Search

We present an analysis of the rest-frame optical ($\lambda \simeq 3100-5600 \,$\r{A}) spectrum of a $\mathrm{log}_{10}(M_*/\mathrm{M_\odot}) = 8.6$ star-forming galaxy at $z=8.271$ (EXCELS-63107) from JWST/NIRSpec medium-resolution observations taken as part of the EXCELS survey. EXCELS-63107 is a compact object consistent with the size of local star-forming cluster complexes ($r_e < 200 \, \rm{pc}$) and has an extremely steep UV continuum measured from JWST/NIRCam photometry ($\beta=-3.3\pm0.3$). The JWST/NIRSpec G395M spectrum of EXCELS-63107 is notable for its strong [OIII]$\lambda4363$ auroral-line emission relative to the [OIII]$\lambda5007$ forbidden line. Via a detailed emission-line and photoionization-modelling analysis, we find that the the observed properties of EXCELS-63107 are consistent with the presence of an ionizing source with an effective temperature of $T_{\rm eff} \gtrsim 80 \, 000\,\rm{K}$ heating ionized gas with a density of $n_e < 10^4 \, \rm{cm}^{-3}$ to a volume-averaged electron temperature of $T_e \simeq 34 \, 000\,\rm{K}$. Crucially, we find that stellar population models assuming a standard IMF are not capable of producing the required heating. We determine an oxygen abundance of ${12+\mathrm{log(O/H)}= 6.89^{+0.26}_{-0.21}}$ which is one of the lowest directly constrained oxygen abundances measured in any galaxy to date, and $\simeq 10 \times$ lower than is typical for $z\simeq8$ galaxies with the same stellar mass. The extremely low metallicity of EXCELS-63107 places it in a regime in which theoretical models expect a transition to a top-heavy IMF, and we speculate that a $\simeq 10-30 \, \times$ excess of $M > 50 \, \rm{M}_{\odot}$ stars is one plausible explanation for its observed properties. However, more exotic scenarios, such as Pop III star formation within a mildly enriched halo, are also consistent with the observations., Comment: 19 pages, 11 figures, submitted to MNRAS

Atomic and trapped-ion systems are the backbone of a new generation of quantum-based positioning, navigation, and timing (PNT) technologies. The miniaturization of such quantum systems offers tremendous technological advantages, especially the reduction of system size, weight, and power consumption. Yet, this has been limited by the absence of compact, standalone photonic integrated circuits (PICs) at the wavelengths suitable for these instruments. Mobilizing such photonic systems requires development of fully integrated, on-chip, active components at sub-micrometer wavelengths. We demonstrate heterogeneous photonic integrated circuits operating at 980 nm based on wafer-scale bonding of InGaAs quantum well active regions to tantalum pentoxide passive components. This high-yield process provides > 95 % surface area yield and enables integration of > 1300 active components on a 76.2 mm (3 inch) silicon wafer. We present a diverse set of functions, including semiconductor optical amplifiers, Fabry-Perot lasers, and distributed feedback lasers with 43 dB side-mode suppression ratio and > 250 GHz single-mode tuning range. We test the precise wavelength control and system level functionality of the on-chip lasers by pumping optical parametric oscillation processes in microring resonators fabricated on the same platform, generating short-wavelength signals at 778 nm and 752 nm. These results provide a pathway to realize fully functional integrated photonic engines for operation of compact quantum sensors based on atomic and trapped-ion systems., Comment: 12 pages, 5 figures

