arXiv subjects

Ioana Ciucă

Publications and source records attributed to Ioana Ciucă.

At least 19 recordsLinked to original sources

arXiv2026

How to Craft the Right Language AI Policy For Your Research Group (Some Assembly Required)

Language AI is rapidly becoming part of the astronomy research ecosystem, prompting research teams to develop policies governing its use. But resources and advice for AI adoption assume that all research groups share the same goals and values. This paper lays out an argument for why there is no single "correct" AI policy for astronomy research groups. Instead, we introduce four research laboratory archetypes with competing research priorities, and we use them to explore how a small laboratory or research group's priorities shape decisions about AI's impact on research productivity, scientist development, scientific integrity, and data governance. Rather than prescribing a universal set of rules, this paper provides a framework for aligning AI policies with a group's scientific values and mission. The objective is to help research leaders decide how Language AI should be used within their particular research environment.

astro-ph.IM↗

arXiv2026

High-fidelity stellar extinction with Gaia and APOGEE -- I. The method and a new extinction curve

The scarcity of high-fidelity extinction measurements remains a bottleneck in deriving accurate stellar properties from Gaia parallaxes. In this work, we aim to derive precision extinction estimates for APOGEE DR19 stars, establishing a new benchmark for Galactic stellar population studies. We first determine reddening by comparing observed colorsr, etrieved from photometric surveys or standardized synthetic magnitudes from Gaia BP/RP spectra, to intrinsic colors predicted via an XGBoost model. The model is trained on minimally reddened stars to infer intrinsic colors and their associated uncertainties, using APOGEE stellar parameters (Teff, logg, [Fe/H], and [alpha/Fe]). The derived reddening values are then converted into extinctions using an anchor ratio of A_BP / A_RP = 1.694 +/- 0.004, derived from red-clump-like stars. Here, we provide extinction measurements in 39 filters across 10 photometric systems and introduce a new empirical extinction curve optimized for broadband passbands. Our extinction estimates (Av) outperform existing results (Bayestar19, StarHorse, SEDEX), achieving a typical precision of 0.03 mag in Av. Notably, we identify systematic deviations of up to 30% between monochromatic and passband-integrated extinction ratios at wavelengths greater than 700 nm. This result highlights the necessity of adopting passband-specific coefficients when correcting extinction to derive stellar parameters. The derived extinction and reddening data are available to the community for download through Zenodo.

astro-ph.SR↗

arXiv2025

Thick-to-thin disc transition and gas disc shrinking induced by the Gaia-Sausage-Enceladus merger

Understanding the Milky Way disc formation requires characterising its structural and kinematic properties as functions of stellar age. Using red giant stars from APOGEE DR17 and Gaia DR3, we model the age-dependent stellar kinematics with a quasi-isothermal distribution function and fit disc parameters as a function of age using non-parametric splines. We identify a transition from thick to thin disc populations around 10 Gyr ago. Stars older than this have short scale lengths ($\sim$1.7 kpc), typical of the thick disc, while younger stars exhibit increasing scale length with decreasing age, consistent with inside-out formation of the thin disc. This transition possibly coincides with the end of the starburst triggered by the Gaia-Sausage-Enceladus (GSE) merger. Stars formed around 10 Gyr ago exhibit a dip in scale length, even shorter than that of the thick disc. Comparison with an Auriga simulation suggests that this scale-length dip reflects gas disc shrinking caused by the transition from a cold to hot gas accretion mode. We propose the following disc formation scenario: (1) the thick disc formed under cold-mode accretion; (2) the GSE merger triggered a starburst and increased the total mass of the Galaxy, causing the transition to hot-mode accretion; (3) rapid gas consumption led to temporary shrinking of the star-forming gas disc; and then (4) thin disc grows in an inside-out fashion, as the size of the star-forming gas disc grows via hot-mode smooth gas accretion.

