Search

Cardio-mechanical models can be used to support clinical decision-making. Unfortunately, the substantial computational effort involved in many cardiac models hinders their application in the clinic, despite the fact that they may provide valuable information. In this work, we present a probabilistic reduced-order modeling (ROM) framework to dramatically reduce the computational effort of such models while providing a credibility interval. In the online stage, a fast-to-evaluate generalized one-fiber model is considered. This generalized one-fiber model incorporates correction factors to emulate patient-specific attributes, such as local geometry variations. In the offline stage, Bayesian inference is used to calibrate these correction factors on training data generated using a full-order isogeometric cardiac model (FOM). A Gaussian process is used in the online stage to predict the correction factors for geometries that are not in the training data. The proposed framework is demonstrated using two examples. The first example considers idealized left-ventricle geometries, for which the behavior of the ROM framework can be studied in detail. In the second example, the ROM framework is applied to scan-based geometries, based on which the application of the ROM framework in the clinical setting is discussed. The results for the two examples convey that the ROM framework can provide accurate online predictions, provided that adequate FOM training data is available. The uncertainty bands provided by the ROM framework give insight into the trustworthiness of its results. Large uncertainty bands can be considered as an indicator for the further population of the training data set.

With a growing number of facilities able to monitor the entire sky and produce light curves with a cadence of days, in recent years there has been an increased rate of detection of sources whose variability deviates from standard behavior, revealing a variety of exotic nuclear transients. The aim of the present study is to disentangle the nature of the transient AT 2021hdr, whose optical light curve used to be consistent with a classic Seyfert 1 nucleus, which was also confirmed by its optical spectrum and high-energy properties. From late 2021, AT 2021hdr started to present sudden brightening episodes in the form of oscillating peaks in the Zwicky Transient Facility (ZTF) alert stream, and the same shape is observed in X-rays and UV from Swift data. The oscillations occur every about 60-90 days with amplitudes of around 0.2 mag in the g and r bands. Very Long Baseline Array (VLBA) observations show no radio emission at milliarcseconds scale. It is argued that these findings are inconsistent with a standard tidal disruption event (TDE), a binary supermassive black hole (BSMBH), or a changing-look active galactic nucleus (AGN); neither does this object resemble previous observed AGN flares, and disk or jet instabilities are an unlikely scenario. Here, we propose that the behavior of AT 2021hdr might be due to the tidal disruption of a gas cloud by a BSMBH. In this scenario, we estimate that the putative binary has a separation of about 0.83 mpc and would merge in about 70000 years. This galaxy is located at 9 kpc from a companion galaxy, and in this work we report this merger for the first time. The oscillations are not related to the companion galaxy., Comment: 11 pages, 5 figures, 2 tables, plus appendices. Accepted for publication in Astronomy & Astrophysics. NASA press release available at https://science.nasa.gov/missions/swift/nasas-swift-studies-gas-churning-monster-black-holes/ doi: 10.1051/0004-6361/202451305

This study explores the semiclassical limit of an integrable-chaotic bosonic many-body quantum system, providing nuanced insights into its behavior. We examine classical-quantum correspondences across different interaction regimes of bosons in a triple-well potential, ranging from the integrable to the self-trapping regime, and including the chaotic one. The close resemblance between the phase-space mean projections of classical trajectories and those of Husimi distributions evokes the Principle of Uniform Semiclassical Condensation (PUSC) of Wigner functions of eigenstates. Notably, the resulting figures also exhibit patterns reminiscent of Jason Gallas's "shrimp" shapes., Comment: 8 pages, 9 figures

Spatially-resolved emission line kinematics are invaluable to investigating fundamental galaxy properties and have become increasingly accessible for galaxies at $z\gtrsim0.5$ through sensitive near-infrared imaging spectroscopy and millimeter interferometry. Kinematic modeling is at the core of the analysis and interpretation of such data sets, which at high-z present challenges due to lower signal-to-noise ratio (S/N) and resolution compared to data of local galaxies. We present and test the 3D fitting functionality of DysmalPy, examining how well it recovers intrinsic disk rotation velocity and velocity dispersion, using a large suite of axisymmetric models, covering a range of galaxy properties and observational parameters typical of $z\sim1$-$3$ star-forming galaxies. We also compare DysmalPy's recovery performance to that of two other commonly used codes, GalPak3D and 3DBarolo, which we use in turn to create additional sets of models to benchmark DysmalPy. Over the ranges of S/N, resolution, mass, and velocity dispersion explored, the rotation velocity is accurately recovered by all tools. The velocity dispersion is recovered well at high S/N, but the impact of methodology differences is more apparent. In particular, template differences for parametric tools and S/N sensitivity for the non-parametric tool can lead to differences up to a factor of 2. Our tests highlight and the importance of deep, high-resolution data and the need for careful consideration of: (1) the choice of priors (parametric approaches), (2) the masking (all approaches) and, more generally, evaluating the suitability of each approach to the specific data at hand. This paper accompanies the public release of DysmalPy., Comment: 38 pages, 21 figures, accepted for publication in ApJ

We present an empirical study investigating how specific properties of preference datasets, such as mixed-quality or noisy data, affect the performance of Preference Optimization (PO) algorithms. Our experiments, conducted in MuJoCo environments, reveal several scenarios where state-of-the-art PO methods experience significant drops in performance. To address this issue, we introduce a novel PO framework based on mirror descent, which can recover existing methods like Direct Preference Optimization (DPO) and Odds-Ratio Preference Optimization (ORPO) for specific choices of the mirror map. Within this framework, we employ evolutionary strategies to discover new loss functions capable of handling the identified problematic scenarios. These new loss functions lead to significant performance improvements over DPO and ORPO across several tasks. Additionally, we demonstrate the generalization capability of our approach by applying the discovered loss functions to fine-tuning large language models using mixed-quality data, where they outperform ORPO.

It is now a common business practice to buy access to large language model (LLM) inference rather than self-host, because of significant upfront hardware infrastructure and energy costs. However, as a buyer, there is no mechanism to verify the authenticity of the advertised service including the serving hardware platform, e.g. that it is actually being served using an NVIDIA H100. Furthermore, there are reports suggesting that model providers may deliver models that differ slightly from the advertised ones, often to make them run on less expensive hardware. That way, a client pays premium for a capable model access on more expensive hardware, yet ends up being served by a (potentially less capable) cheaper model on cheaper hardware. In this paper we introduce \textit{\textbf{hardware and software platform inference (HSPI)}} -- a method for identifying the underlying \GPU{} architecture and software stack of a (black-box) machine learning model solely based on its input-output behavior. Our method leverages the inherent differences of various \GPU{} architectures and compilers to distinguish between different \GPU{} types and software stacks. By analyzing the numerical patterns in the model's outputs, we propose a classification framework capable of accurately identifying the \GPU{} used for model inference as well as the underlying software configuration. Our findings demonstrate the feasibility of inferring \GPU{} type from black-box models. We evaluate HSPI against models served on different real hardware and find that in a white-box setting we can distinguish between different \GPU{}s with between $83.9\%$ and $100\%$ accuracy. Even in a black-box setting we are able to achieve results that are up to three times higher than random guess accuracy.