To optimize interventions for improving wellness, it is essential to understand habits, which wearable devices can measure with greater precision. Using high temporal resolution biometric data taken from the Oura Gen3 ring, we examine daily and weekly sleep and activity patterns of a cohort of young adults (N=582) in their first semester of college. A high compliance rate is observed for both daily and nightly wear, with slight dips in wear compliance observed shortly after waking up and also in the evening. Most students have a late-night chronotype with a median midpoint of sleep at 5AM, with males and those with mental health impairment having more delayed sleep periods. Social jetlag, or the difference in sleep times between free days and school days, is prevalent in our sample. While sleep periods generally shift earlier on weekdays and later on weekends, sleep duration on both weekdays and weekends is shorter than during prolonged school breaks, suggesting chronic sleep debt when school is in session. Synchronized spikes in activity consistent with class schedules are also observed, suggesting that walking in between classes is a widespread behavior in our sample that substantially contributes to physical activity. Lower active calorie expenditure is associated with weekends and a delayed but longer sleep period the night before, suggesting that for our cohort, active calorie expenditure is affected less by deviations from natural circadian rhythms and more by the timing associated with activities. Our study shows that regular sleep and activity routines may be inferred from consumer wearable devices if high temporal resolution and long data collection periods are available.

We present a spectroscopic analysis of two star-forming galaxies at z~5 observed with JWST/NIRSpec as part of the Early eXtragalactic Continuum and Emission Line Science (EXCELS) survey. The detection of the C III]$\lambda\lambda$1906,09, [O II]$\lambda\lambda$3726,29, [O III]$\lambda\lambda$4363,5007, and [N II]$\lambda$6584 nebular emission lines enables investigation of the C/O, N/O, and C/N abundance ratios using the temperature-sensitive method. The two galaxies have stellar masses of log($M_{\star}$/M$_{\odot}$ ) = 8.13$\pm$0.09 and log($M_{\star}$/M$_{\odot}$ )=8.52$\pm$0.13 and corresponding metallicities of Z~0.2Z$_{\odot}$ and Z~0.3Z$_{\odot}$. These metallicities are somewhat higher than is typical for other z>5 galaxies with similar stellar mass and are in fact comparable to high-redshift analogue galaxies at z~0. Both galaxies display evidence for N/O enhancement with respect to the z~0 sample, with log(N/O)=-1.07$\pm$0.17 and log(N/O)=-0.86$\pm$0.15 respectively. In contrast, we find low C abundances, with log(C/O)=-0.82$\pm$0.22 and log(C/O)=-1.02$\pm$0.22, consistent with the predicted yields of core-collapse supernovae. Following the trend observed in other high-redshift sources, we find that the C/N ratios are lower at fixed O/H compared to the majority of local galaxies. In contrast to the top-heavy IMF invoked in some studies to explain low C/N ratios in metal-poor galaxies, we find, via comparison to chemical evolution models, that a standard or bottom-heavy IMF better explains the observed abundance ratios in more enriched systems due to an increase in N-enrichment from intermediate mass (4-7M$_{\odot}$) stars. Our results demonstrate that robust measurements of CNO abundances with JWST can reveal unique enrichment pathways in galaxies as a function of both metallicity and redshift., Comment: 17 pages, 8 Figures, and 3 Tables. Submitted to MNRAS. Comments Welcome

We present the results of a study investigating the galaxy stellar-mass function (GSMF), size-mass relations and morphological properties of star-forming and quiescent galaxies over the redshift range $0.25

We study the homotopy theory of polyhedral products associated to a combinatorial generalisation of manifolds known as pseudomanifolds. As special cases, we show that loop spaces of moment-angle manifolds associated to triangulations of $S^2$ and $S^3$ decompose as a product of spheres and loops on spheres., Comment: 18 pages

