91̽»¨

Abstract:

Neutron star-black hole (NSBH) mergers, detectable via their gravitational-wave (GW) emission, are expected to produce kilonovae (KNe). Four NSBH candidates have been identified and followed-up by more than fifty instruments since the start of the fourth GW observing run (O4), in May 2023, up to July 2024; however, no confirmed associated KN has been detected. This study evaluates ejecta properties from multimessenger observations to understand the absence of detectable KN: we use GW public information and joint observations taken from 05.2023 to 07.2024 (LVK, ATLAS, DECam, GECKO, GOTO, GRANDMA, SAGUARO, TESS, WINTER, ZTF). First, our analysis on follow-up observation strategies shows that, on average, more than 50% of the simulated KNe associated with NSBH mergers reach their peak luminosity around one day after merger in the g, r, i- bands, which is not necessarily covered for each NSBH GW candidate. We also analyze the trade-off between observation efficiency and the intrinsic properties of the KN emission, to understand the impact on how these constraints affect our ability to detect the KN, and underlying ejecta properties for each GW candidate. In particular, we can only confirm the kilonova was not missed for 1% of the GW230529 and S230627c sky localization region, given the large sky localization error of GW230529 and the large distance for S230627c and, their respective KN faint luminosities. More constraining, for S230518h, we infer the dynamical ejecta and postmerger disk wind ejecta mdyn,mwind<0.03M⊙ and the viewing angle θ>25°. Similarly, the nonastrophysical origin of S240422ed is likely further confirmed by the fact that we would have detected even a faint KN at the time and presumed distance of the S240422ed event candidate, within a minimum 45% credible region of the sky area, that can be larger depending on the KN scenario.

Abstract:

The Time Domain Extragalactic Survey (TiDES) conducted on the 4 m Multi-Object Spectroscopic Telescope will perform spectroscopic follow-up of extragalactic transients discovered in the era of the NSF-DOE Vera C. Rubin Observatory. TiDES will conduct a 5 yr survey, covering >14, 000squaredegrees , and use around 250,000 fibre hours to address three main science goals: (i) spectroscopic observations of >30,000 live transients, (ii) comprehensive follow-up of >200,000 host galaxies to obtain redshift measurements, and (iii) repeat spectroscopic observations of active galactic nuclei to enable reverberation mapping studies. The live spectra from TiDES will be used to reveal the diversity and astrophysics of both normal and exotic supernovae across the luminosity-timescale plane. The extensive host-galaxy redshift campaign will allow exploitation of the larger sample of supernovae and improve photometric classification, providing the largest-ever sample of SNe Ia, capable of a sub-2% measurement of the equation-of-state of dark energy. Finally, the TiDES reverberation mapping experiment of 700–1000 AGN will complement the SN Ia sample and extend the Hubble diagram to z ∼ 2.5.

Abstract:

Modern astronomical surveys deliver immense volumes of transient detections, yet distinguishing real astrophysical signals (for example, explosive events) from bogus imaging artefacts remains a challenge. Convolutional neural networks are effectively used for real versus bogus classification; however, their reliance on opaque latent representations hinders interpretability. Here we show that large language models (LLMs) can approach the performance level of a convolutional neural network on three optical transient survey datasets (Pan-STARRS, MeerLICHT and ATLAS) while simultaneously producing direct, human-readable descriptions for every candidate. Using only 15 examples and concise instructions, Google’s LLM, Gemini, achieves a 93% average accuracy across datasets that span a range of resolution and pixel scales. We also show that a second LLM can assess the coherence of the output of the first model, enabling iterative refinement by identifying problematic cases. This framework allows users to define the desired classification behaviour through natural language and examples, bypassing traditional training pipelines. Furthermore, by generating textual descriptions of observed features, LLMs enable users to query classifications as if navigating an annotated catalogue, rather than deciphering abstract latent spaces. As next-generation telescopes and surveys further increase the amount of data available, LLM-based classification could help bridge the gap between automated detection and transparent, human-level understanding.