Showing 1–50 of 225 results for author: Narayanan, D

The AURORA Survey: A New Era of Emission-line Diagrams with JWST/NIRSpec

Authors: Alice E. Shapley, Ryan L. Sanders, Michael W. Topping, Naveen A. Reddy, Danielle A. Berg, Rychard J. Bouwens, Gabriel Brammer, Adam C. Carnall, Fergus Cullen, Romeel Davé, James S. Dunlop, Richard S. Ellis, N. M. Förster Schreiber, Steven R . Furlanetto, Karl Glazebrook, Garth D. Illingworth, Tucker Jones, Mariska Kriek, Derek J. McLeod, Ross J. McLure, Desika Narayanan, Pascal Oesch, Anthony J. Pahl, Max Pettini, Daniel Schaerer , et al. (6 additional authors not shown)

Abstract: We present results on the emission-line properties of z=1.4-7.5 star-forming galaxies in the Assembly of Ultradeep Rest-optical Observations Revealing Astrophysics (AURORA) Cycle 1 JWST/NIRSpec program. Based on its depth, continuous wavelength coverage from 1--5 microns, and medium spectral resolution (R~1000), AURORA includes detections of a large suite of nebular emission lines spanning a broad… ▽ More We present results on the emission-line properties of z=1.4-7.5 star-forming galaxies in the Assembly of Ultradeep Rest-optical Observations Revealing Astrophysics (AURORA) Cycle 1 JWST/NIRSpec program. Based on its depth, continuous wavelength coverage from 1--5 microns, and medium spectral resolution (R~1000), AURORA includes detections of a large suite of nebular emission lines spanning a broad range in rest wavelength. We investigate the locations of AURORA galaxies in multiple different emission-line diagrams, including traditional "BPT" diagrams of [OIII]/Hbeta vs. [NII]/Halpha, [SII]/Halpha, and [OI]/Halpha, and the "ionization-metallicity" diagram of [OIII]/[OII] (O32) vs. ([OIII]+[OII])/Hbeta (R23). We also consider a bluer rest-frame "ionization-metallicity" diagram introduced recently to characterize z>10 galaxies: [NeIII]/[OII] vs. ([NeIII]+[OII])/Hdelta; as well as longer-wavelength diagnostic diagrams extending into the rest-frame near-IR: [OIII]/Hbeta vs. [SIII]/[SII] (S32); and HeI/Pagamma and [SIII]/Pagamma vs. [FeII]/Pabeta. With a significant boost in signal-to-noise and large, representative samples of individual galaxy detections, the AURORA emission-line diagrams presented here definitively confirm a physical picture in which chemically-young, alpha-enhanced, massive stars photoionize the ISM in distant galaxies with a harder ionizing spectrum at fixed nebular metallicity than in their z~0 counterparts. We also uncover previously unseen evolution prior to z~2 in the [OIII]/Hbeta vs. [NII]/Halpha diagram, which motivates deep NIRSpec observations at even higher redshift. Finally, we present the first statistical sample of rest-frame near-IR emission-line diagnostics in star-forming galaxies at high redshift. In order to truly interpret rest-frame near-IR line ratios including [FeII], we must obtain better constraints on dust depletion in the high-redshift ISM. △ Less

Submitted 2 July, 2024; v1 submitted 28 June, 2024; originally announced July 2024.

Comments: 17 pages, 8 figures, submitted to ApJ

Nemotron-4 340B Technical Report

Authors: Nvidia, :, Bo Adler, Niket Agarwal, Ashwath Aithal, Dong H. Anh, Pallab Bhattacharya, Annika Brundyn, Jared Casper, Bryan Catanzaro, Sharon Clay, Jonathan Cohen, Sirshak Das, Ayush Dattagupta, Olivier Delalleau, Leon Derczynski, Yi Dong, Daniel Egert, Ellie Evans, Aleksander Ficek, Denys Fridman, Shaona Ghosh, Boris Ginsburg, Igor Gitman, Tomasz Grzegorzek , et al. (58 additional authors not shown)

Abstract: We release the Nemotron-4 340B model family, including Nemotron-4-340B-Base, Nemotron-4-340B-Instruct, and Nemotron-4-340B-Reward. Our models are open access under the NVIDIA Open Model License Agreement, a permissive model license that allows distribution, modification, and use of the models and its outputs. These models perform competitively to open access models on a wide range of evaluation be… ▽ More We release the Nemotron-4 340B model family, including Nemotron-4-340B-Base, Nemotron-4-340B-Instruct, and Nemotron-4-340B-Reward. Our models are open access under the NVIDIA Open Model License Agreement, a permissive model license that allows distribution, modification, and use of the models and its outputs. These models perform competitively to open access models on a wide range of evaluation benchmarks, and were sized to fit on a single DGX H100 with 8 GPUs when deployed in FP8 precision. We believe that the community can benefit from these models in various research studies and commercial applications, especially for generating synthetic data to train smaller language models. Notably, over 98% of data used in our model alignment process is synthetically generated, showcasing the effectiveness of these models in generating synthetic data. To further support open research and facilitate model development, we are also open-sourcing the synthetic data generation pipeline used in our model alignment process. △ Less

Submitted 17 June, 2024; originally announced June 2024.

An Empirical Study of Mamba-based Language Models

Authors: Roger Waleffe, Wonmin Byeon, Duncan Riach, Brandon Norick, Vijay Korthikanti, Tri Dao, Albert Gu, Ali Hatamizadeh, Sudhakar Singh, Deepak Narayanan, Garvit Kulshreshtha, Vartika Singh, Jared Casper, Jan Kautz, Mohammad Shoeybi, Bryan Catanzaro

Abstract: Selective state-space models (SSMs) like Mamba overcome some of the shortcomings of Transformers, such as quadratic computational complexity with sequence length and large inference-time memory requirements from the key-value cache. Moreover, recent studies have shown that SSMs can match or exceed the language modeling capabilities of Transformers, making them an attractive alternative. In a contr… ▽ More Selective state-space models (SSMs) like Mamba overcome some of the shortcomings of Transformers, such as quadratic computational complexity with sequence length and large inference-time memory requirements from the key-value cache. Moreover, recent studies have shown that SSMs can match or exceed the language modeling capabilities of Transformers, making them an attractive alternative. In a controlled setting (e.g., same data), however, studies so far have only presented small scale experiments comparing SSMs to Transformers. To understand the strengths and weaknesses of these architectures at larger scales, we present a direct comparison between 8B-parameter Mamba, Mamba-2, and Transformer models trained on the same datasets of up to 3.5T tokens. We also compare these models to a hybrid architecture consisting of 43% Mamba-2, 7% attention, and 50% MLP layers (Mamba-2-Hybrid). Using a diverse set of tasks, we answer the question of whether Mamba models can match Transformers at larger training budgets. Our results show that while pure SSMs match or exceed Transformers on many tasks, they lag behind Transformers on tasks which require strong copying or in-context learning abilities (e.g., 5-shot MMLU, Phonebook) or long-context reasoning. In contrast, we find that the 8B Mamba-2-Hybrid exceeds the 8B Transformer on all 12 standard tasks we evaluated (+2.65 points on average) and is predicted to be up to 8x faster when generating tokens at inference time. To validate long-context capabilities, we provide additional experiments evaluating variants of the Mamba-2-Hybrid and Transformer extended to support 16K, 32K, and 128K sequences. On an additional 23 long-context tasks, the hybrid model continues to closely match or exceed the Transformer on average. To enable further study, we release the checkpoints as well as the code used to train our models as part of NVIDIA's Megatron-LM project. △ Less

Submitted 12 June, 2024; originally announced June 2024.

Milky Way and Nearby Galaxies Science with the Single Aperture Large Telescope for Universe Studies (\textit{SALTUS}) Space Observatory

Authors: Rebecca C. Levy, Alexander Tielens, Justin Spilker, Daniel P. Marrone, Desika Narayanan, Christopher K. Walker

Abstract: This paper presents an overview of the Milky Way and nearby galaxies science case for the \textit{Single Aperture Large Telescope for Universe Studies} (SALTUS) far-infrared NASA probe-class mission concept. SALTUS offers enormous gains in spatial resolution and spectral sensitivity over previous far-IR missions, thanks to its cold ($<$40~K) 14-m primary mirror. Key Milky Way and nearby galaxies s… ▽ More This paper presents an overview of the Milky Way and nearby galaxies science case for the \textit{Single Aperture Large Telescope for Universe Studies} (SALTUS) far-infrared NASA probe-class mission concept. SALTUS offers enormous gains in spatial resolution and spectral sensitivity over previous far-IR missions, thanks to its cold ($<$40~K) 14-m primary mirror. Key Milky Way and nearby galaxies science goals for SALTUS focus on understanding the role of star formation in feedback in the Local Universe. In addition to this science case, SALTUS would open a new window to to of Galactic and extragalactic communities in the 2030s, enable fundamentally new questions to be answered, and be a far-IR analog to the near- and mid-IR capabilities of JWST. This paper summarizes the Milky Way and nearby galaxies science case and plans for notional observing programs in both guaranteed and guest (open) time. △ Less

Submitted 25 April, 2024; originally announced April 2024.

Comments: Submitted to JATIS as part of a special issue on SALTUS

High-Redshift Extragalactic Science with the Single Aperture Large Telescope for Universe Studies (SALTUS) Space Observatory

Authors: Justin Spilker, Rebecca C. Levy, Daniel Marrone, Stacey Alberts, Scott C. Chapman, Mark Dickinson, Eiichi Egami, Ryan Endsley, Desika Narayanan, George Rieke, Antony A. Stark, Alexander Tielens, Christopher K. Walker

Abstract: This paper presents an overview of the high-redshift extragalactic science case for the Single Aperture Large Telescope for Universe Studies (SALTUS) far-infrared NASA probe-class mission concept. Enabled by its 14m primary reflector, SALTUS offers enormous gains in spatial resolution and spectral sensitivity over previous far-IR missions. SALTUS would be a versatile observatory capable of respond… ▽ More This paper presents an overview of the high-redshift extragalactic science case for the Single Aperture Large Telescope for Universe Studies (SALTUS) far-infrared NASA probe-class mission concept. Enabled by its 14m primary reflector, SALTUS offers enormous gains in spatial resolution and spectral sensitivity over previous far-IR missions. SALTUS would be a versatile observatory capable of responding to the scientific needs of the extragalactic community in the 2030s, and a natural follow-on to the near- and mid-IR capabilities of JWST. Key early-universe science goals for SALTUS focus on understanding the role of galactic feedback processes in regulating galaxy growth across cosmic time, and charting the rise of metals and dust from the early universe to the present. This paper summarizes these science cases and the performance metrics most relevant for high-redshift observations. △ Less

Submitted 25 April, 2024; originally announced April 2024.

Comments: Submitted to JATIS as part of a special issue on SALTUS. Key performance metrics in Figs. 1 and 2

Tracing the evolutionary pathways of dust and cold gas in high-z quiescent galaxies with SIMBA

Authors: G. Lorenzon, D. Donevski, K. Lisiecki, C. Lovell, M. Romano, D. Narayanan, R. Davé, A. Man, K. E. Whitaker, A. Nanni, A. Long, M. M. Lee, Junais, K. Małek, G. Rodighiero, Q. Li

Abstract: Recent discoveries of copious amounts of dust in quiescent galaxies (QGs) at high redshifts ($z\gtrsim 1-2$) challenge the conventional view that these objects have poor interstellar medium (ISM) in proportion to their stellar mass. We use the SIMBA cosmological simulation to explore the evolution of dust and cold gas content in QGs in relation to the quenching processes affecting them. We track t… ▽ More Recent discoveries of copious amounts of dust in quiescent galaxies (QGs) at high redshifts ($z\gtrsim 1-2$) challenge the conventional view that these objects have poor interstellar medium (ISM) in proportion to their stellar mass. We use the SIMBA cosmological simulation to explore the evolution of dust and cold gas content in QGs in relation to the quenching processes affecting them. We track the changes in the ISM dust abundance across the evolutionary history of QGs identified at $0 \lesssim z \lesssim2$ in the field and cluster environments. The QGs quench via diverse pathways, both rapid and slow, and exhibit a wide range of times elapsed between the quenching event and cold gas removal (from $\sim650$ Myr to $\sim8$ Gyr). We find that quenching modes attributed to the feedback from active galactic nuclei (AGN) do not affect dust and cold gas within the same timescales. Remarkably, QGs may replenish their dust content in the quenched phase primarily due to internal processes and marginally by external factors such as minor mergers. The key mechanism for re-formation of dust is prolonged grain growth on gas-phase metals, it is effective within $\sim100$ Myr after the quenching event, and rapidly increases the dust-to-gas mass ratio in QGs above the standard values ($δ_{\rm DGR}\gtrsim1/100$). As a result, despite heavily depleted cold gas reservoirs, roughly half of QGs maintain little evolution in their ISM dust with stellar age within the first 2 Gyr following the quenching. Overall, we predict that relatively dusty QGs ($M_{\rm dust}/M_{\star}\gtrsim10^{-3}-10^{-4}$) arise from both fast and slow quenchers, and are prevalent in systems of intermediate and low stellar masses ($9<\log(M_{\star}/M_{\odot})<10.5$). This prediction poses an immediate quest for observational synergy between e.g., James Webb Space Telescope (JWST) and the Atacama Large Millimeter Array (ALMA). △ Less

Submitted 16 April, 2024; originally announced April 2024.