We present an analysis of eight star-forming galaxies with $\langle z \rangle = 4.0$ from the JWST EXCELS survey for which we obtain robust chemical abundance estimates for the $\alpha$-elements O, Ne and Ar. The $\alpha$-elements are primarily produced via core-collapse supernovae (CCSNe) which should result in $\alpha$-element abundance ratios that do not vary significantly across cosmic time. However, Type Ia supernovae (SNe Ia) models predict an excess production of Ar relative to O and Ne. The Ar/O abundance ratio can therefore be used as a tracer of the relative enrichment of CCSNe and SNe Ia in galaxies. Our sample significantly increases the number of sources with measurements of ${\rm Ar/O}$ at $z > 2$, and we find that our sample exhibits sub-solar Ar/O ratios on average, with $\rm{Ar/O} = 0.65 \pm 0.10 \, (\rm{Ar/O})_{\odot}$. In contrast, the average Ne/O abundance is fully consistent with the solar ratio, with $\rm{Ne/O} = 1.07 \pm 0.12 \, (\rm{Ne/O})_{\odot}$. Our results support a scenario in which Ar has not had time to build up in the interstellar medium of young high-redshift galaxies, which are dominated by CCSNe enrichment. We show that these abundance estimates are in good agreement with recent Milky Way chemical evolution models, and with Ar/O trends observed for planetary nebulae in the Andromeda galaxy. These results highlight the potential for using multiple element abundance ratios to constrain the chemical enrichment pathways of early galaxies with JWST., Comment: 14 pages, 9 figures. Accepted for publication in MNRAS

We introduce Concept Bottleneck Protein Language Models (CB-pLM), a generative masked language model with a layer where each neuron corresponds to an interpretable concept. Our architecture offers three key benefits: i) Control: We can intervene on concept values to precisely control the properties of generated proteins, achieving a 3 times larger change in desired concept values compared to baselines. ii) Interpretability: A linear mapping between concept values and predicted tokens allows transparent analysis of the model's decision-making process. iii) Debugging: This transparency facilitates easy debugging of trained models. Our models achieve pre-training perplexity and downstream task performance comparable to traditional masked protein language models, demonstrating that interpretability does not compromise performance. While adaptable to any language model, we focus on masked protein language models due to their importance in drug discovery and the ability to validate our model's capabilities through real-world experiments and expert knowledge. We scale our CB-pLM from 24 million to 3 billion parameters, making them the largest Concept Bottleneck Models trained and the first capable of generative language modeling.

Large language models (LLMs) have recently shown significant potential in various biological tasks such as protein engineering and molecule design. These tasks typically involve black-box discrete sequence optimization, where the challenge lies in generating sequences that are not only biologically feasible but also adhere to hard fine-grained constraints. However, LLMs often struggle with such constraints, especially in biological contexts where verifying candidate solutions is costly and time-consuming. In this study, we explore the possibility of employing LLMs as highly-constrained bilevel optimizers through a methodology we refer to as Language Model Optimization with Margin Expectation (LLOME). This approach combines both offline and online optimization, utilizing limited oracle evaluations to iteratively enhance the sequences generated by the LLM. We additionally propose a novel training objective -- Margin-Aligned Expectation (MargE) -- that trains the LLM to smoothly interpolate between the reward and reference distributions. Lastly, we introduce a synthetic test suite that bears strong geometric similarity to real biophysical problems and enables rapid evaluation of LLM optimizers without time-consuming lab validation. Our findings reveal that, in comparison to genetic algorithm baselines, LLMs achieve significantly lower regret solutions while requiring fewer test function evaluations. However, we also observe that LLMs exhibit moderate miscalibration, are susceptible to generator collapse, and have difficulty finding the optimal solution when no explicit ground truth rewards are available., Comment: Supercedes arXiv:2407.00236v1

Autoregressive (AR) Transformer-based sequence models are known to have difficulty generalizing to sequences longer than those seen during training. When applied to text-to-speech (TTS), these models tend to drop or repeat words or produce erratic output, especially for longer utterances. In this paper, we introduce enhancements aimed at AR Transformer-based encoder-decoder TTS systems that address these robustness and length generalization issues. Our approach uses an alignment mechanism to provide cross-attention operations with relative location information. The associated alignment position is learned as a latent property of the model via backprop and requires no external alignment information during training. While the approach is tailored to the monotonic nature of TTS input-output alignment, it is still able to benefit from the flexible modeling power of interleaved multi-head self- and cross-attention operations. A system incorporating these improvements, which we call Very Attentive Tacotron, matches the naturalness and expressiveness of a baseline T5-based TTS system, while eliminating problems with repeated or dropped words and enabling generalization to any practical utterance length., Comment: Submitted to NAACL

