Search

We introduce the THESAN-ZOOM project, a comprehensive suite of high-resolution zoom-in simulations of $14$ high-redshift ($z>3$) galaxies selected from the THESAN simulation volume. This sample encompasses a diverse range of halo masses, with $M_\mathrm{halo} \approx 10^8 - 10^{13}~\mathrm{M}_\odot$ at $z=3$. At the highest-resolution, the simulations achieve a baryonic mass of $142~\mathrm{M}_\odot$ and a gravitational softening length of $17~\mathrm{cpc}$. We employ a state-of-the-art multi-phase interstellar medium (ISM) model that self-consistently includes stellar feedback, radiation fields, dust physics, and low-temperature cooling through a non-equilibrium thermochemical network. Our unique framework incorporates the impact of patchy reionization by adopting the large-scale radiation field topology from the parent THESAN simulation box rather than assuming a spatially uniform UV background. In total, THESAN-ZOOM comprises $60$ simulations, including both fiducial runs and complementary variations designed to investigate the impact of numerical and physical parameters on galaxy properties. The fiducial simulation set reproduces a wealth of high-redshift observational data such as the stellar-to-halo-mass relation, the star-forming main sequence, the Kennicutt-Schmidt relation, and the mass-metallicity relation. While our simulations slightly overestimate the abundance of low-mass and low-luminosity galaxies they agree well with observed stellar and UV luminosity functions at the higher mass end. Moreover, the star-formation rate density closely matches the observational estimates from $z=3-14$. These results indicate that the simulations effectively reproduce many of the essential characteristics of high-redshift galaxies, providing a realistic framework to interpret the exciting new observations from JWST., Comment: 25 pages, 16 figures, Please visit https://thesan-project.com for more details, Comments welcome

We determine forest lease value and optimal harvesting strategies under model parameter uncertainty within stochastic bio-economic models that account for catastrophe risk. Catastrophic events are modeled as a Poisson point process, with a two-factor stochastic convenience yield model capturing the lumber spot price dynamics. Using lumber futures and US wildfire data, we estimate model parameters through a Kalman filter and maximum likelihood estimation and define the model parameter uncertainty set as the 95% confidence region. We numerically determine the forest lease value under catastrophe risk and parameter uncertainty using reflected backward stochastic differential equations (RBSDEs) and establish conservative and optimistic bounds for lease values and optimal stopping boundaries for harvesting, facilitating Monte Carlo simulations. Numerical experiments further explore how parameter uncertainty, catastrophe intensity, and carbon sequestration impact the lease valuation and harvesting decision. In particular, we explore the costs arising from this form of uncertainty in the form of a reduction of the lease value. These are implicit costs that can be attributed to climate risk and will be emphasized through the importance of forestry resources in the energy transition process. We conclude that in the presence of parameter uncertainty, it is better to lean toward a conservative strategy reflecting, to some extent, the worst case than being overly optimistic. Our results also highlight the critical role of convenience yield in determining optimal harvesting strategies.

The shape of the Ly-$\alpha$ transmission in the near zone of the redshift $z=5.9896$ quasar ULAS J0148$+$0600 (hereafter J0148) is consistent with a damping wing arising from an extended neutral hydrogen island in the diffuse intergalactic medium (IGM). Here we use simulations of late-ending reionisation from Sherwood-Relics to assess the expected incidence of quasars with Ly-$\alpha$ and Ly-$\beta$ absorption similar to the observed J0148 spectrum. We find a late end to reionisation at $z=5.3$ is a necessary requirement for reproducing a Ly-$\alpha$ damping wing consistent with J0148. This occurs in $\sim3$ per cent of our simulated spectra for an IGM neutral fraction $\langle x_{\rm HI}\rangle=0.14$ at $z=6$. However, using standard assumptions for the ionising photon output of J0148, the a priori probability of drawing a simulated quasar spectrum with a Ly-$\alpha$ damping wing profile and Ly-$\alpha$ near zone size that simultaneously match J0148 is very low, $p<10^{-3}$. We speculate this is because the ionising emission from J0148 is variable on timescales $t<10^{5}\rm\,yr$, or alternatively that the Ly-$\alpha$ transmission in the J0148 near zone is impacted by the transverse proximity effect from nearby star-forming galaxies or undetected quasars. We also predict the IGM temperature should be $T\sim 4\times 10^{4}\rm\,K$ within a few proper Mpc of the Ly-$\alpha$ near zone edge due to recent HI and HeII photo-heating. Evidence for enhanced thermal broadening in the Ly-$\alpha$ absorption near the damping wing edge would provide further evidence that the final stages of reionisation are occurring at $z<6$., Comment: 12 pages, 6 figures. Submitted to MNRAS