Comments: 21 pages, 10 figures, 1 table, submitted to A&A

The JWST-PRIMAL Legacy Survey. A JWST/NIRSpec reference sample for the physical properties and Lyman-$α$ absorption and emission of $\sim 500$ galaxies at $z=5.5-13.4$

Authors: K. E. Heintz, G. B. Brammer, D. Watson, P. A. Oesch, L. C. Keating, M. J. Hayes, Abdurro'uf, K. Z. Arellano-Córdova, A. C. Carnall, C. R. Christiansen, F. Cullen, R. Davé, P. Dayal, A. Ferrara, K. Finlator, J. P. U. Fynbo, S. R. Flury, V. Gelli, S. Gillman, R. Gottumukkala, K. Gould, T. R. Greve, S. E. Hardin, T. Y. -Y Hsiao, A. Hutter , et al. (23 additional authors not shown)

Abstract: One of the surprising early findings with JWST has been the discovery of a strong "roll-over" or a softening of the absorption edge of Ly$α$ in a large number of galaxies at ($z\gtrsim 6$), in addition to systematic offsets from photometric redshift estimates and fundamental galaxy scaling relations. This has been interpreted as damped Ly$α$ absorption (DLA) wings from high column densities of neu… ▽ More One of the surprising early findings with JWST has been the discovery of a strong "roll-over" or a softening of the absorption edge of Ly$α$ in a large number of galaxies at ($z\gtrsim 6$), in addition to systematic offsets from photometric redshift estimates and fundamental galaxy scaling relations. This has been interpreted as damped Ly$α$ absorption (DLA) wings from high column densities of neutral atomic hydrogen (HI), signifying major gas accretion events in the formation of these galaxies. To explore this new phenomenon systematically, we assemble the JWST/NIRSpec PRImordial gas Mass AssembLy (PRIMAL) legacy survey of 494 galaxies at $z=5.5-13.4$. We characterize this benchmark sample in full and spectroscopically derive the galaxy redshifts, metallicities, star-formation rates, and ultraviolet slopes. We define a new diagnostic, the Ly$α$ damping parameter $D_{\rm Lyα}$ to measure and quantify the Ly$α$ emission strength, HI fraction in the IGM, or local HI column density for each source. The JWST-PRIMAL survey is based on the spectroscopic DAWN JWST Archive (DJA-Spec). All the software, reduced spectra, and spectroscopically derived quantities and catalogs are made publicly available in dedicated repositories. The fraction of strong galaxy DLAs are found to be in the range $65-95\%$ at $z>5.5$. The fraction of strong Ly$α$ emitters (LAEs) is found to increase with decreasing redshift, in qualitative agreement with previous observational results, and are predominantly associated with low-metallicity and UV faint galaxies. By contrast, strong DLAs are observed in galaxies with a variety of intrinsic physical properties. Our results indicate that strong DLAs likely reflect a particular early assembly phase of reionization-era galaxies, at which point they are largely dominated by pristine HI gas accretion. [abridged] △ Less

Submitted 2 April, 2024; originally announced April 2024.

Comments: 20 pages, 18 figures, 3 tables. Submitted to A&A. Comments welcome! All data and catalogs are available through the DAWN JWST Archive (DJA): https://dawn-cph.github.io/dja/ and https://github.com/keheintz/jwst-primal

Nemotron-4 15B Technical Report

Authors: Jupinder Parmar, Shrimai Prabhumoye, Joseph Jennings, Mostofa Patwary, Sandeep Subramanian, Dan Su, Chen Zhu, Deepak Narayanan, Aastha Jhunjhunwala, Ayush Dattagupta, Vibhu Jawa, Jiwei Liu, Ameya Mahabaleshwarkar, Osvald Nitski, Annika Brundyn, James Maki, Miguel Martinez, Jiaxuan You, John Kamalu, Patrick LeGresley, Denys Fridman, Jared Casper, Ashwath Aithal, Oleksii Kuchaiev, Mohammad Shoeybi , et al. (2 additional authors not shown)

Abstract: We introduce Nemotron-4 15B, a 15-billion-parameter large multilingual language model trained on 8 trillion text tokens. Nemotron-4 15B demonstrates strong performance when assessed on English, multilingual, and coding tasks: it outperforms all existing similarly-sized open models on 4 out of 7 downstream evaluation areas and achieves competitive performance to the leading open models in the remai… ▽ More We introduce Nemotron-4 15B, a 15-billion-parameter large multilingual language model trained on 8 trillion text tokens. Nemotron-4 15B demonstrates strong performance when assessed on English, multilingual, and coding tasks: it outperforms all existing similarly-sized open models on 4 out of 7 downstream evaluation areas and achieves competitive performance to the leading open models in the remaining ones. Specifically, Nemotron-4 15B exhibits the best multilingual capabilities of all similarly-sized models, even outperforming models over four times larger and those explicitly specialized for multilingual tasks. △ Less

Submitted 27 February, 2024; v1 submitted 26 February, 2024; originally announced February 2024.

Simba-EoR: Early galaxy formation in the Simba simulation including a new sub-grid interstellar medium model

Authors: Ewan Jones, Britton Smith, Romeel Davé, Desika Narayanan, Qi Li

Abstract: We update the dust model present within the Simba galaxy simulations with a self-consistent framework for the co-evolution of dust and molecular hydrogen populations in the interstellar medium, and use this to explore $z \geq 6$ galaxy evolution. In addition to tracking the evolution of dust and molecular hydrogen abundances, our model fully integrates these species into the Simba simulation, expl… ▽ More We update the dust model present within the Simba galaxy simulations with a self-consistent framework for the co-evolution of dust and molecular hydrogen populations in the interstellar medium, and use this to explore $z \geq 6$ galaxy evolution. In addition to tracking the evolution of dust and molecular hydrogen abundances, our model fully integrates these species into the Simba simulation, explicitly modelling their impact on physical processes such as star formation and cooling through the inclusion of a novel two-phase sub-grid model for interstellar gas. In running two high-resolution simulations down to $z \sim 6$ we find that our Simba-EoR model displays a generally tighter concordance with observational data than fiducial Simba. Additionally we observe that our Simba-EoR models increase star formation activity at early epochs, producing larger dust-to-gas ratios consequently. Finally, we discover a significant population of hot dust at $\sim 100$ K, aligning with contemporaneous observations of high-redshift dusty galaxies, alongside the large $\sim 20$ K population typically identified. △ Less

Submitted 9 February, 2024; originally announced February 2024.

Comments: 19 pages, 12 figures, submitted to MNRAS, comments welcome

The Case for Co-Designing Model Architectures with Hardware

Authors: Quentin Anthony, Jacob Hatef, Deepak Narayanan, Stella Biderman, Stas Bekman, Junqi Yin, Aamir Shafi, Hari Subramoni, Dhabaleswar Panda

Abstract: While GPUs are responsible for training the vast majority of state-of-the-art deep learning models, the implications of their architecture are often overlooked when designing new deep learning (DL) models. As a consequence, modifying a DL model to be more amenable to the target hardware can significantly improve the runtime performance of DL training and inference. In this paper, we provide a set… ▽ More While GPUs are responsible for training the vast majority of state-of-the-art deep learning models, the implications of their architecture are often overlooked when designing new deep learning (DL) models. As a consequence, modifying a DL model to be more amenable to the target hardware can significantly improve the runtime performance of DL training and inference. In this paper, we provide a set of guidelines for users to maximize the runtime performance of their transformer models. These guidelines have been created by carefully considering the impact of various model hyperparameters controlling model shape on the efficiency of the underlying computation kernels executed on the GPU. We find the throughput of models with efficient model shapes is up to 39\% higher while preserving accuracy compared to models with a similar number of parameters but with unoptimized shapes. △ Less

Submitted 30 January, 2024; v1 submitted 25 January, 2024; originally announced January 2024.

TEMPLATES: Direct Abundance Constraints for Two Lensed Lyman-Break Galaxies

Authors: Brian Welch, Grace M. Olivier, Taylor A. Hutchison, Jane R. Rigby, Danielle A. Berg, Manuel Aravena, Matthew B. Bayliss, Jack E. Birkin, Scott C. Chapman, Håkon Dahle, Gourav Khullar, Keunho J. Kim, Guillaume Mahler, Matthew A. Malkan, Desika Narayanan, Kedar A. Phadke, Keren Sharon, J. D. T. Smith, Manuel Solimano, Justin S. Spilker, Joaquin D. Viera, David Vizgan

Abstract: Using integrated spectra for two gravitationally lensed galaxies from the JWST TEMPLATES Early Release Science program, we analyze faint auroral lines, which provide direct measurements of the gas-phase chemical abundance. For the brighter galaxy, SGAS1723$+$34 ($z = 1.3293$), we detect the [OIII]$\lambda4363$, [SIII]$\lambda6312$, and [OII]$λλ$7320,7330 auroral emission lines, and set an upper li… ▽ More Using integrated spectra for two gravitationally lensed galaxies from the JWST TEMPLATES Early Release Science program, we analyze faint auroral lines, which provide direct measurements of the gas-phase chemical abundance. For the brighter galaxy, SGAS1723$+$34 ($z = 1.3293$), we detect the [OIII]$\lambda4363$, [SIII]$\lambda6312$, and [OII]$λλ$7320,7330 auroral emission lines, and set an upper limit for the [NII]$\lambda5755$ line. For the second galaxy, SGAS1226$+$21 ($z = 2.925$), we do not detect any auroral lines, and report upper limits. With these measurements and upper limits, we constrain the electron temperatures in different ionization zones within both of these galaxies. For SGAS1723$+$34, where auroral lines are detected, we calculate direct oxygen and nitrogen abundances, finding an N/O ratio consistent with observations of nearby ($z\sim 0$) galaxies. These observations highlight the potent combination of JWST and gravitational lensing to measure faint emission lines in individual distant galaxies and to directly study the chemical abundance patterns in those galaxies. △ Less

Submitted 23 January, 2024; originally announced January 2024.

Comments: Submitted to the Astrophysical Journal's Focus Issue on the TEMPLATES JWST Early Release Science Program. 15 pages, 6 figures. Comments welcome

Tracing the History of Obscured Star Formation with Cosmological Galaxy Evolution Simulations

Authors: Dhruv T. Zimmerman, Desika Narayanan, Katherine E. Whitaker, Romeel Davè

Abstract: We explore the cosmic evolution of the fraction of dust obscured star formation predicted by the \textsc{simba} cosmological hydrodynamic simulations featuring an on-the-fly model for dust formation, evolution, and destruction. We find that up to $z=2$, our results are broadly consistent with previous observational results of little to no evolution in obscured star formation. However, at $z>2$ we… ▽ More We explore the cosmic evolution of the fraction of dust obscured star formation predicted by the \textsc{simba} cosmological hydrodynamic simulations featuring an on-the-fly model for dust formation, evolution, and destruction. We find that up to $z=2$, our results are broadly consistent with previous observational results of little to no evolution in obscured star formation. However, at $z>2$ we find strong evolution at fixed galaxy stellar mass towards greater amounts of obscured star formation. We explain the trend of increasing obscuration at higher redshifts by greater typical dust column densities along the line of sight to young stars. We additionally see that at a fixed redshift, more massive galaxies have a higher fraction of their star formation obscured, which is explained by increased dust mass fractions at higher stellar masses. Finally, we estimate the contribution of dust-obscured star formation to the total star formation rate budget and find that the dust obscured star formation history (SFH) peaks around $z\sim 2-3$, and becomes subdominant at $z\gtrsim 5$. △ Less

Submitted 12 January, 2024; originally announced January 2024.

Comments: 15 pages, 9 figures, submitted to ApJ; comments welcome

JWST Early Release Science Program TEMPLATES: Targeting Extremely Magnified Panchromatic Lensed Arcs and their Extended Star formation

Authors: Jane R. Rigby, Joaquin D. Vieira, Kedar A. Phadke, Taylor A. Hutchison, Brian Welch, Jared Cathey, Justin S. Spilker, Anthony H. Gonzalez, Prasanna Adhikari, M. Aravena, Matthew B. Bayliss, Jack E. Birkin, Emmy Bursk, Scott C. Chapman, Håkon Dahle, Lauren A. Elicker, Travis C. Fischer, Michael K. Florian, Michael D. Gladders, Christopher C. Hayward, Rose Hewald, Lily A. Kettler, Gourav Khullar, Seonwoo Kim, David R. Law , et al. (11 additional authors not shown)

Abstract: This paper gives an overview of TEMPLATES, a JWST Early Release Science program that targeted four extremely bright, gravitationally lensed galaxies: two extremely dusty, two with low attenuation, as templates for galaxy evolution studies with JWST. TEMPLATES obtains a common set of spectral diagnostics for these 1.3 < z < 4.2 galaxies, in particular H alpha, Paschen alpha, and the rest-frame opti… ▽ More This paper gives an overview of TEMPLATES, a JWST Early Release Science program that targeted four extremely bright, gravitationally lensed galaxies: two extremely dusty, two with low attenuation, as templates for galaxy evolution studies with JWST. TEMPLATES obtains a common set of spectral diagnostics for these 1.3 < z < 4.2 galaxies, in particular H alpha, Paschen alpha, and the rest-frame optical and near-infrared continua. In addition, two of the four targets have JWST coverage of [O III] 5007 Angstrom and H beta; the other two targets have have JWST coverage of PAH 3.3 micron and complementary ALMA data covering the [C II] 158 micron emission line. The science goals of TEMPLATES are to demonstrate attenuation-robust diagnostics of star formation, map the distribution of star formation, compare the young and old stellar populations, and measure the physical conditions of star formation and their spatial variation across the galaxies. In addition, TEMPLATES has technical goals to establish best practices for the Integral Field Units (IFU) within the NIRSpec and MIRI instruments, both in terms of observing strategy and in terms of data reduction. The paper describes TEMPLATES's observing program, scientific and technical goals, data reduction methods, and deliverables, including high-level data products and data reduction cookbooks. △ Less

Submitted 16 December, 2023; originally announced December 2023.