Accurately quantifying the ionizing photon production efficiency ($\xi_\rm{ion}$) of $z>6$ star-forming galaxies (SFGs) is necessary to understand their contribution to reionization. We investigate the ionizing properties of N=289 SFGs selected at $z=6.9-7.6$ from the JWST Cycle-1 imaging programmes; PRIMER and JADES. We use BAGPIPES to consistently infer the equivalent widths of their [OIII]+$\rm{H\beta}$ emission lines ($W_\lambda$) and their physical properties. To supplement our high-redshift galaxies, we measure $W_\lambda$ photometrically for a sample of N=253 $z=3.2-3.6$ SFGs selected from the VANDELS spectroscopic survey. Comparing these samples, we find a strong apparent redshift evolution in their median $W_\lambda$, increasing from $W_\lambda =310\pm25\,\r{A}$ in VANDELS to $W_\lambda =630\pm30\,\r{A}$ in our JWST-based sample. In the JWST sample at $z>7$, we find that $W_\lambda$ correlates with both stellar mass and UV luminosity, with high-mass, $M_{ UV}-$faint galaxies producing systematically weaker emission lines. Moreover, we discover a departure from the standard log-normal shape of the $W_\lambda$ distribution, characterised by a more pronounced tail at lower $W_\lambda$, consistent with increasingly bursty star formation. Using $W_\lambda$ as a proxy for $\xi_\rm{ion}$, and UV spectral slope as a proxy for LyC escape fraction ($f_\rm{esc}$), we find a minority of galaxies with high $\xi_\rm{ion}$ and $f_\rm{esc}$ (e.g., $\rm{log(\xi_{ion}/erg^{-1}Hz})\sim25.6$ and $f_\rm{esc}\sim0.15$). However, we find that the LyC photon budget at $z>7$ is dominated by galaxies with more moderate output, close to the sample median of $\rm{log(\xi_{ion}/erg^{-1}Hz})\sim25.3$ and $f_\rm{esc}\sim0.05$. This is consistent with estimates for the number of LyC photons required to power reionization at $z>7$, with no evidence for over or under-production., Comment: 18 pages, 9 figures + 1 appendix. Submitted to MNRAS

Computational design of protein-binding proteins is a fundamental capability with broad utility in biomedical research and biotechnology. Recent methods have made strides against some target proteins, but on-demand creation of high-affinity binders without multiple rounds of experimental testing remains an unsolved challenge. This technical report introduces AlphaProteo, a family of machine learning models for protein design, and details its performance on the de novo binder design problem. With AlphaProteo, we achieve 3- to 300-fold better binding affinities and higher experimental success rates than the best existing methods on seven target proteins. Our results suggest that AlphaProteo can generate binders "ready-to-use" for many research applications using only one round of medium-throughput screening and no further optimization., Comment: 45 pages, 17 figures

The Fermilab Side-Coupled Linac accelerates H-beam from 116 MeV to 400 MeV through seven 805 MHz modules. Twelve wire scanners are present in the Side Coupled Linac and four are present in the transfer line between the Linac and the Booster synchrotron ring. These wire scanners act as important diagnostic instruments to directly collect information on the beam's transverse distribution. The manipulation of the conditions of wire scanner data collection enables further characterization of the beamline, such as calculating emittance and the Twiss parameters of the beam at select regions, which we present here., Comment: 32nd Linear Accelerator Conference (LINAC 2024)

Recent work by Ram\'irez et al. [2] has introduced Multi-Channel Factor Analysis (MFA) as an extension of factor analysis to multi-channel data that allows for latent factors common to all channels as well as factors specific to each channel. This paper validates the MFA covariance model and analyzes the statistical properties of the MFA estimators. In particular, a thorough investigation of model identifiability under varying latent factor structures is conducted, and sufficient conditions for generic global identifiability of MFA are obtained. The development of these identifiability conditions enables asymptotic analysis of estimators obtained by maximizing a Gaussian likelihood, which are shown to be consistent and asymptotically normal even under misspecification of the latent factor distribution.

