arXiv Science⌕ Search
Home/Archive/arXiv

arXiv · 2504.08583

AstroLLaVA: towards the unification of astronomical data and natural language

Sharaf Zaman·Michael J. Smith·Pranav Khetarpal·Rishabh Chakrabarty·Michele Ginolfi·Marc Huertas-Company·Maja Jabłońska·Sandor Kruk·Matthieu Le Lain·Sergio José Rodríguez Méndez·Dimitrios Tanoglidis
View original source ↗Read source document ↗Export metadata ↓

Abstract

We present AstroLLaVA, a vision language model for astronomy that enables interaction with astronomical imagery through natural dialogue. By fine-tuning the LLaVA model on a diverse dataset of $\sim$30k images with captions and question-answer pairs sourced from NASA's `Astronomy Picture of the Day', the European Southern Observatory, and the NASA/ESA Hubble Space Telescope, we create a model capable of answering open-ended questions about astronomical concepts depicted visually. Our two-stage fine-tuning process adapts the model to both image captioning and visual question answering in the astronomy domain. We demonstrate AstroLLaVA's performance on an astronomical visual question answering benchmark and release the model weights, code, and training set to encourage further open source work in this space. Finally, we suggest a roadmap towards general astronomical data alignment with pre-trained language models, and provide an open space for collaboration towards this end for interested researchers.

Explore related subjects

astro-ph.IMcs.LG

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Sharaf Zaman, Michael J. Smith, Pranav Khetarpal, Rishabh Chakrabarty, Michele Ginolfi, Marc Huertas-Company, Maja Jabłońska, Sandor Kruk, Matthieu Le Lain, Sergio José Rodríguez Méndez, Dimitrios Tanoglidis. 2025-04-11. AstroLLaVA: towards the unification of astronomical data and natural language. https://arxiv.org/abs/2504.08583

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Observations of Binary Stars with the 1.3-m Devasthal Fast Optical Telescope Using Speckle Interferometry: An Attempt

We present a feasibility study of implementing optical interferometry and speckle techniques with the 1.3-m Devasthal Fast Optical Telescope (DFOT) at Aryabhatta Research Institute of Observational Sciences (ARIES), which is currently dedicated to photometric observations. Using the sCMOS camera at the DFOT backend, we perform interferometric speckle observations of six binary stars. Standard Speckle Interferometry algorithms are applied to analyze the recorded speckle images. While this study does not aim to achieve the diffraction limit of DFOT or address a full science-driven resolution case, it serves as a crucial testbed for instrumentation, data acquisition, and analysis of Speckles with DFOT. Although the individual speckle patterns exhibit significant positional variations from frame to frame, the Speckle Interferometry analysis successfully recovers the characteristic three-peak structure of the wide binaries, allowing their relative separations and position angles to be determined, demonstrating the viability of the approach. The measured angular relative separations of $γ$ Leo, $γ$ Vir, and $ζ$ Her are $4.744'' \pm 0.012''$, $3.387'' \pm 0.027''$, and $1.582'' \pm 0.026''$, respectively, with corresponding projected position angles of $126.94^{\circ} \pm 0.14^\circ$, $171.57^{\circ} \pm 0.33^\circ$, and $81.06^{\circ} \pm 0.91^\circ$. These measurements are consistent with previously reported values in the literature, which provides strong motivation for more systematic observations and future implementation of optical interferometry techniques at meter-class telescopes.

astro-ph.IM↗
arXiv2026

Time-Domain Synthesis of Gravitational-Wave Detector Glitches using Class-Conditional Derivative Generative Adversarial Networks

Gravitational-wave detectors such as LIGO, Virgo, and KAGRA are highly sensitive instruments susceptible to many noise sources. Short-duration transient noise events, known as glitches, pose a particular challenge for data analysis pipelines, as they can mimic or obscure astrophysical signals. We present GlitchGAN, a class-conditional generative model built on the Conditional Derivative GAN (cDVGAN) architecture, capable of synthesizing seven glitch types from LIGO's third observing run (O3) directly in the time domain. GlitchGAN generalizes effectively, learning to reproduce a diverse glitch space consistent with high-quality DeepExtractor reconstructions, and can generate hybrid glitch morphologies by interpolating across its class-conditioning vector. It generates 1000 glitches in under 22 seconds on a CPU, suitable for detector simulations, mock data challenges, and pipeline validation. Synthetic glitches are validated using the Gravity Spy classifier and UMAP embeddings, both showing strong agreement with real data. To probe residual distributional differences, we train a separate holdout GlitchGAN model and use a downstream CNN to distinguish held-out real glitches from synthetic ones: detectability is high in a clean representation but drops substantially once both populations are injected into realistic detector noise, the condition under which they would typically be used. Despite this, GlitchGAN-generated glitches remain practically useful: augmenting real training sets with synthetic samples matches simple duplication of real data when data is abundant, and increasingly outperforms it as real data becomes scarce. Finally, we highlight a limitation of magnitude-only spectrograms: magnitude Q-transform classifiers can confidently misclassify physically unrealistic glitches from less robust models, underscoring the need for validation methods that preserve phase information.

astro-ph.IM↗
arXiv2026

SGN: A python framework for stream-processing pipelines

We present the Stream Graph Navigator (SGN), a lightweight Python framework for building streaming data applications. In SGN, stream-processing pipelines are built by connecting computational components into directed acyclic graphs that run within an event loop. The time-series extension of the SGN library, SGN-TS, introduces signal processing methods to handle time series data. Together, SGN and SGN-TS provide the foundation for SGNL, a matched-filtering gravitational-wave search pipeline, and are being adopted by multiple projects across the low-latency gravitational-wave data analysis infrastructure as an extensible and maintainable framework for future gravitational-wave observations.

astro-ph.IM↗