Comments: 28 pages, including 8 figures and 3 tables. Submitted to The Astrophysical Journal's Focus Issue on the TEMPLATES JWST Early Release Science Program, https://iopscience.iop.org/collections/apj-231204-01-templates

Packrat: Automatic Reconfiguration for Latency Minimization in CPU-based DNN Serving

Authors: Ankit Bhardwaj, Amar Phanishayee, Deepak Narayanan, Mihail Tarta, Ryan Stutsman

Abstract: In this paper, we investigate how to push the performance limits of serving Deep Neural Network (DNN) models on CPU-based servers. Specifically, we observe that while intra-operator parallelism across multiple threads is an effective way to reduce inference latency, it provides diminishing returns. Our primary insight is that instead of running a single instance of a model with all available threa… ▽ More In this paper, we investigate how to push the performance limits of serving Deep Neural Network (DNN) models on CPU-based servers. Specifically, we observe that while intra-operator parallelism across multiple threads is an effective way to reduce inference latency, it provides diminishing returns. Our primary insight is that instead of running a single instance of a model with all available threads on a server, running multiple instances each with smaller batch sizes and fewer threads for intra-op parallelism can provide lower inference latency. However, the right configuration is hard to determine manually since it is workload- (DNN model and batch size used by the serving system) and deployment-dependent (number of CPU cores on server). We present Packrat, a new serving system for online inference that given a model and batch size ($B$) algorithmically picks the optimal number of instances ($i$), the number of threads each should be allocated ($t$), and the batch sizes each should operate on ($b$) that minimizes latency. Packrat is built as an extension to TorchServe and supports online reconfigurations to avoid serving downtime. Averaged across a range of batch sizes, Packrat improves inference latency by 1.43$\times$ to 1.83$\times$ on a range of commonly used DNNs. △ Less

Submitted 29 November, 2023; originally announced November 2023.

The ALMA-ALPINE [CII] survey: Kennicutt-Schmidt relation in four massive main-sequence galaxies at z~4.5

Authors: M. Béthermin, C. Accard, C. Guillaume, M. Dessauges-Zavadsky, E. Ibar, P. Cassata, T. Devereaux, A. Faisst, J. Freundlich, G. C. Jones, K. Kraljic, H. Algera, R. O. Amorin, S. Bardelli, M. Boquien, V. Buat, E. Donghia, Y. Dubois, A. Ferrara, Y. Fudamoto, M. Ginolfi, P. Guillard, M. Giavalisco, C. Gruppioni, G. Gururajan , et al. (18 additional authors not shown)

Abstract: The Kennicutt-Schmidt (KS) relation between the gas and the star formation rate (SFR) surface density ($Σ_{\rm gas}$-$Σ_{\rm SFR}$) is essential to understand star formation processes in galaxies. So far, it has been measured up to z~2.5 in main-sequence galaxies. In this letter, we aim to put constraints at z~4.5 using a sample of four massive main-sequence galaxies observed by ALMA at high resol… ▽ More The Kennicutt-Schmidt (KS) relation between the gas and the star formation rate (SFR) surface density ($Σ_{\rm gas}$-$Σ_{\rm SFR}$) is essential to understand star formation processes in galaxies. So far, it has been measured up to z~2.5 in main-sequence galaxies. In this letter, we aim to put constraints at z~4.5 using a sample of four massive main-sequence galaxies observed by ALMA at high resolution. We obtained ~0.3"-resolution [CII] and continuum maps of our objects, which we then converted into gas and obscured SFR surface density maps. In addition, we produced unobscured SFR surface density maps by convolving Hubble ancillary data in the rest-frame UV. We then derived the average $Σ_{\rm SFR}$ in various $Σ_{\rm gas}$ bins, and estimated the uncertainties using a Monte Carlo sampling. Our galaxy sample follows the KS relation measured in main-sequence galaxies at lower redshift and is slightly lower than predictions from simulations. Our data points probe the high end both in terms of $Σ_{\rm gas}$ and $Σ_{\rm gas}$, and gas depletion timescales (285-843 Myr) remain similar to z~2 objects. However, three of our objects are clearly morphologically disturbed, and we could have expected shorter gas depletion timescales (~100 Myr) similar to merger-driven starbursts at lower redshifts. This suggests that the mechanisms triggering starbursts at high redshift may be different than in the low- and intermediate-z Universe. △ Less

Submitted 17 November, 2023; v1 submitted 14 November, 2023; originally announced November 2023.

Comments: 9 pages, 7 figures, 2 tables, accepted by A&A (letter)

Holistic Evaluation of Text-To-Image Models

Authors: Tony Lee, Michihiro Yasunaga, Chenlin Meng, Yifan Mai, Joon Sung Park, Agrim Gupta, Yunzhi Zhang, Deepak Narayanan, Hannah Benita Teufel, Marco Bellagente, Minguk Kang, Taesung Park, Jure Leskovec, Jun-Yan Zhu, Li Fei-Fei, Jiajun Wu, Stefano Ermon, Percy Liang

Abstract: The stunning qualitative improvement of recent text-to-image models has led to their widespread attention and adoption. However, we lack a comprehensive quantitative understanding of their capabilities and risks. To fill this gap, we introduce a new benchmark, Holistic Evaluation of Text-to-Image Models (HEIM). Whereas previous evaluations focus mostly on text-image alignment and image quality, we… ▽ More The stunning qualitative improvement of recent text-to-image models has led to their widespread attention and adoption. However, we lack a comprehensive quantitative understanding of their capabilities and risks. To fill this gap, we introduce a new benchmark, Holistic Evaluation of Text-to-Image Models (HEIM). Whereas previous evaluations focus mostly on text-image alignment and image quality, we identify 12 aspects, including text-image alignment, image quality, aesthetics, originality, reasoning, knowledge, bias, toxicity, fairness, robustness, multilinguality, and efficiency. We curate 62 scenarios encompassing these aspects and evaluate 26 state-of-the-art text-to-image models on this benchmark. Our results reveal that no single model excels in all aspects, with different models demonstrating different strengths. We release the generated images and human evaluation results for full transparency at https://crfm.stanford.edu/heim/v1.1.0 and the code at https://github.com/stanford-crfm/helm, which is integrated with the HELM codebase. △ Less

Submitted 7 November, 2023; originally announced November 2023.

Comments: NeurIPS 2023. First three authors contributed equally

$\texttt{slick}$: Modeling a Universe of Molecular Line Luminosities in Hydrodynamical Simulations

Authors: Karolina Garcia, Desika Narayanan, Gergö Popping, R. Anirudh, Sagan Sutherland, Melanie Kaasinen

Abstract: We present {\sc slick} (the Scalable Line Intensity Computation Kit), a software package that calculates realistic CO, [\ion{C}{1}], and [\ion{C}{2}] luminosities for clouds and galaxies formed in hydrodynamic simulations. Built on the radiative transfer code {\sc despotic}, {\sc slick} computes the thermal, radiative, and statistical equilibrium in concentric zones of model clouds, based on their… ▽ More We present {\sc slick} (the Scalable Line Intensity Computation Kit), a software package that calculates realistic CO, [\ion{C}{1}], and [\ion{C}{2}] luminosities for clouds and galaxies formed in hydrodynamic simulations. Built on the radiative transfer code {\sc despotic}, {\sc slick} computes the thermal, radiative, and statistical equilibrium in concentric zones of model clouds, based on their physical properties and individual environments. We validate our results applying {\sc slick} to the high-resolution run of the {\sc Simba} simulations, testing the derived luminosities against empirical and theoretical/analytic relations. To simulate the line emission from a universe of emitting clouds, we have incorporated random forest machine learning (ML) methods into our approach, allowing us to predict cosmologically evolving properties of CO, [\ion{C}{1}] and [\ion{C}{2}] emission from galaxies such as luminosity functions. We tested this model in 100,000 gas particles, and 2,500 galaxies, reaching an average accuracy of $\sim$99.8\% for all lines. Finally, we present the first model light cones created with realistic and ML-predicted CO, [\ion{C}{1}], and [\ion{C}{2}] luminosities in cosmological hydrodynamical simulations, from $z=0$ to $z=10$. △ Less

Submitted 2 November, 2023; originally announced November 2023.

Comments: 17 pages, 11 figures, comments are welcome

Theoretical strong line metallicity diagnostics for the JWST era

Authors: Prerak Garg, Desika Narayanan, Ryan L. Sanders, Romeel Davè, Gergö Popping, Alice E. Shapley, Daniel P. Stark, Jonathan R. Trump

Abstract: The ratios of strong rest-frame optical emission lines are the dominant indicator of metallicities in high-redshift galaxies. Since typical strong-line based metallicity indicators are calibrated on auroral lines at $z=0$, their applicability for galaxies in the distant Universe is unclear. In this paper, we make use of mock emission line data from cosmological simulations to investigate the calib… ▽ More The ratios of strong rest-frame optical emission lines are the dominant indicator of metallicities in high-redshift galaxies. Since typical strong-line based metallicity indicators are calibrated on auroral lines at $z=0$, their applicability for galaxies in the distant Universe is unclear. In this paper, we make use of mock emission line data from cosmological simulations to investigate the calibration of rest-frame optical emission lines as metallicity indicators at high redshift. Our model, which couples the SIMBA cosmological galaxy formation simulation with cloudy photoionization calculations, includes contributions from HII regions, post-AGB stars and Diffuse Ionized Gas (DIG). We find mild redshift evolution in the 12 indicators that we study, which implies that the dominant physical properties that evolve in our simulations do have a discernible impact on the metallicity calibrations at high redshifts. When comparing our calibrations with high redshift auroral line observations from James Webb Space Telescope we find a slight offset between our model results and the observations and find that a higher ionization parameter at high redshifts can be one of the possible explanations. We explore the physics that drives the shapes of strong-line metallicity relationships and propose calibrations for hitherto unexplored low-metallicity regimes. Finally, we study the contribution of DIG to total line fluxes. We find that the contribution of DIG increases with metallicity at z $\sim$ 0 for singly ionized oxygen and sulfur lines and can be as high as 70% making it crucial to include their contribution when modeling nebular emission. △ Less

Submitted 8 November, 2023; v1 submitted 12 October, 2023; originally announced October 2023.

Comments: 28 pages, 6 figures, typo corrected: in polynomial fits, x represents metallicity and y represents the line ratio

GRID: A Platform for General Robot Intelligence Development

Authors: Sai Vemprala, Shuhang Chen, Abhinav Shukla, Dinesh Narayanan, Ashish Kapoor

Abstract: Developing machine intelligence abilities in robots and autonomous systems is an expensive and time consuming process. Existing solutions are tailored to specific applications and are harder to generalize. Furthermore, scarcity of training data adds a layer of complexity in deploying deep machine learning models. We present a new platform for General Robot Intelligence Development (GRID) to addres… ▽ More Developing machine intelligence abilities in robots and autonomous systems is an expensive and time consuming process. Existing solutions are tailored to specific applications and are harder to generalize. Furthermore, scarcity of training data adds a layer of complexity in deploying deep machine learning models. We present a new platform for General Robot Intelligence Development (GRID) to address both of these issues. The platform enables robots to learn, compose and adapt skills to their physical capabilities, environmental constraints and goals. The platform addresses AI problems in robotics via foundation models that know the physical world. GRID is designed from the ground up to be extensible to accommodate new types of robots, vehicles, hardware platforms and software protocols. In addition, the modular design enables various deep ML components and existing foundation models to be easily usable in a wider variety of robot-centric problems. We demonstrate the platform in various aerial robotics scenarios and demonstrate how the platform dramatically accelerates development of machine intelligent robots. △ Less

Submitted 7 October, 2023; v1 submitted 2 October, 2023; originally announced October 2023.

Identification of a transition from stochastic to secular star formation around $z=9$ with JWST

Authors: L. Ciesla, D. Elbaz, O. Ilbert, V. Buat, B. Magnelli, D. Narayanan, E. Daddi, C. Gómez-Guijarro, R. Arango-Toro

Abstract: Star formation histories (SFH) of early (6$<z<$12) galaxies have been found to be highly stochastic in both simulations and observations, while at $z\lesssim$6 the presence of a main sequence (MS) of star-forming galaxies imply secular processes at play. In this work, we aim at characterising the SFH variability of early galaxies as a function of their stellar mass and redshift. We use the JADES p… ▽ More Star formation histories (SFH) of early (6$<z<$12) galaxies have been found to be highly stochastic in both simulations and observations, while at $z\lesssim$6 the presence of a main sequence (MS) of star-forming galaxies imply secular processes at play. In this work, we aim at characterising the SFH variability of early galaxies as a function of their stellar mass and redshift. We use the JADES public catalogue and derive the physical properties of the galaxies as well as their SFH using the spectral energy distribution modelling code CIGALE. To this aim, we implement a non-parametric SFH with a flat prior allowing for as much stochasticity as possible. We use the SFR gradient, an indicator of the movement of galaxies on the SFR-$M_\ast$ plane, linked to the recent SFH of galaxies. This dynamical approach of the relation between the SFR and stellar mass allows us to show that, at $z>9$, 87% of massive galaxies, ($\log(M_\ast/M_\odot)\gtrsim$9), have SFR gradients consistent with a stochastic star-formation activity during the last 100 Myr, while this fraction drops to 15% at $z<7$. On the other hand, we see an increasing fraction of galaxies with a star-formation activity following a common stream on the SFR-$M_\ast$ plane with cosmic time, indicating that a secular mode of star-formation is emerging. We place our results in the context of the observed excess of UV emission as probed by the UV luminosity function at $z\gtrsim10$, by estimating $σ_{UV}$, the dispersion of the UV absolute magnitude distribution, to be of the order of 1.2mag and compare it with predictions from the literature. In conclusion, we find a transition of star-formation mode happening around $z\sim9$: Galaxies with stochastic SFHs dominates at $z\gtrsim9$, although this level of stochasticity is too low to reach those invoked by recent models to reproduce the observed UV luminosity function. △ Less

Submitted 27 September, 2023; originally announced September 2023.