Zero-shot coordination (ZSC) is a popular setting for studying the ability of reinforcement learning (RL) agents to coordinate with novel partners. Prior ZSC formulations assume the $\textit{problem setting}$ is common knowledge: each agent knows the underlying Dec-POMDP, knows others have this knowledge, and so on ad infinitum. However, this assumption rarely holds in complex real-world settings, which are often difficult to fully and correctly specify. Hence, in settings where this common knowledge assumption is invalid, agents trained using ZSC methods may not be able to coordinate well. To address this limitation, we formulate the $\textit{noisy zero-shot coordination}$ (NZSC) problem. In NZSC, agents observe different noisy versions of the ground truth Dec-POMDP, which are assumed to be distributed according to a fixed noise model. Only the distribution of ground truth Dec-POMDPs and the noise model are common knowledge. We show that a NZSC problem can be reduced to a ZSC problem by designing a meta-Dec-POMDP with an augmented state space consisting of all the ground-truth Dec-POMDPs. For solving NZSC problems, we propose a simple and flexible meta-learning method called NZSC training, in which the agents are trained across a distribution of coordination problems - which they only get to observe noisy versions of. We show that with NZSC training, RL agents can be trained to coordinate well with novel partners even when the (exact) problem setting of the coordination is not common knowledge.

A significant challenge in the study of transient astrophysical phenomena is the identification of bogus events, with human-made Earth-orbiting satellites and debris remain a key contaminant. Existing pipelines effectively identify satellite trails but can miss more complex signatures, such as collections of dots known as satellite glints. In the Rubin Observatory era, the scale of the operations will increase tenfold with respect to its precursor, the Zwicky Transient Facility (ZTF), requiring crucial improvements in classification purity, data compression, pipeline speed and more. We explore the use of the 2D Fast Fourier Transform (FFT) on difference images as a tool to improve satellite detection algorithms. Adopting the single-stamp classification model from the Automatic Learning for the Rapid Classification of Events (ALeRCE) broker as a baseline, we adapt its architecture to receive a cutout of the FFT of the difference image, in addition to the three (science, reference, difference) ZTF image cutouts (hereafter stamps). We study different stamp sizes and resolutions for these four channels, aiming to assess the benefit of including the FFT image, especially in scenarios with data compression and processing speed requirements (e.g., for surveys like the Legacy Survey of Space and Time). The inclusion of the FFT improved satellite detection accuracy, with the most notable increase observed in the model with the smallest field of view (16''), where accuracy rose from 66.9% to 79.7% (a statistically significant improvement of ~13% with a 95% confidence interval of 7.8% to 17.8%). This result demonstrates the effectiveness of FFT in compressing relevant information and extracting features that characterize satellite signatures in larger difference images. We show how FFTs can be leveraged to cull satellite and space debris signatures from alert streams., Comment: 8 pages, 5 figures, submitted to A&A. Abstract shortened here

Proximal policy optimization (PPO) is a widely-used algorithm for on-policy reinforcement learning. This work offers an alternative perspective of PPO, in which it is decomposed into the inner-loop estimation of update vectors, and the outer-loop application of updates using gradient ascent with unity learning rate. Using this insight we propose outer proximal policy optimization (outer-PPO); a framework wherein these update vectors are applied using an arbitrary gradient-based optimizer. The decoupling of update estimation and update application enabled by outer-PPO highlights several implicit design choices in PPO that we challenge through empirical investigation. In particular we consider non-unity learning rates and momentum applied to the outer loop, and a momentum-bias applied to the inner estimation loop. Methods are evaluated against an aggressively tuned PPO baseline on Brax, Jumanji and MinAtar environments; non-unity learning rates and momentum both achieve statistically significant improvement on Brax and Jumanji, given the same hyperparameter tuning budget.

We conceptualize the communication of narratives as a cheap-talk game under model uncertainty. The sender has private information about the true data generating process of publicly observable data. The receiver is uncertain about how to interpret the data, but aware of the sender's incentives to strategically provide interpretations ("narratives") in her favor. We consider a general class of decision rules under ambiguity resolving the receiver's ignorance of the true data generating process, including maximum likelihood expected utility. The set of equilibria is characterized by a positive integer $N$: there is an equilibrium that induces $n$ different actions for each $1\leq n \leq N$. The diverting power of the sender is weaker than with a na\"ive receiver being unaware of the sender's incentives. Surprisingly, the receiver sometimes prefers to be na\"ive.

While large models trained with self-supervised learning on offline datasets have shown remarkable capabilities in text and image domains, achieving the same generalisation for agents that act in sequential decision problems remains an open challenge. In this work, we take a step towards this goal by procedurally generating tens of millions of 2D physics-based tasks and using these to train a general reinforcement learning (RL) agent for physical control. To this end, we introduce Kinetix: an open-ended space of physics-based RL environments that can represent tasks ranging from robotic locomotion and grasping to video games and classic RL environments, all within a unified framework. Kinetix makes use of our novel hardware-accelerated physics engine Jax2D that allows us to cheaply simulate billions of environment steps during training. Our trained agent exhibits strong physical reasoning capabilities, being able to zero-shot solve unseen human-designed environments. Furthermore, fine-tuning this general agent on tasks of interest shows significantly stronger performance than training an RL agent *tabula rasa*. This includes solving some environments that standard RL training completely fails at. We believe this demonstrates the feasibility of large scale, mixed-quality pre-training for online RL and we hope that Kinetix will serve as a useful framework to investigate this further., Comment: The first two authors contributed equally. Project page located at: https://kinetix-env.github.io/

We investigate the design of magnetic ordering in one-dimensional mesoscopic magnetic Ising chains by modulating long-range interactions. These interactions are affected by geometrical modifications to the chain, which adjust the energy hierarchy and the resulting magnetic ground states. Consequently, the magnetic ordering can be tuned between antiferromagnetic and dimer antiferromagnetic phases. These phases are experimentally observed in chains fabricated using both conventional electron-beam lithography and ion implantation techniques, demonstrating the feasibility of controlling magnetic properties at the mesoscale. The ability of attaining these magnetic structures by thermal annealing, underlines the potential of using such systems instead of simulated annealers in tackling combinatorial optimization tasks., Comment: 12 pages, 9 figures

We introduce a multi-turn benchmark for evaluating personalised alignment in LLM-based AI assistants, focusing on their ability to handle user-provided safety-critical contexts. Our assessment of ten leading models across five scenarios (each with 337 use cases) reveals systematic inconsistencies in maintaining user-specific consideration, with even top-rated "harmless" models making recommendations that should be recognised as obviously harmful to the user given the context provided. Key failure modes include inappropriate weighing of conflicting preferences, sycophancy (prioritising user preferences above safety), a lack of attentiveness to critical user information within the context window, and inconsistent application of user-specific knowledge. The same systematic biases were observed in OpenAI's o1, suggesting that strong reasoning capacities do not necessarily transfer to this kind of personalised thinking. We find that prompting LLMs to consider safety-critical context significantly improves performance, unlike a generic 'harmless and helpful' instruction. Based on these findings, we propose research directions for embedding self-reflection capabilities, online user modelling, and dynamic risk assessment in AI assistants. Our work emphasises the need for nuanced, context-aware approaches to alignment in systems designed for persistent human interaction, aiding the development of safe and considerate AI assistants., Comment: Submitted to ICLR 2025 on 01/10/2024