astro-ph.GA↗

arXiv2025

From Queries to Criteria: Understanding How Astronomers Evaluate LLMs

There is growing interest in leveraging LLMs to aid in astronomy and other scientific research, but benchmarks for LLM evaluation in general have not kept pace with the increasingly diverse ways that real people evaluate and use these models. In this study, we seek to improve evaluation procedures by building an understanding of how users evaluate LLMs. We focus on a particular use case: an LLM-powered retrieval-augmented generation bot for engaging with astronomical literature, which we deployed via Slack. Our inductive coding of 368 queries to the bot over four weeks and our follow-up interviews with 11 astronomers reveal how humans evaluated this system, including the types of questions asked and the criteria for judging responses. We synthesize our findings into concrete recommendations for building better benchmarks, which we then employ in constructing a sample benchmark for evaluating LLMs for astronomy. Overall, our work offers ways to improve LLM evaluation and ultimately usability, particularly for use in scientific research.

cs.CL↗

arXiv2025

Supernova Event Dataset: Interpreting Large Language Models' Personality through Critical Event Analysis

Large Language Models (LLMs) are increasingly integrated into everyday applications. As their influence grows, understanding their decision making and underlying personality becomes essential. In this work, we interpret model personality using our proposed Supernova Event Dataset, a novel dataset with diverse articles spanning biographies, historical events, news, and scientific discoveries. We use this dataset to benchmark LLMs on extracting and ranking key events from text, a subjective and complex challenge that requires reasoning over long-range context and modeling causal chains. We evaluate small models like Phi-4, Orca 2, and Qwen 2.5, and large, stronger models such as Claude 3.7, Gemini 2.5, and OpenAI o3, and propose a framework where another LLM acts as a judge to infer each model's personality based on its selection and classification of events. Our analysis shows distinct personality traits: for instance, Orca 2 demonstrates emotional reasoning focusing on interpersonal dynamics, while Qwen 2.5 displays a more strategic, analytical style. When analyzing scientific discovery events, Claude Sonnet 3.7 emphasizes conceptual framing, Gemini 2.5 Pro prioritizes empirical validation, and o3 favors step-by-step causal reasoning. This analysis improves model interpretability, making them user-friendly for a wide range of diverse applications. Project Page - https://www.supernova-event.ai/

cs.CL↗

arXiv2025

Sparks of Science: Hypothesis Generation Using Structured Paper Data

Generating novel and creative scientific hypotheses is a cornerstone in achieving Artificial General Intelligence. Large language and reasoning models have the potential to aid in the systematic creation, selection, and validation of scientifically informed hypotheses. However, current foundation models often struggle to produce scientific ideas that are both novel and feasible. One reason is the lack of a dedicated dataset that frames Scientific Hypothesis Generation (SHG) as a Natural Language Generation (NLG) task. In this paper, we introduce HypoGen, the first dataset of approximately 5500 structured problem-hypothesis pairs extracted from top-tier computer science conferences structured with a Bit-Flip-Spark schema, where the Bit is the conventional assumption, the Spark is the key insight or conceptual leap, and the Flip is the resulting counterproposal. HypoGen uniquely integrates an explicit Chain-of-Reasoning component that reflects the intellectual process from Bit to Flip. We demonstrate that framing hypothesis generation as conditional language modelling, with the model fine-tuned on Bit-Flip-Spark and the Chain-of-Reasoning (and where, at inference, we only provide the Bit), leads to improvements in the overall quality of the hypotheses. Our evaluation employs automated metrics and LLM judge rankings for overall quality assessment. We show that by fine-tuning on our HypoGen dataset we improve the novelty, feasibility, and overall quality of the generated hypotheses. The HypoGen dataset is publicly available at huggingface.co/datasets/UniverseTBD/hypogen-dr1.

cs.CL↗

arXiv2025

A Survey on Hypothesis Generation for Scientific Discovery in the Era of Large Language Models