Comments: Submitted to A&A

Journal ref: A&A 686, A128 (2024)

Star formation efficiency across large-scale galactic environments

Authors: Laya Ghodsi, Allison Man, Darko Donevski, Romeel Davé, Seunghwan Lim, Christopher C. Lovell, Desika Narayanan

Abstract: Environmental effects on the evolution of galaxies have been one of the leading questions in galaxy studies for decades. In this work, we investigate the relationship between the star formation activity of galaxies and their environmental matter density using the cosmological hydrodynamic simulation Simba. The star formation activity indicators we explore include the star formation efficiency (SFE… ▽ More Environmental effects on the evolution of galaxies have been one of the leading questions in galaxy studies for decades. In this work, we investigate the relationship between the star formation activity of galaxies and their environmental matter density using the cosmological hydrodynamic simulation Simba. The star formation activity indicators we explore include the star formation efficiency (SFE), specific star formation rate (sSFR) and molecular hydrogen mass fraction ($f^*_{H_2}$) and the environment is considered as the large-scale environmental matter density, calculated based on the stellar mass of nearby galaxies on a 1 Mpc/h grid using the cloud in cell (CIC) method. Our sample includes galaxies with $9<\log(M_*/M_{\odot})$ at $0<z<4$, divided into three mass bins to disentangle the effects of mass and environment on the galactic star formation activity. For low- to intermediate-mass galaxies at low-redshifts ($z<1.5$), we find that the star formation efficiency of those in high-density regions are $\sim 0.3$ dex lower than those in low-density regions. However, there is no significant environmental dependence of the star formation efficiency for massive galaxies over all our redshift range, and low- to intermediate-mass galaxies at high redshifts ($z > 1.5$). We present a scaling relation for the depletion time of molecular hydrogen (${t_{depl}}=1/SFE$) as a function of galaxy parameters including environmental density. Our findings provide a framework for quantifying the environmental effects on the star formation activities of galaxies as a function of stellar mass and redshift. The most significant environmental dependence is seen at later cosmic times ($z<1.5$) and towards lower stellar masses ($9<\log(M_*/M_{\odot})<10$). Future large galaxy surveys can use this framework to look for the environmental dependence of the star formation activity and examine our predictions. △ Less

Submitted 29 February, 2024; v1 submitted 3 September, 2023; originally announced September 2023.

Comments: 17 pages, 6 figures

Gauging the mass of metals in the gas phase of galaxies from the Local Universe to the Epoch of Reionization

Authors: K. E. Heintz, A. E. Shapley, R. L. Sanders, M. Killi, D. Watson, G. Magdis, F. Valentino, M. Ginolfi, D. Narayanan, T. R. Greve, J. P. U. Fynbo, D. Vizgan, S. N. Wilson

Abstract: The chemical enrichment of dust and metals are vital processes in constraining the star formation history of the universe. Previously, the dust masses of high-redshift star-forming galaxies have been determined through their far-infrared continuum, however, equivalent, and potentially simpler, approaches to determining the metal masses have yet to be explored at $z\gtrsim 2$. Here, we present a ne… ▽ More The chemical enrichment of dust and metals are vital processes in constraining the star formation history of the universe. Previously, the dust masses of high-redshift star-forming galaxies have been determined through their far-infrared continuum, however, equivalent, and potentially simpler, approaches to determining the metal masses have yet to be explored at $z\gtrsim 2$. Here, we present a new method of inferring the metal mass in the interstellar medium (ISM) of galaxies out to $z\approx 8$, using the far-infrared [CII]$-158μ$m emission line as a proxy. We calibrated the [CII]-to-$M_{\rm Z,ISM}$ conversion factor based on a benchmark observational sample at $z\approx 0$, in addition to gamma-ray burst sightlines at $z>2$ and cosmological hydrodynamical simulations of galaxies at $z\approx 0$ and $z\approx 6$. We found a universal scaling across redshifts of $\log (M_{\rm Z,ISM}/M_\odot) = \log (L_{\rm [CII]}/L_\odot) - 0.45,$ with a 0.4 dex scatter, which is constant over more than two orders of magnitude in metallicity. We applied this scaling to recent surveys for [CII] in galaxies at $z\gtrsim 2$ and determined the fraction of metals retained in the gas-phase ISM, $M_{\rm Z,ISM} / M_\star$, as a function of redshift showing that an increasing fraction of metals reside in the ISM of galaxies at higher redshifts. We place further constraints on the cosmic metal mass density in the ISM ($Ω_{\rm Z,ISM}$) at $z\approx 5$ and $\approx 7$, yielding $Ω_{\rm Z,ISM} = 6.6^{+13}_{-4.3}\times 10^{-7}\,M_\odot\, {\rm Mpc}^{-3}$ ($z\approx 5$) and $Ω_{\rm Z,ISM} = 2.0^{+3.5}_{-1.3}\times 10^{-7}\,M_\odot\, {\rm Mpc}^{-3}$ ($z\approx 7$). These results are consistent with the expected metal yields from the integrated star formation history at the respective redshifts. This suggests that the majority of metals produced at $z\gtrsim 5$ are confined to the ISM of galaxies. △ Less

Submitted 28 August, 2023; originally announced August 2023.

Comments: Accepted in A&A, abstract abridged

The cosmic build-up of dust and metals. Accurate abundances from GRB-selected star-forming galaxies at $1.7 < z < 6.3$

Authors: K. E. Heintz, A. De Cia, C. C. Thöne, J. -K. Krogager, R. M. Yates, S. Vejlgaard, C. Konstantopoulou, J. P. U. Fynbo, D. Watson, D. Narayanan, S. N. Wilson, M. Arabsalmani, S. Campana, V. D'Elia, M. De Pasquale, D. H. Hartmann, L. Izzo, P. Jakobsson, C. Kouveliotou, A. Levan, Q. Li, D. B. Malesani, A. Melandri, B. Milvang-Jensen, P. Møller , et al. (16 additional authors not shown)

Abstract: The chemical enrichment of dust and metals in the interstellar medium (ISM) of galaxies throughout cosmic time is one of the key driving processes of galaxy evolution. Here we study the evolution of the gas-phase metallicities, dust-to-gas (DTG), and dust-to-metal (DTM) ratios of 36 star-forming galaxies at $1.7 < z < 6.3$ probed by gamma-ray bursts (GRBs). We compile all GRB-selected galaxies wit… ▽ More The chemical enrichment of dust and metals in the interstellar medium (ISM) of galaxies throughout cosmic time is one of the key driving processes of galaxy evolution. Here we study the evolution of the gas-phase metallicities, dust-to-gas (DTG), and dust-to-metal (DTM) ratios of 36 star-forming galaxies at $1.7 < z < 6.3$ probed by gamma-ray bursts (GRBs). We compile all GRB-selected galaxies with intermediate (R=7000) to high (R>40,000) resolution spectroscopic data for which at least one refractory (e.g. Fe) and one volatile (e.g. S or Zn) element have been detected at S/N>3. This is to ensure that accurate abundances and dust depletion patterns can be obtained. We first derive the redshift evolution of the dust-corrected, absorption-line based gas-phase metallicity [M/H]$_{\rm tot}$ in these galaxies, for which we determine a linear relation with redshift ${\rm [M/H]_{tot}}(z) = (-0.21\pm 0.04)z -(0.47\pm 0.14)$. We then examine the DTG and DTM ratios as a function of redshift and through three orders of magnitude in metallicity, quantifying the relative dust abundance both through the direct line-of-sight visual extinction $A_V$ and the derived depletion level. We use a novel method to derive the DTG and DTM mass ratios for each GRB sightline, summing up the mass of all the depleted elements in the dust-phase. We find that the DTG and DTM mass ratios are both strongly correlated with the gas-phase metallicity and show a mild evolution with redshift as well. While these results are subject to a variety of caveats related to the physical environments and the narrow pencil-beam sightlines through the ISM probed by the GRBs, they provide strong implications for studies of dust masses to infer the gas and metal content of high-redshift galaxies, and particularly demonstrate the large offset from the average Galactic value in the low-metallicity, high-redshift regime. △ Less

Submitted 28 August, 2023; originally announced August 2023.

Comments: Accepted in A&A

Journal ref: A&A 679, A91 (2023)

JWST's TEMPLATES for Star Formation: The First Resolved Gas-Phase Metallicity Maps of Dust-Obscured Star-Forming Galaxies at $z$ $\sim$ 4

Authors: Jack E. Birkin, Taylor A. Hutchison, Brian Welch, Justin S. Spilker, Manuel Aravena, Matthew B. Bayliss, Jared Cathey, Scott C. Chapman, Anthony H. Gonzalez, Gayathri Gururajan, Christopher C. Hayward, Gourav Khullar, Keunho J. Kim, Guillaume Mahler, Matthew A. Malkan, Desika Narayanan, Grace M. Olivier, Kedar A. Phadke, Cassie Reuter, Jane R. Rigby, J. D. T. Smith, Manuel Solimano, Nikolaus Sulzenauer, Joaquin D. Vieira, David Vizgan , et al. (1 additional authors not shown)

Abstract: We present the first spatially resolved maps of gas-phase metallicity for two dust-obscured star-forming galaxies (DSFGs) at $z\sim$ 4, from the JWST TEMPLATES Early Release Science program, derived from NIRSpec integral field unit spectroscopy of the H$α$ and [NII] emission lines. Empirical optical line calibrations are used to determine that the sources are globally enriched to near-solar levels… ▽ More We present the first spatially resolved maps of gas-phase metallicity for two dust-obscured star-forming galaxies (DSFGs) at $z\sim$ 4, from the JWST TEMPLATES Early Release Science program, derived from NIRSpec integral field unit spectroscopy of the H$α$ and [NII] emission lines. Empirical optical line calibrations are used to determine that the sources are globally enriched to near-solar levels. While one source shows elevated [NII]/H$α$ ratios and broad H$α$ emission consistent with the presence of an AGN in a $\gtrsim$1kpc region, we argue that both systems have already undergone significant metal enrichment as a result of their extremely high star formation rates. Utilizing ALMA rest-frame 380$μ$m continuum and [CI]($^3$P$_2$-$^3$P$_1$) line maps we compare the spatial variation of the metallicity and gas-to-dust ratio in the two galaxies, finding the two properties to be anticorrelated on highly resolved spatial scales, consistent with various literature studies of $z\sim$ 0 galaxies. The data are indicative of the enormous potential of JWST to probe the enrichment of the interstellar medium on $\sim$kpc scales in extremely dust-obscured systems at $z\sim$ 4 and beyond. △ Less

Submitted 28 August, 2023; v1 submitted 19 July, 2023; originally announced July 2023.

Comments: 14 pages, 6 figures, accepted for publication in ApJ

TEMPLATES: Characterization of a Merger in the Dusty Lensing SPT0418-47 System

Authors: Jared Cathey, Anthony H. Gonzalez, Sidney Lower, Kedar A. Phadke, Justin Spilker, Manuel Aravena, Jack E. Birkin, Simon Birrer, Scott Chapman, Håkon Dahle, Cristopher C. Hayward, Yashar Hezaveh, Ryley Hill, Taylor A. Hutchison, Guillaume Mahler, Daniel P. Marrone, Desika Narayanan, Alexander Navarre, Cassie Reuter, Jane R. Rigby, Keren Sharon, Manuel Solimano, Nikolaus Sulzenauer, Joaquin Vieira, David Vizgan

Abstract: We present JWST and ALMA results for the lensing system SPT0418-47, which includes a strongly-lensed, dusty star-forming galaxy at redshift z=4.225 and an associated multiply-imaged companion. JWST NIRCam and MIRI imaging observations presented in this paper were acquired as part of the Early Release Science program Targeting Extremely Magnified Panchromatic Lensed Arcs and Their Extended Star For… ▽ More We present JWST and ALMA results for the lensing system SPT0418-47, which includes a strongly-lensed, dusty star-forming galaxy at redshift z=4.225 and an associated multiply-imaged companion. JWST NIRCam and MIRI imaging observations presented in this paper were acquired as part of the Early Release Science program Targeting Extremely Magnified Panchromatic Lensed Arcs and Their Extended Star Formation (TEMPLATES). This data set provides robust, mutiwavelength detection of stellar light in both the main (SPT0418A) and companion (SPT0418B) galaxies, while the ALMA detection of [C II] emission confirms that SPT0418B lies at the same redshift as SPT0418A. From a source plane reconstruction, we infer that the projected physical separation of the two galaxies is $4.42\pm 0.05$ kpc. We derive total magnifications of $μ=29.5\pm1.2$ and $μ=4.2\pm 0.9$ for SPT0418A and SPT0418B, respectively. We use both CIGALE and PROSPECTOR to derive stellar masses. The stellar mass ratio of SPT0418A and SPT0418B is approximately 4 to 1 ($4.5\pm 1.0$ for CIGALE and $4.2^{+1.9}_{-1.6}$ for PROSPECTOR). We also see evidence of extended structure associated with SPT0418A in the lensing reconstruction that is suggestive of a tidal feature. Interestingly, the star formation rates and stellar masses of both galaxies are consistent with the main sequence of star-forming galaxies at this epoch, indicating that this ongoing interaction has not noticeably elevated the star formation levels. △ Less

Submitted 19 July, 2023; originally announced July 2023.

Comments: Submitted to AAS Journals,14 pages, 7 figures

MGit: A Model Versioning and Management System

Authors: Wei Hao, Daniel Mendoza, Rafael da Silva, Deepak Narayanan, Amar Phanishaye

Abstract: Models derived from other models are extremely common in machine learning (ML) today. For example, transfer learning is used to create task-specific models from "pre-trained" models through finetuning. This has led to an ecosystem where models are related to each other, sharing structure and often even parameter values. However, it is hard to manage these model derivatives: the storage overhead of… ▽ More Models derived from other models are extremely common in machine learning (ML) today. For example, transfer learning is used to create task-specific models from "pre-trained" models through finetuning. This has led to an ecosystem where models are related to each other, sharing structure and often even parameter values. However, it is hard to manage these model derivatives: the storage overhead of storing all derived models quickly becomes onerous, prompting users to get rid of intermediate models that might be useful for further analysis. Additionally, undesired behaviors in models are hard to track down (e.g., is a bug inherited from an upstream model?). In this paper, we propose a model versioning and management system called MGit that makes it easier to store, test, update, and collaborate on model derivatives. MGit introduces a lineage graph that records provenance and versioning information between models, optimizations to efficiently store model parameters, as well as abstractions over this lineage graph that facilitate relevant testing, updating and collaboration functionality. MGit is able to reduce the lineage graph's storage footprint by up to 7x and automatically update downstream models in response to updates to upstream models. △ Less

Submitted 14 July, 2023; originally announced July 2023.