Soft robotic manipulators offer operational advantage due to their compliant and deformable structures. However, their inherently nonlinear dynamics presents substantial challenges. Traditional analytical methods often depend on simplifying assumptions, while learning-based techniques can be computationally demanding and limit the control policies to existing data. This paper introduces a novel approach to soft robotic control, leveraging state-of-the-art policy gradient methods within parallelizable synthetic environments learned from data. We also propose a safety oriented actuation space exploration protocol via cascaded updates and weighted randomness. Specifically, our recurrent forward dynamics model is learned by generating a training dataset from a physically safe \textit{mean reverting} random walk in actuation space to explore the partially-observed state-space. We demonstrate a reinforcement learning approach towards closed-loop control through state-of-the-art actor-critic methods, which efficiently learn high-performance behaviour over long horizons. This approach removes the need for any knowledge regarding the robot's operation or capabilities and sets the stage for a comprehensive benchmarking tool in soft robotics control., Comment: soft manipulator, reinforcement learning, learned controllers

Hedin's $GW$ approximation to the electronic self-energy has been impressively successful to calculate quasiparticle energies, such as ionization potentials, electron affinities, or electronic band structures. The success of this fairly simple approximation has been ascribed to the cancellation of the so-called vertex corrections that go beyond $GW$. This claim is mostly based on past calculations using vertex corrections within the crude local-density approximation. Here, we explore a wide variety of non-local vertex corrections in the polarizability and the self-energy, using first-order approximations or infinite summations to all orders. In particular, we use vertices based on statically screened interactions like in the Bethe-Salpeter equation. We demonstrate on realistic molecular systems that the two vertices in Hedin's equation essentially compensate. We further show that consistency between the two vertices is crucial to obtain realistic electronic properties. We finally consider increasingly large clusters and extrapolate that our conclusions would hold for extended systems., Comment: Update of original version: More detailed results including an additional figure, additional references, and another figure with diagrams

We present deep ($\sim$ 20 hr), high-angular resolution Atacama Large Millimeter/submillimeter Array (ALMA) observations of the $\rm CO ~ (4-3)$ and $\rm [CI] ~ (1-0)$ transitions, along with the rest-frame 630 $\mu$m dust continuum, in BX610 --a massive, main-sequence galaxy at the peak epoch of cosmic star formation $(z = 2.21)$. Combined with deep Very Large Telescope (VLT) SINFONI observations of the H$\alpha$ line, we characterize the molecular gas and star formation activity on kiloparsec scales. Our analysis reveals that the excitation of the molecular gas, as traced by the $L'_{\rm CO ~ (4-3)} / L'_{\rm [CI] ~ (1-0)}$ line luminosity ratio, decreases with increasing galactocentric radius. While the line luminosity ratios in the outskirts are similar to those typically found in main-sequence galaxies at $z \sim 1$, the ratios in the central regions of BX610 are comparable to those observed in local starbursts. There is also a giant extra-nuclear star-forming clump in the southwest of BX610 that exhibits high star formation activity, molecular gas abundance, and molecular gas excitation. Furthermore, the central region of BX610 is rich in molecular gas $(M_{\rm mol} / M_{\rm \star} \approx 1)$; however, at the current level of star formation activity, such molecular gas is expected to be depleted in $\sim$ 450 Myr. This, along with recent evidence for rapid inflow toward the center, suggests that BX610 may be experiencing an evolutionary phase often referred to as wet compaction, which is expected to lead to central gas depletion and subsequent inside-out quenching of star formation activity., Comment: Submitted to Astronomy & Astrophysics. 09 pages, 05 figures

Owing to their similarities with giant exoplanets, brown dwarf companions of stars provide insights into the fundamental processes of planet formation and evolution. From their orbits, several brown dwarf companions are found to be more massive than theoretical predictions given their luminosities and the ages of their host stars (e.g. Brandt et al. 2021, Cheetham et al. 2018, Li et al. 2023). Either the theory is incomplete or these objects are not single entities. For example, they could be two brown dwarfs each with a lower mass and intrinsic luminosity (Brandt et al. 2021, Howe et al. 2024). The most problematic example is Gliese 229 B (Nakajima et al. 1995, Oppenheimer et al. 1995), which is at least 2-6 times less luminous than model predictions given its dynamical mass of $71.4\pm0.6$ Jupiter masses ($M_{\rm Jup}$) (Brandt et al. 2021). We observed Gliese 229 B with the GRAVITY interferometer and, separately, the CRIRES+ spectrograph at the Very Large Telescope. Both sets of observations independently resolve Gliese 229 B into two components, Gliese 229 Ba and Bb, settling the conflict between theory and observations. The two objects have a flux ratio of $0.47\pm0.03$ at a wavelength of 2 $\mu$m and masses of $38.1\pm1.0$ and $34.4\pm1.5$ $M_{\rm Jup}$, respectively. They orbit each other every 12.1 days with a semimajor axis of 0.042 astronomical units (AU). The discovery of Gliese 229 BaBb, each only a few times more massive than the most massive planets, and separated by 16 times the Earth-moon distance, raises new questions about the formation and prevalence of tight binary brown dwarfs around stars., Comment: Published in Nature. The Version of Record of this article is located at https://www.nature.com/articles/s41586-024-08064-x

Given the widespread adoption and usage of Large Language Models (LLMs), it is crucial to have flexible and interpretable evaluations of their instruction-following ability. Preference judgments between model outputs have become the de facto evaluation standard, despite distilling complex, multi-faceted preferences into a single ranking. Furthermore, as human annotation is slow and costly, LLMs are increasingly used to make these judgments, at the expense of reliability and interpretability. In this work, we propose TICK (Targeted Instruct-evaluation with ChecKlists), a fully automated, interpretable evaluation protocol that structures evaluations with LLM-generated, instruction-specific checklists. We first show that, given an instruction, LLMs can reliably produce high-quality, tailored evaluation checklists that decompose the instruction into a series of YES/NO questions. Each question asks whether a candidate response meets a specific requirement of the instruction. We demonstrate that using TICK leads to a significant increase (46.4% $\to$ 52.2%) in the frequency of exact agreements between LLM judgements and human preferences, as compared to having an LLM directly score an output. We then show that STICK (Self-TICK) can be used to improve generation quality across multiple benchmarks via self-refinement and Best-of-N selection. STICK self-refinement on LiveBench reasoning tasks leads to an absolute gain of $+$7.8%, whilst Best-of-N selection with STICK attains $+$6.3% absolute improvement on the real-world instruction dataset, WildBench. In light of this, structured, multi-faceted self-improvement is shown to be a promising way to further advance LLM capabilities. Finally, by providing LLM-generated checklists to human evaluators tasked with directly scoring LLM responses to WildBench instructions, we notably increase inter-annotator agreement (0.194 $\to$ 0.256).