Hypothesis generation is a fundamental step in scientific discovery, yet it is increasingly challenged by information overload and disciplinary fragmentation. Recent advances in Large Language Models (LLMs) have sparked growing interest in their potential to enhance and automate this process. This paper presents a comprehensive survey of hypothesis generation with LLMs by (i) reviewing existing methods, from simple prompting techniques to more complex frameworks, and proposing a taxonomy that categorizes these approaches; (ii) analyzing techniques for improving hypothesis quality, such as novelty boosting and structured reasoning; (iii) providing an overview of evaluation strategies; and (iv) discussing key challenges and future directions, including multimodal integration and human-AI collaboration. Our survey aims to serve as a reference for researchers exploring LLMs for hypothesis generation.

cs.CL↗

arXiv2024

Milky Way Disk

Our understanding of the Milky Way disk is rapidly improving with the recent advent of the high quality and vast amount of observational data. We summarize our current view of the structure of the Milky Way disk, such as the masses and sizes of the gas and stellar disks, and the position and motion of the Sun in the disk. We also discuss the different definitions of the thick and thin disks of the Milky Way, the non-axisymmetric structures of the stellar disk, such as the bar and spiral arms, and the radial migration which can be triggered by these non-axisymmetric stellar structures. After the revolutionary data from the European Space Agency's Gaia mission, our view of the Milky Way disk has been transformed to a non-equilibrium system with many complicated structures in stellar kinematic distribution. We also summarize the recent findings of Galactoseismology research. These detailed observational data provide the archaeological information for us to unveil the formation and evolution history of the Milky Way disk, with the aid of the high-resolution numerical simulations of the Milky Way-like galaxy formation. We also discuss the current view of the formation history of the Milky Way disk.

astro-ph.GA↗

arXiv2024

The Multimodal Universe: Enabling Large-Scale Machine Learning with 100TB of Astronomical Scientific Data

We present the MULTIMODAL UNIVERSE, a large-scale multimodal dataset of scientific astronomical data, compiled specifically to facilitate machine learning research. Overall, the MULTIMODAL UNIVERSE contains hundreds of millions of astronomical observations, constituting 100\,TB of multi-channel and hyper-spectral images, spectra, multivariate time series, as well as a wide variety of associated scientific measurements and "metadata". In addition, we include a range of benchmark tasks representative of standard practices for machine learning methods in astrophysics. This massive dataset will enable the development of large multi-modal models specifically targeted towards scientific applications. All codes used to compile the MULTIMODAL UNIVERSE and a description of how to access the data is available at https://github.com/MultimodalUniverse/MultimodalUniverse

astro-ph.IM↗

arXiv2024

New Evidence of Binarity in Young α-Rich Turn-off and Subgiant Stars: Fast Rotation and Strong Magnetic Activity

Young α-rich (YAR) stars within the old Galactic thick disk exhibit a dual characteristic of relative youth determined with asteroseismology and abundance enhancement in α elements measured from high-resolution spectroscopy. The youth origin of YAR stars has been proposed to be binary evolution via mass transfer or stellar mergers. If that is the case, YAR stars should spin rapidly and thus be magnetically active, because they are mass and angular momentum gainers. In this study, to seek this binary footprint we select YAR stars on the main-sequence turn-off or the subgiant branch (MSTO-SGB) from APOGEE DR17, whose ages and projected rotation velocities (vsini) can be precisely measured. With APOGEE vsini and LAMOST spectra, we find that YAR stars are indeed fast rotators and magnetically active. In addition, we observe low [C/N] ratios and high Gaia RUWE in some YAR stars, suggesting that these MSTO-SGB stars probably have experienced mass transfer from red-giant companions. Our findings underscore that magnetic activity can serve as a valuable tool for probing the binary evolution for other chemically peculiar stars, such as red giants with lithium anomalies and carbon-enhanced metal-poor stars.

astro-ph.SR↗

arXiv2024

Cosmological evolution of metallicity correlation functions from the Auriga simulations

