Timothy Beers

Professor

University of Notre Dame
Country flagUnited States

Research Interests

ApplyKite TwinLearns as you apply

Make your next move.

Prepare your application

Review drafts before you send them.

Explore related searches

Articles (21)

Year:

2026

In recent years, large language models (LLMs) have transformed natural language understanding through vast data sets and large-scale parameterization. Inspired by this success, we present SpecCLIP, a foundation model framework that extends LLM-inspired methodologies to stellar spectral analysis. Stellar spectra, akin to structured language, encode rich physical and chemical information about stars. By training foundation models on large-scale spectral data sets, our goal is to learn robust and informative embeddings that support diverse downstream applications. As a proof of concept, SpecCLIP involves pretraining on two spectral types—LAMOST low-resolution and Gaia XP—followed by contrastive alignment using the Contrastive Language–Image Pretraining (CLIP) framework, adapted to associate spectra from different instruments. This alignment is complemented by auxiliary decoders that preserve spectrum-specific information and enable translation (prediction) between spectral types, the former being achieved by maximizing mutual information between embeddings and input spectra. The result is a cross-spectrum framework that enables intrinsic calibration and flexible applications across instruments. We demonstrate that fine-tuning these models on moderate-sized labeled data sets improves adaptability to tasks such as stellar-parameter estimation and chemical-abundance determination. SpecCLIP also enhances the accuracy and precision of parameter estimates benchmarked against external survey data. In addition, its similarity search and cross-spectrum prediction capabilities offer potential for anomaly detection. Our results suggest that contrastively trained foundation models enriched with spectrum-aware decoders can advance precision stellar spectroscopy. Our code SpecCLIP is publicly available on GitHub ✎.

Year:

2026

The rapid neutron-capture process (r-process) is responsible for the creation of roughly half of the elements heavier than iron, including precious metals like silver, gold, and platinum, as well as radioactive elements such as thorium and uranium. Despite its importance, the nature of the astrophysical sites where the r-process occurs, and the detailed mechanisms of its formation, remain elusive. The key to resolving these mysteries lies in the study of chemical signatures preserved in ancient, metal-poor stars. These stars, which formed in the early Universe, retain the chemical fingerprints of early nucleosynthetic events and offer a unique opportunity to trace the origins of r-process elements in the early Galaxy. In this review, we explore the state-of-the-art understanding of r-process nucleosynthesis, focusing on the sites, progenitors, and formation mechanisms. We discuss the role of potential astrophysical sites such as neutron star mergers, core-collapse supernovae, magneto-rotational supernovae, and collapsars, that can play a key role in producing the heavy elements. We also highlight the importance of studying these signatures through high-resolution spectroscopic surveys, stellar archaeology, and multi-messenger astronomy. Recent advancements, such as the gravitational wave event GW170817 and detection of the r-process in the ejecta of its associated kilonovae, have established neutron star mergers as one of the confirmed sites. However, questions remain regarding whether they are the only sites that could have contributed in early epochs or if additional sources are needed to explain the signatures of r-process found in the oldest stars. Additionally, there are strong indications pointing towards additional sources of r-process-rich nuclei in the context of Galactic evolutionary timescales. These are several of the outstanding questions that led to the formation of collaborative efforts such as the R-Process Alliance, which aims to consolidate observational data, modeling techniques, and theoretical frameworks to derive better constraints on deciphering the astrophysical sites and timescales of r-process enrichment in the Galaxy. This review summarizes what has been learned so far, the challenges that remain, and the exciting prospects for future discoveries. The increasing synergy between observational facilities, computational models, and large-scale surveys is poised to transform our understanding of r-process nucleosynthesis in the coming years.