Modern communication networks feature local fast failover mechanisms in the data plane, to swiftly respond to link failures with pre-installed rerouting rules. This paper explores resilient routing meant to tolerate $\leq k$ simultaneous link failures, ensuring packet delivery, provided that the source and destination remain connected. While past theoretical works studied failover routing under static link failures, i.e., links which permanently and simultaneously fail, real-world networks often face link flapping--dynamic down states caused by, e.g., numerous short-lived software-related faults. Thus, in this initial work, we re-investigate the resilience of failover routing against link flapping, by categorizing link failures into static, semi-dynamic (removing the assumption that links fail simultaneously), and dynamic (removing the assumption that links fail permanently) types, shedding light on the capabilities and limitations of failover routing under these scenarios. We show that $k$-edge-connected graphs exhibit $(k-1)$-resilient routing against dynamic failures for $k \leq 5$. We further show that this result extends to arbitrary $k$ if it is possible to rewrite $\log k$ bits in the packet header. Rewriting $3$ bits suffices to cope with $k$ semi-dynamic failures. However, on general graphs, tolerating $2$ dynamic failures becomes impossible without bit-rewriting. Even by rewriting $\log k$ bits, resilient routing cannot resolve $k$ dynamic failures, demonstrating the limitation of local fast rerouting.

We present the properties of a massive, large, dusty, metal-rich, star-forming galaxy at z_spec=6.73. GOODSN-100182 was observed with JWST/NIRSpec as part of the AURORA survey, and is also covered by public multi-wavelength HST and JWST imaging. While the large mass of GOODSN-100182 (~10^10 M_sun) was indicated prior to JWST, NIRCam rest-optical imaging now reveals the presence of an extended disk (r_eff~1.5 kpc). In addition, the NIRSpec R~1000 spectrum of GOODSN-100182 includes the detection of a large suite of rest-optical nebular emission lines ranging in wavelength from [OII]3727 up to [NII]6583. The ratios of Balmer lines suggest significant dust attenuation (E(B-V)_gas=0.40+0.10/-0.09), consistent with the red rest-UV slope inferred for GOODSN-100182 (beta=-0.50+/-0.09). The star-formation rate based on dust-corrected H-alpha emission is log(SFR(H-alpha)/ M_sun/yr)=2.02+0.13/-0.14, well above the z~7 star-forming main sequence in terms of specific SFR. Strikingly, the ratio of [NII]6583/H-alpha emission suggests almost solar metallicity, as does the ratio ([OIII]5007/H-beta)/([NII]6583/H-alpha) and the detection of the faint [FeII]4360 emission feature, whereas the [OIII]5007/[OII]3727 ratio suggests roughly 50% solar metallicity. Overall, the excitation and ionization properties of GOODSN-100182 more closely resemble those of typical star-forming galaxies at z~2-3 rather than z~7. Based on public spectroscopy of the GOODS-N field, we find that GOODSN-100182 resides within a significant galaxy overdensity, and is accompanied by a spectroscopically-confirmed neighbor galaxy. GOODSN-100182 demonstrates the existence of mature, chemically-enriched galaxies within the first billion years of cosmic time, whose properties must be explained by galaxy formation models., Comment: 16 pages, 13 figures, submitted to ApJ

We resolve the multiple images of the binary-lens microlensing event ASASSN-22av using the GRAVITY instrument of the Very Large Telescope Interferometer (VLTI). The light curves show weak binary perturbations, complicating the analysis, but the joint modeling with the VLTI data breaks several degeneracies, arriving at a strongly favored solution. Thanks to precise measurements of angular Einstein radius \theta_E = 0.724 +/- 0.002 mas and microlens parallax, we determine that the lens system consists of two M dwarfs with masses of M_1 = 0.258 +/- 0.008 M_sun and M_2 = 0.130 +/- 0.007 M_sun, a projected separation of r_\perp = 6.83 +/- 0.31 AU and a distance of D_L = 2.29 +/- 0.08 kpc. The successful VLTI observations of ASASSN-22av open up a new path for studying intermediate-separation (i.e., a few AUs) stellar-mass binaries, including those containing dark compact objects such as neutron stars and stellar-mass black holes., Comment: Accepted by ApJ. See the ancillary file for animation associated with Fig. 8

Studying the orbital motion of stars around Sagittarius A* in the Galactic Center provides a unique opportunity to probe the gravitational potential near the supermassive black hole at the heart of our Galaxy. Interferometric data obtained with the GRAVITY instrument at the Very Large Telescope Interferometer (VLTI) since 2016 has allowed us to achieve unprecedented precision in tracking the orbits of these stars. GRAVITY data have been key to detecting the in-plane, prograde Schwarzschild precession of the orbit of the star S2, as predicted by General Relativity. By combining astrometric and spectroscopic data from multiple stars, including S2, S29, S38, and S55 - for which we have data around their time of pericenter passage with GRAVITY - we can now strengthen the significance of this detection to an approximately $10 \sigma$ confidence level. The prograde precession of S2's orbit provides valuable insights into the potential presence of an extended mass distribution surrounding Sagittarius A*, which could consist of a dynamically relaxed stellar cusp comprised of old stars and stellar remnants, along with a possible dark matter spike. Our analysis, based on two plausible density profiles - a power-law and a Plummer profile - constrains the enclosed mass within the orbit of S2 to be consistent with zero, establishing an upper limit of approximately $1200 \, M_\odot$ with a $1 \sigma$ confidence level. This significantly improves our constraints on the mass distribution in the Galactic Center. Our upper limit is very close to the expected value from numerical simulations for a stellar cusp in the Galactic Center, leaving little room for a significant enhancement of dark matter density near Sagittarius A*., Comment: Submitted to A&A on September 17, 2024

Using a combination of HST, JWST, and ALMA data, we perform spatially resolved spectral energy distributions (SED) fitting of fourteen 4

This paper introduces a novel method for creating spectral time series, which can be used for generating synthetic light curves for photometric classification but also for applications like K-corrections and bolometric corrections. This approach is particularly valuable in the era of large astronomical surveys, where it can significantly enhance the analysis and understanding of an increasing number of SNe, even in the absence of extensive spectroscopic data. methods: By employing interpolations based on optimal transport theory, starting from a spectroscopic sequence, we derive weighted average spectra with high cadence. The weights incorporate an uncertainty factor for penalizing interpolations between spectra that show significant epoch differences and lead to a poor match between the synthetic and observed photometry. results: Our analysis reveals that even with phase difference of up to 40 days between pairs of spectra, optical transport can generate interpolated spectral time series that closely resemble the original ones. Synthetic photometry extracted from these spectral time series aligns well with observed photometry. The best results are achieved in the V band, with relative residuals of less than 10% for 87% and 84% of the data for type Ia and II, respectively. For the B, g, R and r bands, the relative residuals are between 65% and 87% within the previously mentioned 10% threshold for both classes. The worse results correspond to the i and I bands where, in the case, of SN~Ia the values drop to 53% and 42%, respectively. conclusions: We introduce a new method for constructing spectral time series for individual SNe starting from a sparse spectroscopic sequence, and demonstrate its capability to produce reliable light curves that can be used for photometric classification., Comment: Accepted for publication in A&A

Anti-viral therapies are typically designed to target only the current strains of a virus. Game theoretically, this corresponds to a short-sighted, or myopic, response. However, therapy-induced selective pressures act on viruses to drive the emergence of mutated strains, against which initial therapies have reduced efficacy. Building on a computational model of binding between antibodies and viral antigens (the Absolut! framework), we design and implement a genetic simulation of viral evolutionary escape. Crucially, this allows our antibody optimisation algorithm to consider and influence the entire escape curve of the virus, i.e. to guide (or "shape") the viral evolution. This is inspired by opponent shaping which, in general-sum learning, accounts for the adaptation of the co-player rather than playing a myopic best response. Hence we call the optimised antibodies shapers. Within our simulations, we demonstrate that our shapers target both current and simulated future viral variants, outperforming the antibodies chosen in a myopic way. Furthermore, we show that shapers exert specific evolutionary pressure on the virus compared to myopic antibodies. Altogether, shapers modify the evolutionary trajectories of viral strains and minimise the viral escape compared to their myopic counterparts. While this is a simplified model, we hope that our proposed paradigm will facilitate the discovery of better long-lived vaccines and antibody therapies in the future, enabled by rapid advancements in the capabilities of simulation tools. Our code is available at https://github.com/olakalisz/antibody-shapers., Comment: Preprint