Outshining by Recent Star Formation Prevents the Accurate Measurement of High-z Galaxy Stellar Masses

Authors: Desika Narayanan, Sidney Lower, Paul Torrey, Gabriel Brammer, Weiguang Cui, Romeel Dave, Kartheik Iyer, Qi Li, Christopher Lovell, Laura Sales, Daniel P. Stark, Federico Marinacci, Mark Vogelsberger

Abstract: In this paper, we demonstrate that the inference of galaxy stellar masses via spectral energy distribution (SED) fitting techniques for galaxies formed in the first billion years after the Big Bang carries fundamental uncertainties owing to the loss of star formation history (SFH) information from the very first episodes of star formation in the integrated spectra of galaxies. While this early sta… ▽ More In this paper, we demonstrate that the inference of galaxy stellar masses via spectral energy distribution (SED) fitting techniques for galaxies formed in the first billion years after the Big Bang carries fundamental uncertainties owing to the loss of star formation history (SFH) information from the very first episodes of star formation in the integrated spectra of galaxies. While this early star formation can contribute substantially to the total stellar mass of high-redshift systems, ongoing star formation at the time of detection outshines the residual light from earlier bursts, hampering the determination of accurate stellar masses. As a result, order of magnitude uncertainties in stellar masses can be expected. We demonstrate this potential problem via direct numerical simulation of galaxy formation in a cosmological context. In detail, we carry out two cosmological simulations with significantly different stellar feedback models which span a significant range in star formation history burstiness. We compute the mock SEDs for these model galaxies at z=7 via 3D dust radiative transfer calculations, and then backwards fit these SEDs with Prospector SED fitting software. The uncertainties in derived stellar masses that we find for z>7 galaxies motivate the development of new techniques and/or star formation history priors to model early Universe star formation. △ Less

Submitted 1 November, 2023; v1 submitted 16 June, 2023; originally announced June 2023.

Comments: ApJ Accepted

Cosmic Sands II: Challenges in Predicting and Measuring High-z Dust Temperatures

Authors: Sidney Lower, Desika Narayanan, Chia-Yu Hu, George C. Privon

Abstract: In the current era of high-z galaxy discovery with JWST and ALMA, our ability to study the stellar populations and ISM conditions in a diverse range of galaxies at Cosmic Dawn has rapidly improved. At the same time, the need to understand the current limitations in modeling galaxy formation processes and physical properties in order to interpret these observations is critical. Here, we study the c… ▽ More In the current era of high-z galaxy discovery with JWST and ALMA, our ability to study the stellar populations and ISM conditions in a diverse range of galaxies at Cosmic Dawn has rapidly improved. At the same time, the need to understand the current limitations in modeling galaxy formation processes and physical properties in order to interpret these observations is critical. Here, we study the challenges in modeling galaxy dust temperatures, both in the context of forward modeling galaxy spectral properties from a hydrodynamical simulation and via backwards modeling galaxy physical properties from mock observations of far-infrared dust emission. Using the Simba model for galaxy formation combined with Powderday radiative transfer, we can accurately predict the evolution of dust at high redshift, though several aspects of the model are essentially free parameters (dust composition, sub-resolution dust in star-forming regions) that dull the predictive power of the model dust temperature distributions. We also highlight the uncertainties in the backwards modeling methods, where we find the commonly used models and assumptions to fit FIR SEDs and infer dust temperatures (e.g., single temperature, optically thin modified blackbody) largely fail to capture the complexity of high-z dusty galaxies. We caution that conclusions inferred from both simulations -- limited by resolution and post-processing techniques -- and observations -- limited by sparse data and simplistic model parameterizations -- are susceptible to unique and nuanced uncertainties that can limit the usefulness of current high-z dust measurements. △ Less

Submitted 4 March, 2024; v1 submitted 12 June, 2023; originally announced June 2023.

Comments: 20 pages main text, 12 figures. Accepted to ApJ, newest version

Observations of neutral carbon in 29 high-z lensed dusty star forming galaxies and the comparison of gas mass tracers

Authors: G. Gururajan, M. Béthermin, N. Sulzenauer, P. Theulé, J. S. Spilker, M. Aravena, S. C. Chapman, A. Gonzalez, T. R. Greve, D. Narayanan, C. Reuter, J. D. Vieira, A. Weiss

Abstract: The nature and evolution of high-redshift dusty star-forming galaxies (high-z DSFGs) remain an open question. Their massive gas reservoirs play an important role in driving the intense star-formation rates hosted in these galaxies. We aim to estimate the molecular gas content of high-z DSFGs by using various gas mass tracers such as the [CI], CO, [CII] emission lines and the dust content. These tr… ▽ More The nature and evolution of high-redshift dusty star-forming galaxies (high-z DSFGs) remain an open question. Their massive gas reservoirs play an important role in driving the intense star-formation rates hosted in these galaxies. We aim to estimate the molecular gas content of high-z DSFGs by using various gas mass tracers such as the [CI], CO, [CII] emission lines and the dust content. These tracers need to be well calibrated as they are all limited by uncertainties on factors such as aCO, XCI, aCII and GDR, thereby affecting the determination of the gas mass accurately. The main goal of our work is to check the consistency between the gas mass tracers and cross-calibrate the uncertain factors. We observe the two [CI] line transitions for 29 SPT-SMGs with the ALMA-ACA. Additionally, we also present new APEX observations of [CII] line for 9 of these galaxies. We find a nearly linear relation between the infrared luminosity and [CI] luminosity if we fit the starbursts and main-sequence galaxies separately. We measure a median [CI]-derived excitation temperature of 34.5+/-2.1 K. We probe the properties of the interstellar medium (ISM) such as density and radiation field intensity using [CI] to mid- or high-J CO lines and [CI] to infrared luminosity ratio, and find similar values to the SMG populations in literature. Finally, the gas masses estimated from [CI], CO, dust, and [CII] do not exhibit any significant trend with the infrared luminosity or the dust temperature. We provide the various cross-calibrations between these tracers. Our study confirms that [CI] is a suitable tracer of the molecular gas content, and shows an overall agreement between all the classical gas tracers used at high redshift. However, their absolute calibration and thus the gas depletion timescale measurements remain uncertain. △ Less

Submitted 5 June, 2023; originally announced June 2023.

Comments: Accepted for publication in A&A, 25 pages, 11 figures, 6 tables

Journal ref: A&A 676, A89 (2023)

Spatial variations in aromatic hydrocarbon emission in a dust-rich galaxy

Authors: Justin S. Spilker, Kedar A. Phadke, Manuel Aravena, Melanie Archipley, Matthew B. Bayliss, Jack E. Birkin, Matthieu Bethermin, James Burgoyne, Jared Cathey, Scott C. Chapman, Hakon Dahle, Anthony H. Gonzalez, Gayathri Gururajan, Christopher C. Hayward, Yashar D. Hezaveh, Ryley Hill, Taylor A. Hutchison, Keunho J. Kim, Seonwoo Kim, David Law, Ronan Legin, Matthew A. Malkan, Daniel P. Marrone, Eric J. Murphy, Desika Narayanan , et al. (13 additional authors not shown)

Abstract: Dust grains absorb half of the radiation emitted by stars throughout the history of the universe, re-emitting this energy at infrared wavelengths. Polycyclic aromatic hydrocarbons (PAHs) are large organic molecules that trace millimeter-size dust grains and regulate the cooling of the interstellar gas within galaxies. Observations of PAH features in very distant galaxies have been difficult due to… ▽ More Dust grains absorb half of the radiation emitted by stars throughout the history of the universe, re-emitting this energy at infrared wavelengths. Polycyclic aromatic hydrocarbons (PAHs) are large organic molecules that trace millimeter-size dust grains and regulate the cooling of the interstellar gas within galaxies. Observations of PAH features in very distant galaxies have been difficult due to the limited sensitivity and wavelength coverage of previous infrared telescopes. Here we present JWST observations that detect the 3.3um PAH feature in a galaxy observed less than 1.5 billion years after the Big Bang. The high equivalent width of the PAH feature indicates that star formation, rather than black hole accretion, dominates the infrared emission throughout the galaxy. The light from PAH molecules, large dust grains, and stars and hot dust are spatially distinct from one another, leading to order-of-magnitude variations in the PAH equivalent width and the ratio of PAH to total infrared luminosity across the galaxy. The spatial variations we observe suggest either a physical offset between the PAHs and large dust grains or wide variations in the local ultraviolet radiation field. Our observations demonstrate that differences in the emission from PAH molecules and large dust grains are a complex result of localized processes within early galaxies. △ Less

Submitted 5 June, 2023; originally announced June 2023.

Comments: Published in Nature 5 June 2023 at https://www.nature.com/articles/s41586-023-05998-6. MIRI MRS reduction notebook is available at https://github.com/jwst-templates

Cheaply Evaluating Inference Efficiency Metrics for Autoregressive Transformer APIs

Authors: Deepak Narayanan, Keshav Santhanam, Peter Henderson, Rishi Bommasani, Tony Lee, Percy Liang

Abstract: Large language models (LLMs) power many state-of-the-art systems in natural language processing. However, these models are extremely computationally expensive, even at inference time, raising the natural question: when is the extra cost of deploying a larger model worth the anticipated boost in capabilities? Better understanding this tradeoff fundamentally could benefit from an inference efficienc… ▽ More Large language models (LLMs) power many state-of-the-art systems in natural language processing. However, these models are extremely computationally expensive, even at inference time, raising the natural question: when is the extra cost of deploying a larger model worth the anticipated boost in capabilities? Better understanding this tradeoff fundamentally could benefit from an inference efficiency metric that is both (i) easily comparable across models from different providers, and (ii) representative of the true cost of running queries in an isolated performance environment. Unfortunately, access to LLMs today is largely restricted to black-box text generation APIs and raw runtimes measured through this interface do not satisfy these desiderata: model providers can apply various software and hardware optimizations orthogonal to the model, and models served on shared infrastructure are susceptible to performance contention. To circumvent these problems, we propose a new metric for comparing inference efficiency across models. This metric puts models on equal footing as though they were served (i) on uniform hardware and software, and (ii) without performance contention. We call this metric the \emph{idealized runtime}, and we propose a methodology to efficiently estimate this metric for autoregressive Transformer models. We also propose cost-aware variants that incorporate the number of accelerators needed to serve the model. Using these metrics, we compare ten state-of-the-art LLMs to provide the first analysis of inference efficiency-capability tradeoffs; we make several observations from this analysis, including the fact that the superior inference runtime performance of certain APIs is often a byproduct of optimizations within the API rather than the underlying model. Our methodology also facilitates the efficient comparison of different software and hardware stacks. △ Less

Submitted 3 May, 2023; originally announced May 2023.

In pursuit of giants: II. Evolution of dusty quiescent galaxies over the last six billion years from the hCOSMOS survey

Authors: Darko Donevski, Ivana Damjanov, Ambra Nanni, Allison Man, Marika Giulietti, Michael Romano, Andrea Lapi, Desika Narayanan, Romeel Davé, Irene Shivaei, Jubee Sohn, Junais, Lara Pantoni, Qi Li

Abstract: Quantifying changes in galaxies' interstellar medium (ISM) abundance after quenching star formation is an important aspect of galaxy evolution, but it is poorly constrained beyond the local universe. We characterise the dust-related properties in 548 quiescent galaxies observed at $0.1<z<0.6$ as part of the hCOSMOS spectroscopic survey. This is the largest sample of quiescent galaxies at intermedi… ▽ More Quantifying changes in galaxies' interstellar medium (ISM) abundance after quenching star formation is an important aspect of galaxy evolution, but it is poorly constrained beyond the local universe. We characterise the dust-related properties in 548 quiescent galaxies observed at $0.1<z<0.6$ as part of the hCOSMOS spectroscopic survey. This is the largest sample of quiescent galaxies at intermediate redshifts, for which the co-evolution of dust, metals and stars have been estimated. We reveal the complex relations between the key markers of galaxies' dust life-cycles, such as specific dust mass ($M_{\rm dust}$/$M_{\rm \star}$), with gas-metallicity ($Z_{\rm gas}$), time since quenching ($t_{\rm quench}$), stellar age and size. We find morphology to be important factor of a large scatter ($\sim2$ orders of magnitude) in $M_{\rm dust}/M_{\rm \star}$. Through modelling the star formation histories of our objects, we derive a broad dynamical range of post-quenching timescales ($60\:\rm Myr<t_{\rm quench}<3.2\:\rm Gyr$). We find that $M_{\rm dust}/M_{\rm \star}$ is the highest in recently quenched systems ($t_{\rm quench}<500$ Myr), but its further evolution is non-monotonic as a consequence of diverse pathways for prolonged dust formation, or removal on various timescales. Our data are well reproduced by the SIMBA cosmological simulation and chemical models that include dust growth in the ISM. While this process is prevalent in dusty quiescent galaxies, $\sim15\%$ of objects show signs of external dust acquisition, most likely via minor mergers. Our results strongly suggest that prolonged dust production on a timescale $0.5-1\:\rm Gyr$ since quenching may be common in dusty quiescent galaxies at intermediate redshifts, even if their gas reservoirs are heavily exhausted (i.e., cold gas fraction $<1-5\%$). △ Less

Submitted 12 April, 2023; originally announced April 2023.

Comments: Main text 18 pages, 9 figures; Submitted to A&A; Comments welcome!