We study the cosmological evolution of the two-point correlation functions of galactic gas-phase metal distributions using the 28 simulated galaxies from the Auriga Project. Using mock observations of the $z = 0$ snapshots to mimic our past work, we show that the correlation functions of the simulated mock observations are well matched to the correlation functions measured from local galaxy surveys. This comparison suggests that the simulations capture the processes important for determining metal correlation lengths, the key parameter in metallicity correlation functions. We investigate the evolution of metallicity correlations over cosmic time using the true simulation data, showing that individual galaxies undergo no significant systematic evolution in their metal correlation functions from $z\sim 3$ to today. In addition, the fluctuations in metal correlation length are correlated with but lag ahead fluctuations in star formation rate. This suggests that re-arrangement of metals within galaxies occurs at a higher cadence than star formation activity, and is more sensitive to the changes of environment, such as galaxy mergers, gas inflows / outflows, and fly-bys.

astro-ph.GA↗

arXiv2024

3D NLTE Lithium abundances for late-type stars in GALAH DR3

Lithium's susceptibility to burning in stellar interiors makes it an invaluable tracer for delineating the evolutionary pathways of stars, offering insights into the processes governing their development. Observationally, the complex Li production and depletion mechanisms in stars manifest themselves as Li plateaus, and as Li-enhanced and Li-depleted regions of the HR diagram. The Li-dip represents a narrow range in effective temperature close to the main-sequence turn-off, where stars have slightly super-solar masses and strongly depleted Li. To study the modification of Li through stellar evolution, we measure 3D non-local thermodynamic equilibrium (NLTE) Li abundance for 581 149 stars released in GALAH DR3. We describe a novel method that fits the observed spectra using a combination of 3D NLTE Li line profiles with blending metal line strength that are optimized on a star-by-star basis. Furthermore, realistic errors are determined by a Monte Carlo nested sampling algorithm which samples the posterior distribution of the fitted spectral parameters. The method is validated by recovering parameters from a synthetic spectrum and comparing to 26 stars in the Hypatia catalogue. We find 228 613 Li detections, and 352 536 Li upper limits. Our abundance measurements are generally lower than GALAH DR3, with a mean difference of 0.23 dex. For the first time, we trace the evolution of Li-dip stars beyond the main sequence turn-off and up the subgiant branch. This is the first 3D NLTE analysis of Li applied to a large spectroscopic survey, and opens up a new era of precision analysis of abundances for large surveys.

astro-ph.SR↗

arXiv2023

Experimenting with Large Language Models and vector embeddings in NASA SciX

Open-source Large Language Models enable projects such as NASA SciX (i.e., NASA ADS) to think out of the box and try alternative approaches for information retrieval and data augmentation, while respecting data copyright and users' privacy. However, when large language models are directly prompted with questions without any context, they are prone to hallucination. At NASA SciX we have developed an experiment where we created semantic vectors for our large collection of abstracts and full-text content, and we designed a prompt system to ask questions using contextual chunks from our system. Based on a non-systematic human evaluation, the experiment shows a lower degree of hallucination and better responses when using Retrieval Augmented Generation. Further exploration is required to design new features and data augmentation processes at NASA SciX that leverages this technology while respecting the high level of trust and quality that the project holds.

cs.CL↗

arXiv2023

Galactic ChitChat: Using Large Language Models to Converse with Astronomy Literature

We demonstrate the potential of the state-of-the-art OpenAI GPT-4 large language model to engage in meaningful interactions with Astronomy papers using in-context prompting. To optimize for efficiency, we employ a distillation technique that effectively reduces the size of the original input paper by 50\%, while maintaining the paragraph structure and overall semantic integrity. We then explore the model's responses using a multi-document context (ten distilled documents). Our findings indicate that GPT-4 excels in the multi-document domain, providing detailed answers contextualized within the framework of related research findings. Our results showcase the potential of large language models for the astronomical community, offering a promising avenue for further exploration, particularly the possibility of utilizing the models for hypothesis generation.