We present a method for the additive fabrication of planar magnetic nanoarrays with minimal surface roughness. Synthesis is accomplished by combining electron-beam lithography, used to generate nanometric patterned masks, with ion implantation in thin films. By implanting $^{56}$Fe$^{+}$ ions, we are able to introduce magnetic functionality in a controlled manner into continuous Pd thin films, achieving 3D spatial resolution down to a few tens of nanometers. Our results demonstrate the application of this technique in fabricating square artificial spin ice lattices, which exhibit well-defined magnetization textures and interactions among the patterned magnetic elements., Comment: 5 pages, 4 figures, Supplemental Material

Large Language Models still struggle in challenging scenarios that leverage structured data, complex reasoning, or tool usage. In this paper, we propose Source2Synth: a new method that can be used for teaching LLMs new skills without relying on costly human annotations. Source2Synth takes as input a custom data source and produces synthetic data points with intermediate reasoning steps grounded in real-world sources. Source2Synth improves the dataset quality by discarding low-quality generations based on their answerability. We demonstrate the generality of this approach by applying it to two challenging domains: we test reasoning abilities in multi-hop question answering (MHQA), and tool usage in tabular question answering (TQA). Our method improves performance by 25.51% for TQA on WikiSQL and 22.57% for MHQA on HotPotQA compared to the fine-tuned baselines.

The interaction between surface acoustic waves and magnetization offers an efficient route for electrically controlling magnetic states. Here, we demonstrate the excitation of magnetoacoustic waves in galfenol, a highly magnetostrictive alloy made of iron (72%) and gallium (28%). We quantify the amplitude of the induced magnetization oscillations using magnetic imaging in an X-ray photoelectron microscope and estimate the dynamic magnetoelastic constants through micromagnetic simulations. Our findings demonstrate the potential of galfenol for magnonic applications and reveal that, despite strong magnetoelastic coupling, magnetic interactions and spin-wave dispersion relations significantly influence the overall amplitude of magnetoacoustic waves., Comment: 6 pages, 4 figures

Graph and network visualization supports exploration, analysis and communication of relational data arising in many domains: from biological and social networks, to transportation and powergrid systems. With the arrival of AI-based question-answering tools, issues of trustworthiness and explainability of generated answers motivate a greater role for visualization. In the context of graphs, we see the need for visualizations that can convince a critical audience that an assertion about the graph under analysis is valid. The requirements for such representations that convey precisely one specific graph property are quite different from standard network visualization criteria which optimize general aesthetics and readability. In this paper, we aim to provide a comprehensive introduction to visual proofs of graph properties and a foundation for further research in the area. We present a framework that defines what it means to visually prove a graph property. In the process, we introduce the notion of a visual certificate, that is, a specialized faithful graph visualization that leverages the viewer's perception, in particular, pre-attentive processing (e.g. via pop-out effects), to verify a given assertion about the represented graph. We also discuss the relationships between visual complexity, cognitive load and complexity theory, and propose a classification based on visual proof complexity. Finally, we provide examples of visual certificates for problems in different visual proof complexity classes., Comment: Appears in the Proceedings of the 32nd International Symposium on Graph Drawing and Network Visualization (GD 2024)

Human intelligence emerged through the process of natural selection and evolution on Earth. We investigate what it would take to re-create this process in silico. While past work has often focused on low-level processes (such as simulating physics or chemistry), we instead take a more targeted approach, aiming to evolve agents that can accumulate open-ended culture and technologies across generations. Towards this, we present JaxLife: an artificial life simulator in which embodied agents, parameterized by deep neural networks, must learn to survive in an expressive world containing programmable systems. First, we describe the environment and show that it can facilitate meaningful Turing-complete computation. We then analyze the evolved emergent agents' behavior, such as rudimentary communication protocols, agriculture, and tool use. Finally, we investigate how complexity scales with the amount of compute used. We believe JaxLife takes a step towards studying evolved behavior in more open-ended simulations. Our code is available at https://github.com/luchris429/JaxLife

MICADO is a first light instrument for the Extremely Large Telescope (ELT), set to start operating later this decade. It will provide diffraction limited imaging, astrometry, high contrast imaging, and long slit spectroscopy at near-infrared wavelengths. During the initial phase operations, adaptive optics (AO) correction will be provided by its own natural guide star wavefront sensor. In its final configuration, that AO system will be retained and complemented by the laser guide star multi-conjugate adaptive optics module MORFEO (formerly known as MAORY). Among many other things, MICADO will study exoplanets, distant galaxies and stars, and investigate black holes, such as Sagittarius A* at the centre of the Milky Way. After their final design phase, most components of MICADO have moved on to the manufacturing and assembly phase. Here we summarize the final design of the instrument and provide an overview about its current manufacturing status and the timeline. Some lessons learned from the final design review process will be presented in order to help future instrumentation projects to cope with the challenges arising from the substantial differences between projects for 8-10m class telescopes (e.g. ESO-VLT) and the next generation Extremely Large Telescopes (e.g. ESO-ELT). Finally, the expected performance will be discussed in the context of the current landscape of astronomical observatories and instruments. For instance, MICADO will have similar sensitivity as the James Webb Space Telescope (JWST), but with six times the spatial resolution., Comment: Proceedings of the SPIE, Volume 13096, id. 1309611 11 pp. (2024)

What data or environments to use for training to improve downstream performance is a longstanding and very topical question in reinforcement learning. In particular, Unsupervised Environment Design (UED) methods have gained recent attention as their adaptive curricula promise to enable agents to be robust to in- and out-of-distribution tasks. This work investigates how existing UED methods select training environments, focusing on task prioritisation metrics. Surprisingly, despite methods aiming to maximise regret in theory, the practical approximations do not correlate with regret but with success rate. As a result, a significant portion of an agent's experience comes from environments it has already mastered, offering little to no contribution toward enhancing its abilities. Put differently, current methods fail to predict intuitive measures of ``learnability.'' Specifically, they are unable to consistently identify those scenarios that the agent can sometimes solve, but not always. Based on our analysis, we develop a method that directly trains on scenarios with high learnability. This simple and intuitive approach outperforms existing UED methods in several binary-outcome environments, including the standard domain of Minigrid and a novel setting closely inspired by a real-world robotics problem. We further introduce a new adversarial evaluation procedure for directly measuring robustness, closely mirroring the conditional value at risk (CVaR). We open-source all our code and present visualisations of final policies here: https://github.com/amacrutherford/sampling-for-learnability.