Year:

2025

The China Space Station Telescope (CSST) will conduct a deep and wide imaging survey in the near-ultraviolet, u , g , r , i , z , and y bands. In this work, using theoretical data synthesized from the BOSZ spectra of R. C. Bohlin et al., along with observational data constructed from different sources, we present two methods for estimating stellar parameters from CSST-like photometry. One approach is to estimate metallicity [M/H] and surface gravity log g simultaneously by using the metallicity- and log g -dependent stellar loci. Tests with theoretical data (without photometric errors) result in precisions of 0.088 dex and 0.083 dex for [M/H] and log g , respectively. With 0.01 mag photometric errors, precision is degraded by about a factor of 2, due to degeneracy in [M/H] and log g . Tests with observational data, although with larger photometric errors, result in precisions of 0.10 dex and 0.39 dex for [Fe/H] and log g , respectively, thanks to the strong correlation between stellar colors and log g in real data. The other approach is the giant-dwarf loci method to obtain classifications and metallicity estimates. With the same observational data, it achieves a better [Fe/H] precision of 0.084 dex, due to the stronger constraints imposed on log g . The method also performs well in distinguishing giants from dwarfs, particularly for red or metal-poor giants. This work demonstrates the clear potential of the CSST data, paving the way for stellar-parameter estimates for many billions of stars.

Year:

2025

We present a strictly line-by-line differential analysis of a moderately r -process-enhanced star ( r -I: HD 107752) with respect to a strongly r -process-enhanced star ( r -II: CS 31082-0001) to investigate the possible common origin of their heavy-element nucleosynthesis with high-precision abundances. This study employs European Southern Observatory data archive high-resolution and high-signal-to-noise spectra taken with the Ultraviolet and Visual Echelle Spectrograph Very Large Telescope spectrograph. Considering only the lines in common in both spectra, we estimate differential abundances of 16 light/Fe-peak elements and 15 neutron-capture elements. Abundances of O, Al, Pr, Gd, Dy, Ho, Er, and a detection of Tm in HD 107752, are presented for the first time. We found three distinct features in the differential-abundance pattern. Nearly equal abundances of light elements up to Zn are present in both stars, indicating a common origin for these elements; in addition to no noticable odd–even differential pattern. Regarding the neutron-capture elements, the r -I star exhibits mildly depleted light r -process elements and more depleted heavier r -process elements relative to r -II star. We also show that, among the r -I and r -II stars, the ratio of lighter-to-heavier r -process elements (e.g., [(Sr, Y, Zr)/Eu]) exhibits a decreasing trend with respect to the overall r -process enhancement, forming a continuous sequence from the r -I to the r -II stars. Finally, we discuss the necessity of multiple sites for the formation of r -I stars.

Year:

2025

The origin of very metal-poor (VMP; [Fe/H] ≤ −2.0) stars on planar orbits has been the subject of great attention since their first discovery. However, prior to the release of the Gaia BP/RP (XP) spectra and large photometric samples such as SkyMapper, SAGES, J-PLUS, and S-PLUS, most studies have been limited due to their small sample sizes or strong selection effects. Here, we crossmatch photometric metallicities derived from Gaia XP synthetic photometry and geometric distances from Bailer-Jones et al., and select 12,000 VMP stars (1604 dwarfs and 10,396 giants) with available high-quality astrometry. After calculating dynamical parameter estimates using AGAMA , we employ the nonnegative matrix factorization technique to the v ϕ distribution across bins in Z max (the maximum height above or below the Galactic plane during the stellar orbit). We find three primary populations of the selected VMP stars: halo, disk system, and the Gaia Sausage/Enceladus structure. The fraction of disk-like stars decreases with increasing Z max (as expected), although it is still ∼20% for stars with Z max ∼ 3 kpc. Similar results emerge from the application of the Hayden criterion, which separates stellar populations on the basis of their orbital inclination angles relative to the Galactic plane. We argue that such high fractions of disk-like stars indicate that they are an independent component, rather than originating solely from Galactic building blocks or heating by minor mergers. We suggest that most of these VMP stars are members of the hypothesized “primordial” disk.

Year:

2025

We design an uncertainty-aware cost-sensitive neural network (UA-CSNet) to estimate metallicities from dereddened and corrected Gaia BP/RP (XP) spectra for giant stars. This method accounts for both stochastic errors in the input spectra and the imbalanced density distribution in [Fe/H] values. With a specialized architecture and training strategy, the UA-CSNet improves the precision of the predicted metallicities, especially for very metal-poor (VMP; [Fe/H] ≤ −2.0) stars. With the PASTEL catalog as the training sample, our model can estimate metallicities down to [Fe/H] ∼ −4. We compare our estimates with a number of external catalogs and conduct tests using star clusters, finding overall good agreement. We also confirm that our estimates for VMP stars are unaffected by carbon enhancement. Applying the UA-CSNet, we obtain reliable and precise metallicity estimates for approximately 20 million giant stars, including 360,000 VMP stars and 50,000 extremely metal-poor ([Fe/H] ≤ −3.0) stars. The resulting catalog is publicly available via the Chinese Virtual Observatory at doi: 10.12149/101604. This work highlights the potential of low-resolution spectra for metallicity estimation and provides a valuable data set for studying the formation and chemodynamical evolution of our Galaxy.