Journal ref: A&A 678, A35 (2023)

Polycraft World AI Lab (PAL): An Extensible Platform for Evaluating Artificial Intelligence Agents

Authors: Stephen A. Goss, Robert J. Steininger, Dhruv Narayanan, Daniel V. Olivença, Yutong Sun, Peng Qiu, Jim Amato, Eberhard O. Voit, Walter E. Voit, Eric J. Kildebeck

Abstract: As artificial intelligence research advances, the platforms used to evaluate AI agents need to adapt and grow to continue to challenge them. We present the Polycraft World AI Lab (PAL), a task simulator with an API based on the Minecraft mod Polycraft World. Our platform is built to allow AI agents with different architectures to easily interact with the Minecraft world, train and be evaluated in… ▽ More As artificial intelligence research advances, the platforms used to evaluate AI agents need to adapt and grow to continue to challenge them. We present the Polycraft World AI Lab (PAL), a task simulator with an API based on the Minecraft mod Polycraft World. Our platform is built to allow AI agents with different architectures to easily interact with the Minecraft world, train and be evaluated in multiple tasks. PAL enables the creation of tasks in a flexible manner as well as having the capability to manipulate any aspect of the task during an evaluation. All actions taken by AI agents and external actors (non-player-characters, NPCs) in the open-world environment are logged to streamline evaluation. Here we present two custom tasks on the PAL platform, one focused on multi-step planning and one focused on navigation, and evaluations of agents solving them. In summary, we report a versatile and extensible AI evaluation platform with a low barrier to entry for AI researchers to utilize. △ Less

Submitted 27 January, 2023; originally announced January 2023.

Comments: 27 pages, 5 figures

A Framework for Modeling Polycyclic Aromatic Hydrocarbon Emission in Galaxy Evolution Simulations

Authors: Desika Narayanan, J. D. Smith, Brandon Hensley, Qi Li, Chia-Yu Hu, Karin Sandstrom, Paul Torrey, Mark Vogelsberger, Federico Marinacci, Laura Sales

Abstract: We present a new methodology for simulating mid-infrared emission from polycyclic aromatic hydrocarbons (PAHs) in galaxy evolution simulations. To do this, we combine theoretical models of PAH emission features as they respond to varying interstellar radiation fields, grain size distributions, and ionization states with a new on-the-fly model for dust evolution in hydrodynamic galaxy simulations.… ▽ More We present a new methodology for simulating mid-infrared emission from polycyclic aromatic hydrocarbons (PAHs) in galaxy evolution simulations. To do this, we combine theoretical models of PAH emission features as they respond to varying interstellar radiation fields, grain size distributions, and ionization states with a new on-the-fly model for dust evolution in hydrodynamic galaxy simulations. We apply these models to 3 idealized arepo galaxy evolution simulations within the smuggle physics framework. We use these simulations to develop numerical experiments investigating the buildup of PAH masses and luminosities in galaxies in idealized analogs of the Milky Way, a dwarf galaxy, and starburst disk. Our main results follow. Galaxies with high specific star formation rates have increased feedback energy per unit mass, and are able to efficiently shatter dust grains, driving up the fraction of ultra small grains. At the same time, in our model large radiation fields per unit gas density convert aliphatic grains into aromatics. The fraction of dust grains in the form of PAHs (q_PAH) can be understood as a consequence of these processes, and in our model PAHs form primarily from interstellar processing (shattering) of larger grains rather than from the growth of smaller grains. We find that the hardness of the radiation field plays a larger role than variations in the grain size distribution in setting the total integrated PAH luminosities, though cosmological simulations are necessary to fully investigate the complex interplay of processes that drive PAH band luminosities in galaxies. Finally, we highlight feature PAH strength variations, cautioning against the usage of emission templates with constant feature strength ratios. △ Less

Submitted 17 January, 2023; originally announced January 2023.

Comments: Submitted to ApJ; comments welcome

$\rm [C_{II}]$ 158 $\rm μm$ emission as an indicator of galaxy star formation rate

Authors: Lichen Liang, Robert Feldmann, Norman Murray, Desika Narayanan, Christopher C. Hayward, Daniel Anglés-Alcázar, Luigi Bassini, Alexander J. Richings, Claude-André Faucher-Giguère, Dongwoo T. Chung, Jennifer Y. H. Chan, Doǧa Tolgay, Onur Çatmabacak, Dušan Kereš, Philip F. Hopkins

Abstract: Observations of local star-forming galaxies (SFGs) show a tight correlation between their singly ionized carbon line luminosity ($L_{\rm [C_{II}]}$) and star formation rate (SFR), suggesting that $L_{\rm [C_{II}]}$ may be a useful SFR tracer for galaxies. Some other galaxy populations, however, are found to have lower $L_{\rm [C_{II}]}{}/{}\rm SFR$ than the local SFGs, including the infrared-lumin… ▽ More Observations of local star-forming galaxies (SFGs) show a tight correlation between their singly ionized carbon line luminosity ($L_{\rm [C_{II}]}$) and star formation rate (SFR), suggesting that $L_{\rm [C_{II}]}$ may be a useful SFR tracer for galaxies. Some other galaxy populations, however, are found to have lower $L_{\rm [C_{II}]}{}/{}\rm SFR$ than the local SFGs, including the infrared-luminous, starburst galaxies at low and high redshifts, as well as some moderately star-forming galaxies at the epoch of re-ionization (EoR). The origin of this `$\rm [C_{II}]$ deficit' is unclear. In this work, we study the $L_{\rm [C_{II}]}$-SFR relation of galaxies using a sample of $z=0-8$ galaxies with $M_*\approx10^7-5\times10^{11}\,M_\odot$ extracted from cosmological volume and zoom-in simulations from the Feedback in Realistic Environments (FIRE) project. We find a simple analytic expression for $L_{\rm [C_{II}]}$/SFR of galaxies in terms of the following parameters: mass fraction of $\rm [C_{II}]$-emitting gas ($f_{\rm [C_{II}]}$), gas metallicity ($Z_{\rm gas}$), gas density ($n_{\rm gas}$) and gas depletion time ($t_{\rm dep}{}={}M_{\rm gas}{}/{}\rm SFR$). We find two distinct physical regimes, where $t_{\rm dep}$ ($Z_{\rm gas}$) is the main driver of the $\rm [C_{II}]$ deficit in $\rm H_2$-rich ($\rm H_2$-poor) galaxies. The observed $\rm [C_{II}]$ deficit of IR-luminous galaxies and early EoR galaxies, corresponding to the two different regimes, is due to short gas depletion time and low gas metallicity, respectively. Our result indicates that $\rm [C_{II}]$ deficit is a common phenomenon of galaxies, and caution needs to be taken when applying a constant $L_{\rm [C_{II}]}$-to-SFR conversion factor derived from local SFGs to estimate cosmic SFR density at high redshifts and interpret data from upcoming $\rm [C_{II}]$ line intensity mapping experiments. △ Less

Submitted 6 December, 2023; v1 submitted 10 January, 2023; originally announced January 2023.

Comments: 44 page, 24 figures, 8 tables and 8 appendices, accepted for publication in MNRAS. Key figures: fig. 16 & 17. Comments are welcome

Extended Lyman-$α$ emission towards the SPT2349-56 protocluster at $z=4.3$

Authors: Yordanka Apostolovski, Manuel Aravena, Timo Anguita, Matthieu Bethermin, James Burgoyne, Scott Chapman, Carlos De Breuck, Anthony Gonzalez, Max Gronke, Lucia Guaita, Yashar Hezaveh, Ryley Hill, Sreevani Jarugula, Evelyn Johnston, Matt Malkan, Desika Narayanan, Cassie Reuter, Manuel Solimano, Justin Spilker, Nikolaus Sulzenauer, Joaquin Vieira, David Vizgan, Axel Weiß

Abstract: Context. Deep spectroscopic surveys with the Atacama Large Millimeter/submillimeter Array (ALMA) have revealed that some of the brightest infrared sources in the sky correspond to concentrations of dusty star-forming galaxies (DSFG) at high redshift. Among these, the SPT2349-56 protocluster system at z = 4.304 is amongst the most extreme examples due to its high source density and integrated star… ▽ More Context. Deep spectroscopic surveys with the Atacama Large Millimeter/submillimeter Array (ALMA) have revealed that some of the brightest infrared sources in the sky correspond to concentrations of dusty star-forming galaxies (DSFG) at high redshift. Among these, the SPT2349-56 protocluster system at z = 4.304 is amongst the most extreme examples due to its high source density and integrated star formation rate. Aims. We conducted a deep Lyman-$α$ line emission survey around SPT2349-56 using the Multi-Unit Spectroscopic Explorer (MUSE) at Very Large Telescope (VLT) in order to characterize this uniquely dense environment. Methods. Taking advantage of the deep three-dimensional nature of this survey, we performed a sensitive search for Lyman-$α$ emitters (LAEs) toward the core and northern extension of the protocluster, which correspond to the brightest infrared regions in this field. Using a smoothed narrowband image extracted from the MUSE datacube around the protocluster redshift, we searched for possible extended structures. Results. We identify only three LAEs at z = 4.3 in this field, in concordance with expectations for blank-fields, and an extended Lyman-$α$ structure spatially associated with core of the protocluster. All the previously-identified DSFGs in this field are undetected in Lyman-$α$ emission, consistent with the conspicuous dust obscuration in these systems. We find an extended Lyman-$α$ structure, about $60 \times 60$ kpc$^{2}$ in size, and located 56 kpc west of the protocluster core. Three DSFGs coincide spatially with the location of this structure. We conclude that either the three co-spatial DSFGs or the protocluster core itself are feeding ionizing photons to the Lyman-$α$ structure. △ Less

Submitted 3 January, 2023; originally announced January 2023.

Comments: 14 pages, 9 figures, 2 tables, submitted to A&A

Cosmic Sands: The Origin of Dusty, Star-forming Galaxies in the Epoch of Reionization

Authors: Sidney Lower, Desika Narayanan, Qi Li, Romeel Davé

Abstract: We present the Cosmic Sands suite of cosmological zoom-in simulations based on the Simba galaxy formation model in order to study the build up of the first massive and dusty galaxies in the early Universe. Residing in the most massive halos, we find that the compact proto-massive galaxies undergo nearly continuous mergers with smaller subhalos, boosting star formation rates (SFRs) and the build up… ▽ More We present the Cosmic Sands suite of cosmological zoom-in simulations based on the Simba galaxy formation model in order to study the build up of the first massive and dusty galaxies in the early Universe. Residing in the most massive halos, we find that the compact proto-massive galaxies undergo nearly continuous mergers with smaller subhalos, boosting star formation rates (SFRs) and the build up of stellar mass. The galaxies are already appreciably chemically evolved by z=10, with modeled dust masses comparable to those inferred from observations in the same epoch. We track gas accretion onto the galaxies to understand how extreme SFRs can be sustained by these early systems. We find that smooth gas accretion can maintain SFRs above 250 M$_{\odot}$ / yr but to achieve SFRs that boost galaxies well above the main sequence, a larger perturbation like a gas-rich major merger is necessary to trigger a starburst episode. Post-processing the Cosmic Sands simulations with dust radiative transfer, we find that while the infrared luminosities of the most dust rich galaxies are comparable to local ULIRGs, they are substantially dimmer than classical z=2 sub-millimeter galaxies. We end with a discussion on the possible reasons for this discrepancy at the highest masses and the future work we intend to carry out to study the chemical enrichment of the earliest dusty galaxies. △ Less

Submitted 5 December, 2022; originally announced December 2022.

Comments: Submitted to ApJ, main text 17 pages, 12 figures. Comments welcome!

Merger Signatures are Common, but not Universal, In Massive, Recently-Quenched Galaxies at z~0.7

Authors: Margaret Verrico, David J. Setton, Rachel Bezanson, Jenny E. Greene, Katherine A. Suess, Andy D. Goulding, Justin S. Spilker, Mariska Kriek, Robert Feldmann, Desika Narayanan, Vincenzo Donofrio, Gourav Khullar

Abstract: We present visual classifications of merger-induced tidal disturbances in 143 $\rm{M}_* \sim 10^{11}\rm{M}_\odot$ post-starburst galaxies at z$\sim$0.7 identified in the SQuIGG$\vec{L}$E Sample. This sample spectroscopically selects galaxies from the Sloan Digital Sky Survey that have stopped their primary epoch of star formation within the past $\sim$500 Myrs. Visual classifications are performed… ▽ More We present visual classifications of merger-induced tidal disturbances in 143 $\rm{M}_* \sim 10^{11}\rm{M}_\odot$ post-starburst galaxies at z$\sim$0.7 identified in the SQuIGG$\vec{L}$E Sample. This sample spectroscopically selects galaxies from the Sloan Digital Sky Survey that have stopped their primary epoch of star formation within the past $\sim$500 Myrs. Visual classifications are performed on Hyper Suprime Cam (HSC) i-band imaging. We compare to a control sample of mass- and redshift-matched star-forming and quiescent galaxies from the Large Early Galaxy Census and find that post-starburst galaxies are more likely to be classified as disturbed than either category. This corresponds to a factor of $3.6^{+2.9}_{-1.3}$ times the disturbance rate of older quiescent galaxies and $2.1^{+1.9}_{-.73}$ times the disturbance rate of star-forming galaxies. Assuming tidal features persist for $\lesssim500$ Myr, this suggests merging is coincident with quenching in a significant fraction of these post-starbursts. Galaxies with tidal disturbances are younger on average than undisturbed post-starburst galaxies in our sample, suggesting tidal features from a major merger may have faded over time. This may be exacerbated by the fact that, on average, the undisturbed subset is fainter, rendering low surface brightness tidal features harder to identify. However, the presence of ten young ($\lesssim150$ Myr since quenching) undisturbed galaxies suggests that major mergers are not the only fast physical mechanism that shut down the primary epoch of star formation in massive galaxies at intermediate redshift. △ Less

Submitted 17 March, 2023; v1 submitted 29 November, 2022; originally announced November 2022.