We show that every planar graph has a monotone topological 2-page book embedding where at most (4n-10)/5 (of potentially 3n-6) edges cross the spine, and every edge crosses the spine at most once; such an edge is called a biarc. We can also guarantee that all edges that cross the spine cross it in the same direction (e.g., from bottom to top). For planar 3-trees we can further improve the bound to (3n-9)/4, and for so-called Kleetopes we obtain a bound of at most (n-8)/3 edges that cross the spine. The bound for Kleetopes is tight, even if the drawing is not required to be monotone. A Kleetope is a plane triangulation that is derived from another plane triangulation T by inserting a new vertex v_f into each face f of T and then connecting v_f to the three vertices of f., Comment: Appears in the Proceedings of the 32nd International Symposium on Graph Drawing and Network Visualization (GD 2024)

We present a study on the static magnetic properties of individual $\text{GaAs}$-$\text{Fe}_{\text{33}}\text{Co}_{\text{67}}$ core-shell nanorods. X-ray Magnetic Circular Dichroism combined with Photoemission Electron Microscopy (XMCD-PEEM) and Scanning Transmission X-ray Microscopy (STXM) were used to investigate the magnetic nanostructures. The magnetic layer is purposely designed to establish a magnetic easy axis neither along a high symmetry nor mirror axes to promote 3D magnetic helical order on the curved surface. In practice, two types of magnetic textures with in-plane magnetization were found inside the nanostructures' facets: magnetic domains with almost longitudinal or almost perpendicular magnetization with respect to the axis of the tube. We observe that a magnetic field applied perpendicular to the long axis of the nanostructure can add an azimuthal component of the magnetization to the previously almost longitudinal magnetization.

Analytical results on the correlation functions of strongly correlated many-body systems are rare in the literature and their importance cannot be overstated. We present determinant representations for the space-, time-, and temperature-dependent correlation functions of the strongly interacting one-dimensional Hubbard model in the presence of an external trapping potential. These representations are exact and valid in both equilibrium and nonequilibrium scenarios like the ones initiated by a sudden change of the confinement potential. In addition, they can be implemented numerically very easily significantly outperforming other numerical approaches. As applications of our results we investigate the single particle spectral functions of systems with harmonic trapping and show that dynamical quasicondensation occurs for both fermionic and bosonic spin-$1/2$ systems released from a Mott insulator state., Comment: 10 pages, 2 figures, RevTeX 4.2

The Mixture of Experts (MoE) framework has become a popular architecture for large language models due to its superior performance over dense models. However, training MoEs from scratch in a large-scale regime is prohibitively expensive. Existing methods mitigate this by pre-training multiple dense expert models independently and using them to initialize an MoE. This is done by using experts' feed-forward network (FFN) to initialize the MoE's experts while merging other parameters. However, this method limits the reuse of dense model parameters to only the FFN layers, thereby constraining the advantages when "upcycling" these models into MoEs. We propose BAM (Branch-Attend-Mix), a simple yet effective method that addresses this shortcoming. BAM makes full use of specialized dense models by not only using their FFN to initialize the MoE layers but also leveraging experts' attention parameters fully by initializing them into a soft-variant of Mixture of Attention (MoA) layers. We explore two methods for upcycling attention parameters: 1) initializing separate attention experts from dense models including all attention parameters for the best model performance; and 2) sharing key and value parameters across all experts to facilitate for better inference efficiency. To further improve efficiency, we adopt a parallel attention transformer architecture to MoEs, which allows the attention experts and FFN experts to be computed concurrently. Our experiments on seed models ranging from 590 million to 2 billion parameters demonstrate that BAM surpasses baselines in both perplexity and downstream task performance, within the same computational and data constraints.

One of the grand challenges of artificial general intelligence is developing agents capable of conducting scientific research and discovering new knowledge. While frontier models have already been used as aides to human scientists, e.g. for brainstorming ideas, writing code, or prediction tasks, they still conduct only a small part of the scientific process. This paper presents the first comprehensive framework for fully automatic scientific discovery, enabling frontier large language models to perform research independently and communicate their findings. We introduce The AI Scientist, which generates novel research ideas, writes code, executes experiments, visualizes results, describes its findings by writing a full scientific paper, and then runs a simulated review process for evaluation. In principle, this process can be repeated to iteratively develop ideas in an open-ended fashion, acting like the human scientific community. We demonstrate its versatility by applying it to three distinct subfields of machine learning: diffusion modeling, transformer-based language modeling, and learning dynamics. Each idea is implemented and developed into a full paper at a cost of less than $15 per paper. To evaluate the generated papers, we design and validate an automated reviewer, which we show achieves near-human performance in evaluating paper scores. The AI Scientist can produce papers that exceed the acceptance threshold at a top machine learning conference as judged by our automated reviewer. This approach signifies the beginning of a new era in scientific discovery in machine learning: bringing the transformative benefits of AI agents to the entire research process of AI itself, and taking us closer to a world where endless affordable creativity and innovation can be unleashed on the world's most challenging problems. Our code is open-sourced at https://github.com/SakanaAI/AI-Scientist

We use JWST/NIRSpec observations from the Assembly of Ultradeep Rest-optical Observations Revealing Astrophysics (AURORA) survey to constrain the shape of the nebular attenuation curve of a star-forming galaxy at z=4.41, GOODSN-17940. We utilize 11 unblended HI recombination lines to derive the attenuation curve spanning optical to near-infrared wavelengths (3751-9550 \r{A}). We then leverage a high-S/N spectroscopic detection of the rest-frame ultraviolet continuum in combination with rest-UV photometric measurements to constrain the shape of the curve at ultraviolet wavelengths. While this UV constraint is predominantly based on stellar emission, the large measured equivalent widths of H$\alpha$ and H$\beta$ indicate that GOODSN-17940 is dominated by an extremely young stellar population <10 Myr in age such that the UV stellar continuum experiences the same attenuation as the nebular emission. The resulting combined nebular attenuation curve spans 1400-9550 \r{A} and has a shape that deviates significantly from commonly assumed dust curves in high-redshift studies. Relative to the Milky Way, SMC, and Calzetti curves, the new curve has a steeper slope at long wavelengths ($\lambda>5000$ \r{A}) while displaying a similar slope across blue-optical wavelengths ($\lambda=3750-5000$ \r{A}). In the ultraviolet, the new curve is shallower than the SMC and Calzetti curves and displays no significant 2175 \r{A} bump. This work demonstrates that the most commonly assumed dust curves are not appropriate for all high-redshift galaxies. These results highlight the ability to derive nebular attenuation curves for individual high-redshift sources with deep JWST/NIRSpec spectroscopy, thereby improving the accuracy of physical properties inferred from nebular emission lines., Comment: 21 pages, 6 figures, 1 table, submitted to ApJ