Year:

2025

R -process-enhanced (RPE) stars are rare and typically metal-poor ([Fe/H] +0.70) located in the Milky Way disk, with [Fe/H] of −0.34 and −0.80, respectively. These two stars are selected from the Large Sky Area Multi-Object Fiber Spectroscopic Telescope medium-resolution ( R ∼ 7500) spectroscopic survey. Follow-up high-resolution ( R ∼ 25,000) observations were conducted with the High Optical Resolution Spectrograph installed on the 10.4 m Gran Telescopio Canarias. We perform the determination of elemental abundances and calculate the orbital parameters. We find that they are r -II stars with elemental abundances in agreement with the solar r -process pattern. These two objects are chemically and dynamically consistent with membership in the Galactic disk and exhibit no evidence of being part of accreted systems.

Year:

2025

Based on a template-matching method, we estimate the barium (Ba) abundances for stellar spectra from the Large Sky Area Multi-Object Fiber Spectroscopic Telescope (LAMOST) Medium-Resolution Spectroscopic Survey (MRS). The Ba abundances of 198,011 stars have been determined from MRS spectra with signal-to-noise ratios (S/N) > 40 combined with the stellar atmospheric parameters from the LAMOST Low-Resolution Spectroscopic Survey DR9 by the LAMOST Stellar Parameter Pipeline. The uncertainties in the Ba abundances from the LAMOST MRS spectra are less than 0.3 dex when S/N exceeds 40, which align closely with the results based on the high-resolution UVES spectra from the Gaia-ESO survey obtained by spectral synthesis. Further analysis of Ba abundances from repeated observations of the same stars reveals that random errors related to spectral quality remain below 0.3 dex at the same S/N, with a systematic overestimation for the low-S/N spectra. This extensive sample of stellar Ba abundances will enhance studies of the s -, i -, and r -processes, and deepen our understanding of the chemical-evolution history of the Milky Way.

Year:

2025

We present an updated catalog of stellar parameters, including effective temperature, luminosity classification, and metallicity, for over fifty million stars from the SkyMapper Southern Survey (SMSS) DR4 and Gaia DR3. The accuracy of the derived parameters remains consistent with those achieved with SMSS DR2 using the same methods. Thanks to the advancements in SMSS DR4, photometric-metallicity estimates are now available for an unprecedented number of metal-poor stars. The catalog includes over 13 million metal-poor ([Fe/H] ≤ −1) stars, nearly three million very metal-poor ([Fe/H] ≤ −2.0) stars, and approximately 120,000 extremely metal-poor ([Fe/H] ≤ −3.0) stars—representing an increase by a factor of 4–6 compared to SMSS DR2. This catalog, combined with other stellar parameters obtained through our efforts, will be made available at the China-VO and Zenodo.

Year:

2025

Year:

2025

Year:

2025

Year:

2025

Year:

2025

Year:

2024

Year:

2024

Year:

2024

Year:

2024

Year:

2024

Year:

2024

Year:

2024

Collaborators (11)

Yuan-Sen Ting

Associate Professor

Ohio State University

UNITED STATES
View details

Huawei Zhang

Associate Professor

Peking University

CHINA
View details

Xuesong Wang

Assistant Professor

Tsinghua University

CHINA
View details

Antonio Kanaan

Associate Professor

Universidade Federal de Santa Catarina

BRAZIL
View details

Ian Roederer

Associate Professor

North Carolina State University

UNITED STATES
View details

Wei Zhu

Assistant Professor

Tsinghua University

CHINA
View details

Vinicius Placco

US National Gemini Office Head

NSF NOIRLab

UNITED STATES
View details

Terese Hansen

Researcher

Stockholm University

SWEDEN
View details

Masaomi Tanaka

Tohoku University

JAPAN
View details

Anna Frebel

Professor

Massachusetts Institute of Technology

UNITED STATES
View details

Lile Wang

Assistant Professor

Peking University

CHINA
View details