Comments: 13 pages, 8 figures, accepted to ApJ

MegaBlocks: Efficient Sparse Training with Mixture-of-Experts

Authors: Trevor Gale, Deepak Narayanan, Cliff Young, Matei Zaharia

Abstract: We present MegaBlocks, a system for efficient Mixture-of-Experts (MoE) training on GPUs. Our system is motivated by the limitations of current frameworks, which restrict the dynamic routing in MoE layers to satisfy the constraints of existing software and hardware. These formulations force a tradeoff between model quality and hardware efficiency, as users must choose between dropping tokens from t… ▽ More We present MegaBlocks, a system for efficient Mixture-of-Experts (MoE) training on GPUs. Our system is motivated by the limitations of current frameworks, which restrict the dynamic routing in MoE layers to satisfy the constraints of existing software and hardware. These formulations force a tradeoff between model quality and hardware efficiency, as users must choose between dropping tokens from the computation or wasting computation and memory on padding. To address these limitations, we reformulate MoE computation in terms of block-sparse operations and develop new block-sparse GPU kernels that efficiently handle the dynamism present in MoEs. Our approach never drops tokens and maps efficiently to modern hardware, enabling end-to-end training speedups of up to 40% over MoEs trained with the state-of-the-art Tutel library and 2.4x over DNNs trained with the highly-optimized Megatron-LM framework. △ Less

Submitted 28 November, 2022; originally announced November 2022.

Holistic Evaluation of Language Models

Authors: Percy Liang, Rishi Bommasani, Tony Lee, Dimitris Tsipras, Dilara Soylu, Michihiro Yasunaga, Yian Zhang, Deepak Narayanan, Yuhuai Wu, Ananya Kumar, Benjamin Newman, Binhang Yuan, Bobby Yan, Ce Zhang, Christian Cosgrove, Christopher D. Manning, Christopher Ré, Diana Acosta-Navas, Drew A. Hudson, Eric Zelikman, Esin Durmus, Faisal Ladhak, Frieda Rong, Hongyu Ren, Huaxiu Yao , et al. (25 additional authors not shown)

Abstract: Language models (LMs) are becoming the foundation for almost all major language technologies, but their capabilities, limitations, and risks are not well understood. We present Holistic Evaluation of Language Models (HELM) to improve the transparency of language models. First, we taxonomize the vast space of potential scenarios (i.e. use cases) and metrics (i.e. desiderata) that are of interest fo… ▽ More Language models (LMs) are becoming the foundation for almost all major language technologies, but their capabilities, limitations, and risks are not well understood. We present Holistic Evaluation of Language Models (HELM) to improve the transparency of language models. First, we taxonomize the vast space of potential scenarios (i.e. use cases) and metrics (i.e. desiderata) that are of interest for LMs. Then we select a broad subset based on coverage and feasibility, noting what's missing or underrepresented (e.g. question answering for neglected English dialects, metrics for trustworthiness). Second, we adopt a multi-metric approach: We measure 7 metrics (accuracy, calibration, robustness, fairness, bias, toxicity, and efficiency) for each of 16 core scenarios when possible (87.5% of the time). This ensures metrics beyond accuracy don't fall to the wayside, and that trade-offs are clearly exposed. We also perform 7 targeted evaluations, based on 26 targeted scenarios, to analyze specific aspects (e.g. reasoning, disinformation). Third, we conduct a large-scale evaluation of 30 prominent language models (spanning open, limited-access, and closed models) on all 42 scenarios, 21 of which were not previously used in mainstream LM evaluation. Prior to HELM, models on average were evaluated on just 17.9% of the core HELM scenarios, with some prominent models not sharing a single scenario in common. We improve this to 96.0%: now all 30 models have been densely benchmarked on the same core scenarios and metrics under standardized conditions. Our evaluation surfaces 25 top-level findings. For full transparency, we release all raw model prompts and completions publicly for further analysis, as well as a general modular toolkit. We intend for HELM to be a living benchmark for the community, continuously updated with new scenarios, metrics, and models. △ Less

Submitted 1 October, 2023; v1 submitted 16 November, 2022; originally announced November 2022.

Comments: Authored by the Center for Research on Foundation Models (CRFM) at the Stanford Institute for Human-Centered Artificial Intelligence (HAI). Project page: https://crfm.stanford.edu/helm/v1.0

Journal ref: Published in Transactions on Machine Learning Research (TMLR), 2023

BLOOM: A 176B-Parameter Open-Access Multilingual Language Model

Authors: BigScience Workshop, :, Teven Le Scao, Angela Fan, Christopher Akiki, Ellie Pavlick, Suzana Ilić, Daniel Hesslow, Roman Castagné, Alexandra Sasha Luccioni, François Yvon, Matthias Gallé, Jonathan Tow, Alexander M. Rush, Stella Biderman, Albert Webson, Pawan Sasanka Ammanamanchi, Thomas Wang, Benoît Sagot, Niklas Muennighoff, Albert Villanova del Moral, Olatunji Ruwase, Rachel Bawden, Stas Bekman, Angelina McMillan-Major , et al. (369 additional authors not shown)

Abstract: Large language models (LLMs) have been shown to be able to perform new tasks based on a few demonstrations or natural language instructions. While these capabilities have led to widespread adoption, most LLMs are developed by resource-rich organizations and are frequently kept from the public. As a step towards democratizing this powerful technology, we present BLOOM, a 176B-parameter open-access… ▽ More Large language models (LLMs) have been shown to be able to perform new tasks based on a few demonstrations or natural language instructions. While these capabilities have led to widespread adoption, most LLMs are developed by resource-rich organizations and are frequently kept from the public. As a step towards democratizing this powerful technology, we present BLOOM, a 176B-parameter open-access language model designed and built thanks to a collaboration of hundreds of researchers. BLOOM is a decoder-only Transformer language model that was trained on the ROOTS corpus, a dataset comprising hundreds of sources in 46 natural and 13 programming languages (59 in total). We find that BLOOM achieves competitive performance on a wide variety of benchmarks, with stronger results after undergoing multitask prompted finetuning. To facilitate future research and applications using LLMs, we publicly release our models and code under the Responsible AI License. △ Less

Submitted 27 June, 2023; v1 submitted 9 November, 2022; originally announced November 2022.

The interstellar medium distribution, gas kinematics, and system dynamics of the far-infrared luminous quasar SDSS J2310+1855 at $z=6.0$

Authors: Yali Shao, Ran Wang, Axel Weiss, Jeff Wagg, Chris L. Carilli, Michael A. Strauss, Fabian Walter, Pierre Cox, Xiaohui Fan, Karl M. Menten, Desika Narayanan, Dominik Riechers, Frank Bertoldi, Alain Omont, Linhua Jiang

Abstract: We present ALMA sub-kpc- to kpc-scale resolution observations of the [CII], CO(9-8), and OH$^{+}$\,($1_{1}$--$0_{1}$) lines along with their dust continuum emission toward the FIR luminous quasar SDSS J231038.88+185519.7 at $z = 6.0031$. The [CII] brightness follows a flat distribution with a Sersic index of 0.59. The CO(9-8) line and the dust continuum can be fit with an unresolved nuclear compon… ▽ More We present ALMA sub-kpc- to kpc-scale resolution observations of the [CII], CO(9-8), and OH$^{+}$\,($1_{1}$--$0_{1}$) lines along with their dust continuum emission toward the FIR luminous quasar SDSS J231038.88+185519.7 at $z = 6.0031$. The [CII] brightness follows a flat distribution with a Sersic index of 0.59. The CO(9-8) line and the dust continuum can be fit with an unresolved nuclear component and an extended Sersic component with a Sersic index of ~1. The dust temperature drops with distance from the center. The effective radius of the dust continuum is smaller than that of the line emission and the dust mass surface density, but is consistent with that of the star formation rate surface density. The OH$^{+}$\,($1_{1}$--$0_{1}$) line shows a P-Cygni profile with an absorption, which may indicate an outflow with a neutral gas mass of $(6.2\pm1.2)\times10^{8} M_{\odot}$ along the line of sight. We employed a 3D tilted ring model to fit the [CII] and CO(9-8) data cubes. The two lines are both rotation dominated and trace identical disk geometries and gas motions. We decompose the circular rotation curve measured from the kinematic model fit to the [CII] line into four matter components (black hole, stars, gas, and dark matter). The quasar-starburst system is dominated by baryonic matter inside the central few kiloparsecs. We constrain the black hole mass to be $2.97^{+0.51}_{-0.77}\times 10^{9}\,M_{\odot}$; this is the first time that the dynamical mass of a black hole has been measured at $z\sim6$. A massive stellar component (on the order of $10^{9}\,M_{\odot}$) may have already existed when the Universe was only ~0.93 Gyr old. The relations between the black hole mass and the baryonic mass of this quasar indicate that the central supermassive black hole may have formed before its host galaxy. [Abridged version. Please see the full abstract in the manuscript.] △ Less

Submitted 21 October, 2022; originally announced October 2022.

Comments: Accepted for publication in A&A

Journal ref: A&A 668, A121 (2022)

Investigating the [C$\,{\rm \scriptsize II}$]-to-H$\,{\rm \scriptsize I}$ conversion factor and the H$\,{\rm \scriptsize I}$ gas budget of galaxies at $z\approx 6$ with hydrodynamical simulations

Authors: David Vizgan, Kasper E. Heintz, Thomas R. Greve, Desika Narayanan, Romeel Davé, Karen P. Olsen, Gergö Popping, Darach Watson

Abstract: One of the most fundamental baryonic matter components of galaxies is the neutral atomic hydrogen (H$\,{\rm \scriptsize I}$). At low redshifts, this component can be traced directly through the 21-cm transition, but to infer H$\,{\rm \scriptsize I}$ gas content of the most distant galaxies, a viable tracer is needed. We here investigate the fidelity of the fine structure transition of the (… ▽ More One of the most fundamental baryonic matter components of galaxies is the neutral atomic hydrogen (H$\,{\rm \scriptsize I}$). At low redshifts, this component can be traced directly through the 21-cm transition, but to infer H$\,{\rm \scriptsize I}$ gas content of the most distant galaxies, a viable tracer is needed. We here investigate the fidelity of the fine structure transition of the ($^2P_{3/2} - ^2P_{1/3}$) transition of singly-ionized carbon [C$\,{\rm \scriptsize II}$] at $158\,μ$m as a proxy for H$\,{\rm \scriptsize I}$ in a set simulated galaxies at $z\approx 6$, following the work by Heintz et al. (2021). We select 11,125 star-forming galaxies from the SIMBA simulations, with far-infrared line emissions post-processed and modeled within the SIGAME framework. We find a strong connection between [C$\,{\rm \scriptsize II}$] and H$\,{\rm \scriptsize I}$, with the relation between this [C$\,{\rm \scriptsize II}$]-to-H$\,{\rm \scriptsize I}$ relation ($β_{\rm [C\,{\rm \scriptsize II}]}$) being anti-correlated with the gas-phase metallicity of the simulated galaxies. We further use these simulations to make predictions for the total baryonic matter content of galaxies at $z\approx 6$, and specifically the HI gas mass fraction. We find mean values of $M_{\rm HI}/M_\star = 1.4$, and $M_{\rm HI}/M_{\rm bar,tot} = 0.45$. These results provide strong evidence for H$\,{\rm \scriptsize I}$ being the dominant baryonic matter component by mass in galaxies at $z\approx 6$. △ Less

Submitted 7 October, 2022; originally announced October 2022.

Comments: 7 pages, 3 figures. Accepted for publication by ApJL

Star Formation Suppresion by Tidal Removal of Cold Molecular Gas from an Intermediate-Redshift Massive Post-Starburst Galaxy

Authors: Justin S. Spilker, Katherine A. Suess, David J. Setton, Rachel Bezanson, Robert Feldmann, Jenny E. Greene, Mariska Kriek, Sidney Lower, Desika Narayanan, Margaret Verrico

Abstract: Observations and simulations have demonstrated that star formation in galaxies must be actively suppressed to prevent the formation of over-massive galaxies. Galactic outflows driven by stellar feedback or supermassive black hole accretion are often invoked to regulate the amount of cold molecular gas available for future star formation, but may not be the only relevant quenching processes in all… ▽ More Observations and simulations have demonstrated that star formation in galaxies must be actively suppressed to prevent the formation of over-massive galaxies. Galactic outflows driven by stellar feedback or supermassive black hole accretion are often invoked to regulate the amount of cold molecular gas available for future star formation, but may not be the only relevant quenching processes in all galaxies. We present the discovery of vast molecular tidal features extending up to 64 kpc outside of a massive z=0.646 post-starburst galaxy that recently concluded its primary star-forming episode. The tidal tails contain (1.2 +/- 0.1)x10^10 Msun of molecular gas, 47 +/- 5 % of the total cold gas reservoir of the system. Both the scale and magnitude of the molecular tidal features are unprecedented compared to all known nearby or high-redshift merging systems. We infer that the cold gas was stripped from the host galaxies during the merger, which is most likely responsible for triggering the initial burst phase and the subsequent suppression of star formation. While only a single example, this result shows that galaxy mergers can regulate the cold gas contents in distant galaxies by directly removing a large fraction of the molecular gas fuel, and plausibly suppress star formation directly, a qualitatively different physical mechanism than feedback-driven outflows. △ Less

Submitted 29 August, 2022; originally announced August 2022.

Comments: Published in ApJL, 8 pages, 5 figures

Recovering the star formation histories of recently-quenched galaxies: the impact of model and prior choices

Authors: Katherine A. Suess, Joel Leja, Benjamin D. Johnson, Rachel Bezanson, Jenny E. Greene, Mariska Kriek, Sidney Lower, Desika Narayanan, David J. Setton, Justin S. Spilker

Abstract: Accurate models of the star formation histories (SFHs) of recently-quenched galaxies can provide constraints on when and how galaxies shut down their star formation. The recent development of "non-parametric" SFH models promises the flexibility required to make these measurements. However, model and prior choices significantly affect derived SFHs, particularly for post-starburst galaxies (PSBs) wh… ▽ More Accurate models of the star formation histories (SFHs) of recently-quenched galaxies can provide constraints on when and how galaxies shut down their star formation. The recent development of "non-parametric" SFH models promises the flexibility required to make these measurements. However, model and prior choices significantly affect derived SFHs, particularly for post-starburst galaxies (PSBs) which have sharp changes in their recent SFH. In this paper, we create mock PSBs, then use the Prospector SED fitting software to test how well four different SFH models recover key properties. We find that a two-component parametric model performs well for our simple mock galaxies, but is sensitive to model mismatches. The fixed- and flexible-bin non-parametric models included in Prospector are able to rapidly quench a major burst of star formation, but systematically underestimate the post-burst age by up to 200 Myr. We develop a custom SFH model that allows for additional flexibility in the recent SFH. Our flexible non-parametric model is able to constrain post-burst ages with no significant offset and just ~90 Myr of scatter. Our results suggest that while standard non-parametric models are able to recover first-order quantities of the SFH (mass, SFR, average age), accurately recovering higher-order quantities (burst fraction, quenching time) requires careful consideration of model flexibility. These mock recovery tests are a critical part of future SFH studies. Finally, we show that our new, public SFH model is able to accurately recover the properties of mock star-forming and quiescent galaxies and is suitable for broader use in the SED fitting community. △ Less

Submitted 6 July, 2022; originally announced July 2022.