We present the design and observations of low resolution JWST/NIRSpec PRISM spectroscopy from the Ultradeep NIRSpec and NIRCam ObserVations before the Epoch of Reionization (UNCOVER) Cycle 1 JWST Treasury program. Targets are selected using JWST/NIRCam photometry from UNCOVER and other programs, and cover a wide range of categories and redshifts to ensure the legacy value of the survey. These categories include the first galaxies at $z\gtrsim10$, faint galaxies during the Epoch of Reionization ($z\sim6-8$), high redshift AGN ($z\gtrsim6$), Population III star candidates, distant quiescent and dusty galaxies ($1\lesssim z \lesssim 6$), and filler galaxies sampling redshift--color--magnitude space from $z\sim 0.1-13$. Seven NIRSpec MSA masks across the extended Abell 2744 cluster were observed, along with NIRCam parallel imaging in 8 filters (F090W, F115W, F150W, F200W, F277W, F356W, F410M, F444W, F480M) over a total area of ~26 arcmin$^2$, overlapping existing HST coverage from programs including the Hubble Frontier Fields and BUFFALO. We successfully observed 553 objects down to $m_{\mathrm{F444W}}\sim30\mathrm{AB}$, and by leveraging mask overlaps, we reach total on-target exposure times ranging from 2.4-16.7h. We demonstrate the success rate and distribution of confirmed redshifts, and also highlight the rich information revealed by these ultradeep spectra for a subset of our targets. An updated lens model of Abell 2744 is also presented, including 14 additional spectroscopic redshifts and finding a total cluster mass of $M_{\mathrm{SL}}=(2.1\pm0.3)\times10^{15}\,\mathrm{M}_{\odot}$. We publicly release reduced 1D and 2D spectra for all objects observed in Summer 2023 along with a spectroscopic redshift catalog and the updated lens model of the cluster (https://jwst-uncover.github.io/DR4.html)., Comment: 19 pages, 9 figures, 4 tables, submitted to ApJ, comments welcome! Data available at: https://jwst-uncover.github.io/DR4.html (v2: figure format correction)

Galaxy scaling relations provide insights into the processes that drive galaxy evolution. The extension of these scaling relations into the dwarf galaxy regime is of particular interest. This is because dwarf galaxies represent a crucial stage in galaxy evolution, and understanding them could also shed light on their role in reionising the early Universe. There is currently no consensus on the processes that dominate the evolution of dwarfs. In this work we constrain the atomic gas sequence (stellar mass vs. atomic gas fraction) and mass-metallicity relation (stellar mass vs. gas phase metallicity) from dwarf ($10^{6.5}$ $\textrm{M}_{\odot}$) to massive ($10^{11.5}$ $\textrm{M}_{\odot}$) galaxies in the local Universe. The combined optical and 21-cm spectroscopic observations of the DESI and ALFALFA surveys allow us to simultaneously constrain both scaling relations. We find a slope change of the atomic gas sequence at a stellar mass of $\sim 10^{9} ~\textrm{M}_{\odot}$. We also find that the shape and scatter of the atomic gas sequence and mass-metallicity relation are strongly linked for both dwarfs and more massive galaxies. Consequently, the low mass slope change of the atomic gas sequence is imprinted onto the mass-metallicity relation of dwarf galaxies. The mass scale of the measured slope change is consistent with a predicted escape velocity threshold below which low mass galaxies experience significant supernova-driven gas loss, as well as with a reduction in cold gas accretion onto more massive galaxies., Comment: 16 pages, 10 figures, submitted to MNRAS

We investigate the spatial extent of the [CII] line emission in a sample of 34 galaxies at $z=4-6$ from the ALMA-CRISTAL Survey. By modeling the [CII] line emission in the visibility data directly, we derive the effective radius of [CII] line emission assuming exponential distribution. These measurements comprise not only isolated galaxies but also interacting systems, identified thanks to the high spatial resolution of the data. The [CII] line radius ranges from 0.5 to 3.5 kpc with an average value of 1.9 kpc. We compare the [CII] sizes with the sizes of UV and FIR continua, which were measured from the HST F160W and ALMA Band-7 continuum images, respectively. We confirm that the [CII] line emission is more spatially extended than the continuum emission, with average size ratios of $R_{e,[CII]}/R_{e,UV}=2.90$ and $R_{e,[CII]}/R_{e,FIR}=1.54$, although about half of the FIR-detected sample show comparable spatial extent between [CII] line and FIR continuum emission ($R_{e,[CII]}\approx R_{e, FIR}$). The residual visibility data of the best-fit model do not show evidence of flux excesses either individually or in stacking analysis. This indicates that the [CII] line emission in star-forming galaxies can be characterized by an extended exponential disk profile. Overall, our results suggest that the spatial extent of [CII] line emission can primarily be explained by photodissociation regions associated with star formation activity, while the contribution from diffuse neutral medium (atomic gas) and the effects of mergers may further expand the [CII] line distributions, causing their variations among our sample. We report the correlations between the [CII] line, dust, and Lya line properties, which may be in line with our scenario. Future 3D-analysis of Lya and Ha lines will shed light on the association of the extended [CII] line emission with atomic gas and outflows., Comment: Submitted to A&A, 21 pages, 14 figures, 4 tables

The dust- and gas-rich protoplanetary disks around young stellar systems play a key role in star and planet formation. While considerable progress has recently been made in probing these disks on large scales of a few tens of astronomical units (au), the central au needs to be more investigated. We aim at unveiling the physical processes at play in the innermost regions of the strongly accreting T Tauri Star S CrA N by means of near-infrared interferometric observations. The K-band continuum emission is well reproduced with an azimuthally-modulated dusty ring. As the star alone cannot explain the size of this sublimation front, we propose that magnetospheric accretion is an important dust-heating mechanism leading to this continuum emission. The differential analysis of the Hydrogen Br$\gamma$ line is in agreement with radiative transfer models combining magnetospheric accretion and disk winds. Our observations support an origin of the Br$\gamma$ line from a combination of (variable) accretion-ejection processes in the inner disk region.

We present new VLTI/GRAVITY near-infrared interferometric measurements of the angular size of the innermost hot dust continuum for 14 type 1 AGNs. The angular sizes are resolved on scales of ~0.7 mas and the inferred ring radii range from 0.028 to 1.33 pc, comparable to those reported previously and a factor 10-20 smaller than the mid-infrared sizes in the literature. Combining our new data with previously published values, we compile a sample of 25 AGN with bolometric luminosity ranging from $10^{42}$ to $10^{47} \rm erg~s^{-1}$, with which we study the radius-luminosity (R-L) relation for the hot dust structure. Our interferometric measurements of radius are offset by a factor 2 from the equivalent relation derived through reverberation mapping. Using a simple model to explore the dust structure's geometry, we conclude that this offset can be explained if the 2 um emitting surface has a concave shape. Our data show that the slope of the relation is in line with the canonical $R \propto L^{0.5}$ when using an appropriately non-linear correction for bolometric luminosity. In contrast, using optical luminosity or applying a constant bolometric correction to it results in a significant deviation in the slope, suggesting a potential luminosity dependence on the spectral energy distribution. Over four orders of magnitude in luminosity, the intrinsic scatter around the R-L relation is 0.2 dex, suggesting a tight correlation between innermost hot dust structure size and the AGN luminosity., Comment: 14 pages, 5 figures, 4 tables, accepted for publication in A&A