Recent work has suggested that, during reionisation, spatial variations in the ionising radiation field should produce enhanced Ly ${\alpha}$ forest transmission at distances of tens of comoving Mpc from high-redshift galaxies. We demonstrate that the Sherwood-Relics suite of hybrid radiation-hydrodynamical simulations are qualitatively consistent with this interpretation. The shape of the galaxy--Ly ${\alpha}$ transmission cross-correlation is sensitive to both the mass of the haloes hosting the galaxies and the volume averaged fraction of neutral hydrogen in the IGM, $\bar{x}_{\rm HI}$. The reported excess Ly ${\alpha}$ forest transmission on scales r ~ 10 cMpc at $\langle z \rangle \approx 5.2$ -- as measured using C IV absorbers as proxies for high-redshift galaxies -- is quantitatively reproduced by Sherwood-Relics at z = 6 if we assume the galaxies that produce ionising photons are hosted in haloes with mass $M_{\rm h}\geq 10^{10}~h^{-1}\,{\rm M}_\odot$. However, this redshift mismatch is equivalent to requiring $\bar{x}_{\rm HI}\sim 0.1$ at $z\simeq 5.2$, which is inconsistent with the observed Ly ${\alpha}$ forest effective optical depth distribution. We speculate this tension may be partly resolved if the minimum C IV absorber host halo mass at z > 5 is larger than $M_{\rm h}=10^{10}~h^{-1}\,{\rm M}_\odot$. After reionisation completes, relic IGM temperature fluctuations will continue to influence the shape of the cross-correlation on scales of a few comoving Mpc at $4 \leq z \leq 5$. Constraining the redshift evolution of the cross-correlation over this period may therefore provide further insight into the timing of reionisation., Comment: 15 pages, 14 figures. Submitted to MNRAS

Interfacing light from solid-state single-photon sources with scalable and robust room-temperature quantum memories has been a long-standing challenge in photonic quantum information technologies due to inherent noise processes and time-scale mismatches between the operating conditions of solid-state and atomic systems. Here, we demonstrate on-demand storage and retrieval of single photons from a semiconductor quantum dot device in a room-temperature atomic vapor memory. A deterministically fabricated InGaAs quantum dot light source emits single photons at the wavelength of the cesium D1 line at 895\,nm which exhibit an inhomogeneously broadened linewidth of 5.1(7)\,GHz and are subsequently stored in a low-noise ladder-type cesium vapor memory. We show control over the interaction between the single photons and the atomic vapor, allowing for variable retrieval times of up to 19.8(3)\,ns at an internal efficiency of $\eta_\mathrm{int}=0.6(1)\%$. Our results significantly expand the application space of both room-temperature vapor memories and semiconductor quantum dots in future quantum network architectures.

We present the first constraints on primordial magnetic fields from the Lyman-$\alpha$ forest using full cosmological hydrodynamic simulations. At the scales and redshifts probed by the data, the flux power spectrum is extremely sensitive to the extra power induced by primordial magnetic fields in the linear matter power spectrum, at a scale that we parametrize with $k_{\rm peak}$. We rely on a set of more than a quarter million flux models obtained by varying thermal, reionization histories and cosmological parameters. We find a hint of extra power that is well fitted by the PMF model with $B\sim 0.2$ nG, corresponding to $k_{\rm peak}\sim 20$ Mpc$^{-1}$. However, when applying very conservative assumptions on the modelling of the noise, we obtain a 3$\sigma$ C.L. lower limit $k_{\rm peak}> 30$ Mpc$^{-1}$ which translates into the tightest bounds on the strength of primordial intergalactic magnetic fields: $B < 0.30$ nG (for fixed, nearly scale-invariant $n_{\rm B}=-2.9$)., Comment: 5 pages + Supplemental material

In quantum computing systems the quantum states of qubits can be modified among others by applying light pulses. In order to achieve low computing error rates these pulses have to be precisely shaped in magnitude and phase. In practical applications, both acousto-optic (AOM) and electro-optic modulators (EOM) are used for this purpose. The advantages of EOMs, in particular Mach-Zehnder modulators (MZMs), include e.g. higher bandwidth, compactness, good integrability and better noise performance. On the other hand, EOMs are challenging regarding their control voltage regulation as they experience a bias drift, i.e. voltage shifts in the modulator's operating point. Such a shift significantly impairs the quality of the modulation, which is why EOMs usually require a bias control loop to track the DC operating point. This work addresses the estimation of the instantaneous bias voltage using a small pilot tone, where the optical output power of the EOM is tracked by a photodetector feedback signal.