Comments: 24 pages, 10 figures. Posted after referee report; comments welcome. Key results in Figures 8 & 9. New SFH model publicly available as "continuity_psb_sfh" in Prospector: https://github.com/bd-j/prospector

The State of the Molecular Gas in Post-Starburst Galaxies

Authors: K. Decker French, Adam Smercina, Kate Rowlands, Akshat Tripathi, Ann I. Zabludoff, J. D. Smith, Desika Narayanan, Yujin Yang, Yancy Shirley, Katey Alatalo

Abstract: The molecular gas in galaxies traces both the fuel for star formation and the processes that can enhance or suppress star formation. Observations of the molecular gas state can thus point to when and why galaxies stop forming stars. In this study, we present ALMA observations of the molecular gas in galaxies evolving through the post-starburst phase. These galaxies have low current star formation… ▽ More The molecular gas in galaxies traces both the fuel for star formation and the processes that can enhance or suppress star formation. Observations of the molecular gas state can thus point to when and why galaxies stop forming stars. In this study, we present ALMA observations of the molecular gas in galaxies evolving through the post-starburst phase. These galaxies have low current star formation rates, regardless of the SFR tracer used, with recent starbursts ending within the last 600 Myr. We present CO (3-2) observations for three post-starburst galaxies, and dense gas HCN/HCO+/HNC (1-0) observations for six (four new) post-starburst galaxies. The post-starbursts have low excitation traced by the CO spectral line energy distribution (SLED) up to CO (3-2), more similar to early-type than starburst galaxies. The low excitation indicates that lower density rather than high temperatures may suppress star formation during the post-starburst phase. One galaxy displays a blueshifted outflow traced by CO (3-2). MaNGA observations show that the ionized gas velocity is disturbed relative to the stellar velocity field, with a blueshifted component aligned with the molecular gas outflow, suggestive of a multiphase outflow. Low ratios of HCO+/CO, indicating low fractions of dense molecular gas relative to the total molecular gas, are seen throughout post-starburst phase, except for the youngest post-starburst galaxy considered here. These observations indicate that the impact of any feedback or quenching processes may be limited to low excitation and weak outflows in the cold molecular gas during the post-starburst phase. △ Less

Submitted 12 December, 2022; v1 submitted 15 April, 2022; originally announced April 2022.

Comments: 33 pages, 24 figures, 4 tables, ApJ in press

Kiloparsec-scale imaging of the CO(1-0)-traced cold molecular gas reservoir in a z~3.4 submillimeter galaxy

Authors: Marta Frias Castillo, Matus Rybak, Jacqueline Hodge, Paul van der Werf, Dominik A. Riechers, Daniel Vieira, Gabriela Calistro Rivera, Laura N. Martinez-Ramirez, Fabian Walter, Erwin de Blok, Desika Narayanan, Jeff Wagg

Abstract: We present a high-resolution study of the cold molecular gas as traced by CO(1-0) in the unlensed z$\sim$3.4 submillimeter galaxy SMM J13120+4242, using multi-configuration observations with the Karl G. Jansky Very Large Array (JVLA). The gas reservoir, imaged on 0.39" ($\sim$3 kpc) scales, is resolved into two components separated by $\sim$11 kpc with a total extent of 16 $\sim$3 kpc. Despite the… ▽ More We present a high-resolution study of the cold molecular gas as traced by CO(1-0) in the unlensed z$\sim$3.4 submillimeter galaxy SMM J13120+4242, using multi-configuration observations with the Karl G. Jansky Very Large Array (JVLA). The gas reservoir, imaged on 0.39" ($\sim$3 kpc) scales, is resolved into two components separated by $\sim$11 kpc with a total extent of 16 $\sim$3 kpc. Despite the large spatial extent of the reservoir, the observations show a CO(1-0) FWHM linewidth of only 267 $\pm$ 64 km s$^{-1}$. We derive a revised line luminosity of L'$_\mathrm{CO(1-0)}$ = (10 $\pm$ 3) $\times$ 10$^{10}$ K km s$^{-1}$ pc$^2$ and a molecular gas mass of M$_\mathrm{gas}$ = (13 $\pm$ 3) $\times$ 10$^{10}$ ($α_\mathrm{CO}$/1) M$_{\odot}$. Despite the presence of a velocity gradient (consistent with previous resolved CO(6-5) imaging), the CO(1-0) imaging shows evidence for significant turbulent motions which are preventing the gas from fully settling into a disk. The system likely represents a merger in an advanced stage. Although the dynamical mass is highly uncertain, we use it to place an upper limit on the CO-to-H$_2$ mass conversion factor $α_\mathrm{CO}$ of 1.4. We revisit the SED fitting, finding that this galaxy lies on the very massive end of the main sequence at z = 3.4. Based on the low gas fraction, short gas depletion time and evidence for a central AGN, we propose that SMM J13120 is in a rapid transitional phase between a merger-driven starburst and an unobscured quasar. The case of SMM J13120 highlights the how mergers may drive important physical changes in galaxies without pushing them off the main sequence. △ Less

Submitted 29 March, 2022; originally announced March 2022.

Comments: 19 pages, 12 figures, accepted for publication in ApJ

Chaotic and Clumpy Galaxy Formation in an Extremely Massive Reionization-Era Halo

Authors: Justin S. Spilker, Christopher C. Hayward, Daniel P. Marrone, Manuel Aravena, Matthieu Bethermin, James Burgoyne, Scott C. Chapman, Thomas R. Greve, Gayathri Gururajan, Yashar D. Hezaveh, Ryley Hill, Katrina C. Litke, Christopher C. Lovell, Matthew A. Malkan, Eric J. Murphy, Desika Narayanan, Kedar A. Phadke, Cassie Reuter, Antony A. Stark, Nikolaus Sulzenauer, Joaquin D. Vieira, David Vizgan, Axel Weiss

Abstract: The SPT0311-58 system at z=6.900 is an extremely massive structure within the reionization epoch, and offers a chance to understand the formation of galaxies in an extreme peak in the primordial density field. We present 70mas Atacama Large Millimeter/submillimeter Array observations of the dust continuum and CII 158um emission in the central pair of galaxies and reach physical resolution ~100-350… ▽ More The SPT0311-58 system at z=6.900 is an extremely massive structure within the reionization epoch, and offers a chance to understand the formation of galaxies in an extreme peak in the primordial density field. We present 70mas Atacama Large Millimeter/submillimeter Array observations of the dust continuum and CII 158um emission in the central pair of galaxies and reach physical resolution ~100-350pc, among the most detailed views of any reionization-era system to date. The observations resolve the source into at least a dozen kiloparsec-size clumps. The global kinematics and high turbulent velocity dispersion within the galaxies present a striking contrast to recent claims of dynamically cold thin-disk kinematics in some dusty galaxies just 800Myr later at z~4. We speculate that both gravitational interactions and fragmentation from massive parent disks have likely played a role in the overall dynamics and formation of clumps in the system. Each clump individually is comparable in mass to other 6<z<8 galaxies identified in rest-UV/optical deep field surveys, but with star formation rates elevated by ~3-5x. Internally, the clumps themselves bear close resemblance to greatly scaled-up versions of virialized cloud-scale structures identified in low-redshift galaxies. Our observations are qualitatively similar to the chaotic and clumpy assembly within massive halos seen in simulations of high-redshift galaxies. △ Less

Submitted 28 March, 2022; originally announced March 2022.

Comments: Accepted for publication in ApJ Letters, 9 pages, 5 figures. Interactive versions of Fig. 2 and data products are available at https://github.com/spt-smg/publicdata

The Compact Structures of Massive $z\sim0.7$ Post-Starburst Galaxies in the SQuIGG$\vec{L}$E Sample

Authors: David J. Setton, Margaret Verrico, Rachel Bezanson, Jenny E. Greene, Katherine A. Suess, Andy D. Goulding, Justin S. Spilker, Mariska Kriek, Robert Feldmann, Desika Narayanan, Khalil Hall-Hooper, Erin Kado-Fong

Abstract: We present structural measurements of 145 spectroscopically selected intermediate-redshift (z$\sim$0.7), massive ($M_\star \sim 10^{11} \ M_\odot$) post-starburst galaxies from the SQuIGG$\vec{L}$E Sample measured using wide-depth Hyper Suprime-Cam i-band imaging. This deep imaging allows us to probe the sizes and structures of these galaxies, which we compare to a control sample of star forming a… ▽ More We present structural measurements of 145 spectroscopically selected intermediate-redshift (z$\sim$0.7), massive ($M_\star \sim 10^{11} \ M_\odot$) post-starburst galaxies from the SQuIGG$\vec{L}$E Sample measured using wide-depth Hyper Suprime-Cam i-band imaging. This deep imaging allows us to probe the sizes and structures of these galaxies, which we compare to a control sample of star forming and quiescent galaxies drawn from the LEGA-C Survey. We find that post-starburst galaxies systematically lie $\sim0.1$ dex below the quiescent mass-size (half-light radius) relation, with a scatter of $\sim0.2$ dex. This finding is bolstered by non-parametric measures, such as the Gini coefficient and the concentration, which also reveal these galaxies to have more compact light profiles than both quiescent and star-forming populations at similar mass and redshift. The sizes of post-starburst galaxies show either negative or no correlation with the time since quenching, such that more recently quenched galaxies are larger or similarly sized. This empirical finding disfavors the formation of post-starburst galaxies via a purely central burst of star formation that simultaneously shrinks the galaxy and shuts off star formation. We show that the central densities of post-starburst and quiescent galaxies at this epoch are very similar, in contrast with their effective radii. The structural properties of z$\sim$0.7 post-starburst galaxies match those of quiescent galaxies that formed in the early universe, suggesting that rapid quenching in the present epoch is driven by a similar mechanism to the one at high redshift. △ Less

Submitted 13 June, 2022; v1 submitted 16 March, 2022; originally announced March 2022.

Comments: 28 pages, 19 figures, 2 appendices, updated to accepted ApJ version

Tracing Molecular Gas in z $\simeq$ 6 Galaxies with [C${\rm \scriptsize II}$]

Authors: David Vizgan, Thomas R. Greve, Karen P. Olsen, Anita Zanella, Desika Narayanan, Romeel Davè, Georgios E. Magdis, Gergö Popping, Francesco Valentino, Kasper E. Heintz

Abstract: We investigate the fine-structure [C${\rm \scriptsize II}$] line at $158\,μ$m as a molecular gas tracer by analyzing the relationship between molecular gas mass ($M_{\rm mol}$) and [C${\rm \scriptsize II}$] line luminosity ($L_{\rm [CII]}$) in 11,125 $z\simeq 6$ star-forming, main sequence galaxies from the SIMBA simulations, with line emission modeled by SÍGAME. Though most ($\sim 50-100\,\%$) of… ▽ More We investigate the fine-structure [C${\rm \scriptsize II}$] line at $158\,μ$m as a molecular gas tracer by analyzing the relationship between molecular gas mass ($M_{\rm mol}$) and [C${\rm \scriptsize II}$] line luminosity ($L_{\rm [CII]}$) in 11,125 $z\simeq 6$ star-forming, main sequence galaxies from the SIMBA simulations, with line emission modeled by SÍGAME. Though most ($\sim 50-100\,\%$) of the gas mass in our simulations is ionized, the bulk ($> 50\,\%$) of the [C${\rm \scriptsize II}$] emission comes from the molecular phase. We find a sub-linear (slope $0.78\pm 0.01$) $\log L_{\rm [CII]}-\log M_{\rm mol}$ relation, in contrast with the linear relation derived from observational samples of more massive, metal-rich galaxies at $z \lesssim 6$. We derive a median [C${\rm \scriptsize II}$]-to-$M_{\rm mol}$ conversion factor of $α_{\rm [CII]} \simeq 18\,{\rm M_{\rm \odot}/L_{\rm \odot}}$. This is lower than the average value of $\simeq 30\,{\rm M_{\rm \odot}/L_{\rm \odot}}$ derived from observations, which we attribute to lower gas-phase metallicities in our simulations. Thus, a lower, luminosity-dependent, conversion factor must be applied when inferring molecular gas masses from [C${\rm \scriptsize II}$] observations of low-mass galaxies. For our simulations, [C${\rm \scriptsize II}$] is a better tracer of the molecular gas than CO $J=1-0$, especially at the lowest metallicities, where much of the gas is 'CO-dark'. We find that $L_{\rm [CII]}$ is more tightly correlated with $M_{\rm mol}$ than with star-formation rate (${\rm SFR}$), and both the $\log L_{\rm [CII]}-\log M_{\rm mol}$ and $\log L_{\rm [CII]}-\log {\rm SFR}$ relations arise from the Kennicutt-Schmidt relation. Our findings suggest that $L_{\rm [CII]}$ is a promising tracer of the molecular gas at the earliest cosmic epochs. △ Less

Submitted 10 March, 2022; originally announced March 2022.

Comments: 13 pages, 9 figures. Accepted for publication in ApJ

Journal ref: 2022 ApJ 929 92