We present new JWST/NIRSpec IFU observations of the J1000+0234 system at $z=4.54$, the dense core of a galaxy protocluster hosting a massive, dusty star forming galaxy (DSFG) with a low luminosity radio counterpart. The new data reveals two extended, high equivalent width (EW$_0 > 1000$ {\AA}) nebulae at each side of the DSFG disk along its minor axis (namely O3-N and O3-S). On one hand, O3-N's spectrum shows a prominent FWHM $\sim1300$ km s$^{-1}$ broad and blueshifted component, suggesting an outflow origin. On the other hand, O3-S stretches over parsec and has a velocity gradient that spans $800$ km s$^{-1}$ but no evidence of a broad component. Both sources, however, seem to be powered at least partially by an active galactic nucleus (AGN), so we classify them as extended emission-line regions (EELRs). The strongest evidence comes from the detection of the high-ionization [Ne V] $\lambda3427$ line toward O3-N, which paired with the non-detection of hard X-rays implies an obscuring column density above the Compton-thick regime. In O3-S, the [Ne V] line is not detected, but we measure a He II well above the expectation for star formation. We interpret this as O3-S being externally irradiated by the AGN, akin to the famous Hanny's Voorwerp object in the local Universe. In addition, more classical line ratio diagnostics (e.g. [O III]/H$\beta$ vs [N II]/H$\alpha$) put the DSFG itself in the AGN region of the diagrams, and hence the most probable host of the AGN. These results showcase the ability of JWST of unveiling highly obscured AGN at high redshifts., Comment: 5 pages, 4 figures plus 5 appendices (incl. 3 extra figures and one table). Submitted to A&A on July 17th 2024

We present new $\lambda_{\rm rest}=77$ $\mu$m dust continuum observations from the ALMA of HZ10 (CRISTAL-22), a dusty main-sequence galaxy at $z$=5.66 as part of the [CII] Resolved Ism in STar-forming Alma Large program, CRISTAL. The high angular resolution of the ALMA Band 7 and new Band 9 data($\sim{0}''.4$) reveals the complex structure of HZ10, which comprises two main components (HZ10-C and HZ10-W) and a bridge-like dusty emission between them (the Bridge). We model the dust spectral energy distribution (SED) to constrain the physical conditions of the interstellar medium (ISM) and its variations among the different components identified in HZ10. We find that HZ10-W (the more UV-obscured component) has an SED dust temperature of $T_{\rm SED}$$\sim$51.2$\pm13.1$ K; this is $\sim$5 K higher (although still consistent) than that of the central component and previous global estimations for HZ10. Our new ALMA data allow us to reduce by a factor of $\sim$2.3 the uncertainties of global $T_{\rm SED}$ measurements compared to previous studies. Interestingly, HZ10-W shows a lower [CII]/FIR ratio compared to the other two components (although still within the uncertainties), suggesting a harder radiation field destroying polycyclic aromatic hydrocarbon associated with [CII] emission (e.g., active galactic nuclei or young stellar populations). While HZ10-C appears to follow the tight IRX-$\beta_{\rm UV}$ relation seen in local UV-selected starburst galaxies and high-$z$ star-forming galaxies, we find that both HZ10-W and the Bridge depart from this relation and are well described by dust-screen models with holes in front of a hard UV radiation field. This suggests that the UV emission (likely from young stellar populations) is strongly attenuated in the more dusty components of the HZ10 system.

While reinforcement learning (RL) holds great potential for decision making in the real world, it suffers from a number of unique difficulties which often need specific consideration. In particular: it is highly non-stationary; suffers from high degrees of plasticity loss; and requires exploration to prevent premature convergence to local optima and maximize return. In this paper, we consider whether learned optimization can help overcome these problems. Our method, Learned Optimization for Plasticity, Exploration and Non-stationarity (OPEN), meta-learns an update rule whose input features and output structure are informed by previously proposed solutions to these difficulties. We show that our parameterization is flexible enough to enable meta-learning in diverse learning contexts, including the ability to use stochasticity for exploration. Our experiments demonstrate that when meta-trained on single and small sets of environments, OPEN outperforms or equals traditionally used optimizers. Furthermore, OPEN shows strong generalization across a distribution of environments and a range of agent architectures., Comment: AutoRL Workshop at ICML 2024

Q-learning played a foundational role in the field reinforcement learning (RL). However, TD algorithms with off-policy data, such as Q-learning, or nonlinear function approximation like deep neural networks require several additional tricks to stabilise training, primarily a replay buffer and target networks. Unfortunately, the delayed updating of frozen network parameters in the target network harms the sample efficiency and, similarly, the replay buffer introduces memory and implementation overheads. In this paper, we investigate whether it is possible to accelerate and simplify TD training while maintaining its stability. Our key theoretical result demonstrates for the first time that regularisation techniques such as LayerNorm can yield provably convergent TD algorithms without the need for a target network, even with off-policy data. Empirically, we find that online, parallelised sampling enabled by vectorised environments stabilises training without the need of a replay buffer. Motivated by these findings, we propose PQN, our simplified deep online Q-Learning algorithm. Surprisingly, this simple algorithm is competitive with more complex methods like: Rainbow in Atari, R2D2 in Hanabi, QMix in Smax, PPO-RNN in Craftax, and can be up to 50x faster than traditional DQN without sacrificing sample efficiency. In an era where PPO has become the go-to RL algorithm, PQN reestablishes Q-learning as a viable alternative.

We present the stellar metallicities and multi-element abundances (C, Mg, Si, Ca, Ti, Cr, and Fe) of 15 massive (log $M/M_\odot=10.2-11.2$) quiescent galaxies at z=1-3, derived from ultradeep JWST-SUSPENSE spectra. Compared to quiescent galaxies at z~0, these galaxies exhibit a deficiency of 0.26$\pm0.04$ dex in [C/H], 0.16$\pm0.03$ dex in [Fe/H], and 0.07$\pm0.04$ dex in [Mg/H], implying rapid formation and quenching before significant enrichment from asymptotic giant branch stars and Type Ia supernovae. Additionally, we find that galaxies forming at higher redshift consistently show higher [Mg/Fe] and lower [Fe/H] and [Mg/H], regardless of their observed redshift. The evolution in [Fe/H] and [C/H] is therefore primarily driven by lower-redshift samples naturally including galaxies with longer star-formation timescales. In contrast, the lower [Mg/H] likely reflects earlier-forming galaxies expelling larger gas reservoirs during their quenching phase. Consequently, the mass-metallicity relation, primarily reflecting [Mg/H], is somewhat lower at z=1-3 compared to the lower redshift relation. Finally, we compare our results to standard stellar population modeling approaches employing solar abundance patterns and non-parametric star-formation histories (using Prospector). Our SSP-equivalent ages agree with the mass-weighted ages from Prospector, while the metallicities disagree significantly. Nonetheless, the metallicities better reflect [Fe/H] than total [Z/H]. We also find that star-formation timescales inferred from elemental abundances are significantly shorter than those from Prospector, and we discuss the resulting implications for the early formation of massive galaxies., Comment: Accepted to ApJ; 22 pages, 6 figures, 3 tables

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., Comment: 17 pages, 8 figures, submitted to ApJ

Mixtures of Experts (MoEs) have gained prominence in (self-)supervised learning due to their enhanced inference efficiency, adaptability to distributed training, and modularity. Previous research has illustrated that MoEs can significantly boost Deep Reinforcement Learning (DRL) performance by expanding the network's parameter count while reducing dormant neurons, thereby enhancing the model's learning capacity and ability to deal with non-stationarity. In this work, we shed more light on MoEs' ability to deal with non-stationarity and investigate MoEs in DRL settings with "amplified" non-stationarity via multi-task training, providing further evidence that MoEs improve learning capacity. In contrast to previous work, our multi-task results allow us to better understand the underlying causes for the beneficial effect of MoE in DRL training, the impact of the various MoE components, and insights into how best to incorporate them in actor-critic-based DRL networks. Finally, we also confirm results from previous work.