We aim to construct a machine-learning approach that allows for a pixel-by-pixel reconstruction of the intergalactic medium (IGM) density field for various warm dark matter (WDM) models using the Lyman-alpha forest. With this regression machinery, we constrain the mass of a potential WDM particle from observed Lyman-alpha sightlines directly from the density field. We design and train a Bayesian neural network on the supervised regression task of recovering the optical depth-weighted density field $\Delta_\tau$ as well as its reconstruction uncertainty from the Lyman-alpha forest flux field. We utilise the Sherwood-Relics simulation suite at $4.1\leq z \leq 5.0$ as the main training and validation dataset. Leveraging the density field recovered by our neural network, we construct an inference pipeline to constrain the WDM particle masses based on the probability distribution function of the density fields. We find that our trained Bayesian neural network can accurately recover within a $1\sigma$ error $\geq 85\%$ of the density field pixels from a validation simulated dataset that encompasses multiple WDM and thermal models of the IGM. When predicting on Lyman-alpha skewers generated using the alternative hydrodynamical code Nyx not included in the training data, we find a $1\sigma$ accuracy rate $\geq 75\%$. We consider 2 samples of observed Lyman-alpha spectra from the UVES and GHOST instruments, at $z=4.4$ and $z=4.9$ respectively and fit the density fields recovered by our Bayesian neural network to constrain WDM masses. We find lower bounds on the WDM particle mass of $m_{\mathrm{WDM}} \geq 3.8$ KeV and $m_{\mathrm{WDM}} \geq 2.2$ KeV at $2\sigma$ confidence, respectively. We are able to match current state-of-the-art WDM particle mass constraints using up to 40 times less observational data than Markov Chain Monte Carlo techniques based on the Lyman-alpha forest power spectrum., Comment: 12 pages, Submitted to A&A

We present a suite of zoom-in cosmological simulations of Milky Way-like galaxies with a prominent disc component and a strong bar in their centre, based on a subsample of barred galaxies from the TNG50 magneto-hydrodynamic simulation. We modify the physical models that regulate star formation, namely, supernova feedback and black hole quasar feedback, to examine how they affect the disc and bar formation. We find that the morphology remains consistent in all the galaxies that develop a massive stellar disc ($>10^{10}\Msun$), which is dominant in comparison with the bulge mass. The black hole quasar feedback models used in this work do not affect bar formation, although they can affect the properties of the bar. The energy released by the supernovae causes a delay in the time of bar formation and, in models with strong feedback, stops bar formation. This could be understood since supernova feedback affects the assembly of the disc and bulge and their structural properties, such as the mass content, size, and radial velocity dispersion. We compared our predictions to three bar instability criteria proposed in the literature. We find that galaxies with varying supernovae and black hole quasar feedback satisfy these criteria at the moment of bar formation, except in extreme cases where the galaxy does not have or has weak supernova feedback, where for some galaxies, they fail to predict the existence (or no existence) of bars. Our findings provide insights into the physical process behind bar formation while putting constraints on the analytic prescriptions that predict bar formation., Comment: 21 pages, 11 figures, submitted to A&A. Comments are welcome

Observations of Faraday rotation and synchrotron emission in galaxy clusters imply large-scale magnetic fields with $\mu\mathrm{G}$ strengths possibly extending back to $z=4$. Non-radiative cosmological simulations of galaxy clusters show a comparably slow magnetic field growth that only saturates at late times. We include galaxy formation physics and find a significantly accelerated magnetic field growth. We identify three crucial stages in the magnetic field evolution. 1) At high redshift, the central dominant galaxy serves as the prime agent that magnetizes not only its immediate vicinity but also most of the forming protocluster through a combination of a small-scale dynamo induced by gravitationally driven, compressive turbulence and stellar and active galactic nuclei feedback that distribute the magnetic field via outflows. 2) This process continues as other galaxies merge into the forming cluster in subsequent epochs, thereby transporting their previously amplified magnetic field to the intracluster medium (ICM) through ram pressure stripping and galactic winds. 3) At lower redshift, gas accretion and frequent cluster mergers trigger additional small-scale dynamo processes, thereby preventing the decay of the magnetic field and fostering the increase of the magnetic coherence scale. We show that the magnetic field observed today in the weakly collisional ICM is consistently amplified on collisional scales. Initially, this occurs in the collisional interstellar medium during protocluster assembly, and later in the ICM on the magnetic coherence scale, which always exceeds the particle mean free path, supporting the use of magneto-hydrodynamics for studying the cluster dynamo. We generate synthetic Faraday rotation measure observations of protoclusters, highlighting the potential for studying magnetic field growth during the onset of cluster formation at cosmic dawn., Comment: 23 pages, 15 figures, submitted to A&A