This article develops a methodology allowing application of the complete machinery of particle-based inference methods upon what we call the class of continuous-discrete State Space Models (CD-SSMs). Such models correspond to a latent continuous-time It\^o diffusion process which is observed with noise at discrete time instances. Due to the continuous-time nature of the hidden signal, standard Feynman-Kac formulations and their accompanying particle-based approximations have to overcome several challenges, arising mainly due to the following considerations: (i) finite-time transition densities of the signal are typically intractable; (ii) ancestors of sampled signals are determined w.p.~1, thus cannot be resampled; (iii) diffusivity parameters given a sampled signal yield Dirac distributions. We overcome all above issues by introducing a framework based on carefully designed proposals and transformations thereof. That is, we obtain new expressions for the Feynman-Kac model that accommodate the effects of a continuous-time signal and overcome induced degeneracies. The constructed formulations will enable use of the full range of particle-based algorithms for CD-SSMs: for filtering/smoothing and parameter inference, whether online or offline. Our framework is compatible with guided proposals in the filtering steps that are essential for efficient algorithmic performance in the presence of informative observations or in higher dimensions, and is applicable for a very general class of CD-SSMs, including the case when the signal is modelled as a hypo-elliptic diffusion. Our methods can be immediately incorporated to available software packages for particle-based algorithms., Comment: 29 pages

We show that the loop space of a moment-angle complex associated to a $2$-dimensional simplicial complex decomposes as a finite type product of spheres, loops on spheres, and certain indecomposable spaces which appear in the loop space decomposition of Moore spaces. We also give conditions on certain subcomplexes under which, localised away from sufficiently many primes, the loop space of a moment-angle complex decomposes as a finite type product of spheres and loops on spheres., Comment: 22 pages, comments welcome

There is a growing body of work seeking to replicate the success of machine learning (ML) on domains like computer vision (CV) and natural language processing (NLP) to applications involving biophysical data. One of the key ingredients of prior successes in CV and NLP was the broad acceptance of difficult benchmarks that distilled key subproblems into approachable tasks that any junior researcher could investigate, but good benchmarks for biophysical domains are rare. This scarcity is partially due to a narrow focus on benchmarks which simulate biophysical data; we propose instead to carefully abstract biophysical problems into simpler ones with key geometric similarities. In particular we propose a new class of closed-form test functions for biophysical sequence optimization, which we call Ehrlich functions. We provide empirical results demonstrating these functions are interesting objects of study and can be non-trivial to solve with a standard genetic optimization baseline.

We propose a novel technique for guidance of buoyancy-controlled vehicles in uncertain under-ice ocean flows. In-situ melt rate measurements collected at the grounding zone of Antarctic ice shelves, where the ice shelf meets the underlying bedrock, are essential to constrain models of future sea level rise. Buoyancy-controlled vehicles, which control their vertical position in the water column through internal actuation but have no means of horizontal propulsion, offer an affordable and reliable platform for such in-situ data collection. However, reaching the grounding zone requires vehicles to traverse tens of kilometers under the ice shelf, with approximate position knowledge and no means of communication, in highly variable and uncertain ocean currents. To address this challenge, we propose a partially observable MDP approach that exploits model-based knowledge of the under-ice currents and, critically, of their uncertainty, to synthesize effective guidance policies. The approach uses approximate dynamic programming to model uncertainty in the currents, and QMDP to address localization uncertainty. Numerical experiments show that the policy can deliver up to 88.8% of underwater vehicles to the grounding zone -- a 33% improvement compared to state-of-the-art guidance techniques, and a 262% improvement over uncontrolled drifters. Collectively, these results show that model-based under-ice guidance is a highly promising technique for exploration of under-ice cavities, and has the potential to enable cost-effective and scalable access to these challenging and rarely observed environments., Comment: Presented at IROS 2021