cs.CL↗

arXiv2023

AstroLLaMA: Towards Specialized Foundation Models in Astronomy

Large language models excel in many human-language tasks but often falter in highly specialized domains like scholarly astronomy. To bridge this gap, we introduce AstroLLaMA, a 7-billion-parameter model fine-tuned from LLaMA-2 using over 300,000 astronomy abstracts from arXiv. Optimized for traditional causal language modeling, AstroLLaMA achieves a 30% lower perplexity than Llama-2, showing marked domain adaptation. Our model generates more insightful and scientifically relevant text completions and embedding extraction than state-of-the-arts foundation models despite having significantly fewer parameters. AstroLLaMA serves as a robust, domain-specific model with broad fine-tuning potential. Its public release aims to spur astronomy-focused research, including automatic paper summarization and conversational agent development.

astro-ph.IM↗

arXiv2023

Harnessing the Power of Adversarial Prompting and Large Language Models for Robust Hypothesis Generation in Astronomy

This study investigates the application of Large Language Models (LLMs), specifically GPT-4, within Astronomy. We employ in-context prompting, supplying the model with up to 1000 papers from the NASA Astrophysics Data System, to explore the extent to which performance can be improved by immersing the model in domain-specific literature. Our findings point towards a substantial boost in hypothesis generation when using in-context prompting, a benefit that is further accentuated by adversarial prompting. We illustrate how adversarial prompting empowers GPT-4 to extract essential details from a vast knowledge base to produce meaningful hypotheses, signaling an innovative step towards employing LLMs for scientific research in Astronomy.

astro-ph.IM↗

arXiv2023

Chasing the impact of the Gaia-Sausage-Enceladus merger on the formation of the Milky Way thick disc

We employ our Bayesian Machine Learning framework BINGO (Bayesian INference for Galactic archaeOlogy) to obtain high-quality stellar age estimates for 68,360 red giant and red clump stars present in the 17th data release of the Sloan Digital Sky Survey, the APOGEE-2 high-resolution spectroscopic survey. By examining the denoised age-metallicity relationship of the Galactic disc stars, we identify a drop in metallicity with an increase in [Mg/Fe] at an early epoch, followed by a chemical enrichment episode with increasing [Fe/H] and decreasing [Mg/Fe]. This result is congruent with the chemical evolution induced by an early-epoch gas-rich merger identified in the Milky Way-like zoom-in cosmological simulation Auriga. In the initial phase of the merger of Auriga 18 there is a drop in metallicity due to the merger diluting the metal content and an increase in the [Mg/Fe] of the primary galaxy. Our findings suggest that the last massive merger of our Galaxy, the Gaia-Sausage-Enceladus, was likely a significant gas-rich merger and induced a starburst, contributing to the chemical enrichment and building of the metal-rich part of the thick disc at an early epoch.

astro-ph.GA↗

arXiv2021

Unveiling the Distinct Formation Pathways of the Inner and Outer Discs of the Milky Way with Bayesian Machine Learning

We develop a Bayesian Machine Learning framework called BINGO (Bayesian INference for Galactic archaeOlogy) centred around a Bayesian neural network. After being trained on the APOGEE and \emph{Kepler} asteroseismic age data, BINGO is used to obtain precise relative stellar age estimates with uncertainties for the APOGEE stars. We carefully construct a training set to minimise bias and apply BINGO to a stellar population that is similar to our training set. We then select the 17,305 stars with ages from BINGO and reliable kinematic properties obtained from \textit{Gaia} DR2. By combining the age and chemo-kinematical information, we dissect the Galactic disc stars into three components, namely, the thick disc (old, high-[$α$/Fe], [$α$/Fe] $\gtrsim$ 0.12), the thin disc (young, low-[$α$/Fe]) and the Bridge, which is a region between the thick and thin discs. Our results indicate that the thick disc formed at an early epoch only in the inner region, and the inner disc smoothly transforms to the thin disc. We found that the outer disc follows a different chemical evolution pathway from the inner disc. The outer metal-poor stars only start forming after the compact thick disc phase has completed and the star-forming gas disc extended outwardly with metal-poor gas accretion. We found that in the Bridge region the range of [Fe/H] becomes wider with decreasing age, which suggests that the Bridge region corresponds to the transition phase from the smaller chemically well-mixed thick to a larger thin disc with a metallicity gradient.

astro-ph.GA↗