ANN_QSO, p_ANN_star
PS           g, r, i, z, g - r, g - i, g - z, r - i, r - z, i - z
WISE      W1, W2, W3, W4, W1 - W2, W1 - W3, W1 - W4, W2 - W3, W2 - W4, W3 - W4
GAIA       g_mean_mag, parallax, pmra, pmdec, bp_mean_mag, rp_mean_mag, bp_rp_excess_factor
The exact column names can be found in the script “features.py” in the Github repository.

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.

Follow-up of gravitational-wave events by wide-field surveys is a crucial tool for the discovery of electromagnetic counterparts to gravitational wave sources, such as kilonovae. Machine learning tools can play an important role in aiding search efforts. We have developed a public tool to predict kilonova light curves using simulated low-latency alert data from the International Gravitational Wave Network during observing runs 4 (O4) and 5 (O5). It uses a bidirectional long-short-term memory (LSTM) model to forecast kilonova light curves from binary neutron star and neutron star-black hole mergers in the Zwicky Transient Facility (ZTF) and Rubin Observatory's Legacy Survey of Space and Time filters. The model achieves a test mean squared error (MSE) of 0.19 for ZTF filters and 0.22 for Rubin filters, calculated by averaging the squared error over all time steps, filters, and light curves in the test set. We verify the performance of the model against merger events followed-up by the ZTF partnership during O4a and O4b. We also analyze the effect of incorporating skymaps and constraints on physical features such as ejecta mass through a hybrid convolutional neural network and LSTM model. Using ejecta mass, the performance of the model improves to an MSE of 0.1. However, using full skymap information results in slightly lower model performance. Our models are publicly available and can help to add important information to help plan follow-up of candidate events discovered by current and next-generation public surveys.