We assess the impact of QSOs on the high redshift (z > 4) Intergalactic Medium using Monte Carlo realisations of QSO populations and the HeIII regions they generate, applied to the Sherwood-Relics simulations, allowing for uncertainties in the QSO luminosity function, its evolution, and QSO spectra and ages. While QSO luminosity functions based on optical-infrared selection are unable to reproduce the broadening HI Lyman-alpha optical depth distributions at z > 5, much broader distributions are found for the higher numbers of QSOs based on x-ray selection, suggesting a large QSO contribution to the ultra-violet background at z > 5 may offer an alternative to late reionization models to account for the broad HI Lyman-alpha optical depth distributions. Realisations using QSOs based on the higher QSO counts also much better recover the measured pixel flux auto-correlation function at z > 5. The HeIII regions from QSO sources according to both types of luminosity function suppress the pixel flux power spectrum on small scales, k > 0.02 s/km, while enhancing it on larger, both by amounts of up to tens of percent at z > 4, with the magnitude increasing with the intergalactic HeIII filling factor and the boost in temperature within the HeIII regions., Comment: Accepted for publication in MNRAS; 10 pages, 33 figures

Augmented Reality (AR) and Virtual Reality (VR), require miniaturized displays with ultrahigh pixel densities. Here, we demonstrate an individually addressable subwavelength OLED pixel based on a nanoscale electrode capable of supporting plasmonic modes. Our approach is based on the notion that when scaling down pixel size, the 2D planar geometry of conventional organic light-emitting diodes (OLEDs) evolves into a significantly more complex 3D geometry governed by sharp nanoelectrode contours. These cause (i) spatially imbalanced charge carrier transport and recombination, resulting in a low quantum efficiency, and (ii) filament growth, leading to rapid device failure. Here, we circumvent such effects by selectively covering sharp electrode contours with an insulating layer, while utilizing a nano-aperture in flat areas of the electrode. We thereby ensure controlled charge carrier injection and recombination at the nanoscale and suppress filament growth. As a proof of principle, we first demonstrate stable and efficient hole injection from Au nanoelectrodes in hole-only devices with above 90 % pixel yield and longtime operation stability and then a complete vertical OLED pixel with an individually addressable nanoelectrode (300 x 300 nm$^{2}$), highlighting the potential to further leverage plasmonic nanoantenna effects to enhance the performance and functionality of nano-OLEDs.

We investigate early, $z > 3$, galaxy formation in a cosmological zoom-in simulation of a close, early-forming Milky Way (MW) analog extracted from TNG50 simulation and re-simulated with detailed modeling of cold interstellar medium (ISM) formation, coupled with on-the-fly UV radiative transfer, turbulence-regulated star formation, and stellar feedback. In our enhanced-physics simulation, the galaxy develops a bi-stable ISM structure (warm, with $T \sim 10^4$ K, and cold, with $T < 100$ K) and exhibits significantly more efficient, early, and bursty star formation than in TNG. Notably, the stellar disk of this MW progenitor forms extremely early, around $z\sim6-7$, and exhibits chemo-kinematic properties consistent with the low-metallicity population of the MW stars. The disk forms rapidly, on a timescale of $\sim$0.2 Gyr which is significantly shorter than the timescale implied by the observable chemo-kinematic signatures of disk spinup, $\sim$0.7 Gyr, due to the scatter in the age-metallicity relation. The rotational support of the gas disk and the location of the galaxy on the main sequence are consistent with early disk galaxies observed by JWST and ALMA at $z\sim4-7$, suggesting that some of these galaxies could be progenitors of MW-like systems. Remarkably, the variation of the global star formation rate (SFR) before disk formation is similar to the observed SFR scatter at these early times. Our findings underscore the critical role of modeling a turbulent cold ISM and turbulence-regulated star formation and feedback in driving early SFR variability, while at the same time enabling early disk formation, without destroying it with overly efficient stellar feedback., Comment: 17 pages + appendix, 12 figures, 2 tables. To be submitted. Comments are welcome

The relative contribution of emission from stellar sources and accretion onto supermassive black holes to reionization has been brought into focus again by the apparent high abundance of faint AGN at $4\lesssim z\lesssim11$ uncovered by JWST. We investigate the contribution of these faint AGN to hydrogen and the early stages of helium reionization using the GPU-based radiative transfer code ATON-HE by post-processing a cosmological hydrodynamical simulation from the SHERWOOD-RELICS suite of simulations. We study four models of reionization: two previously studied galaxy-only late-end reionization models and two new models -- a QSO-assisted model and a QSO-only model. In the QSO-assisted model, 1\% of the haloes host AGN with a 10~Myr lifetime, and the AGN luminosities are scaled such that the AGN contribution to the hydrogen-ionizing emissivity is 20\% of that contributed by galaxies. In the QSO-only model, quasars account for all the hydrogen-ionizing emissivity, with 10\% of the haloes hosting AGN, each with a 10 Myr lifetime. All models are calibrated to the observed mean Lyman-$\alpha$ forest transmission at $5\lesssim z\lesssim6.2$. We find that the QSO-assisted model requires an emissivity factor of $1.8$ lower than the galaxy-only models towards the end of reionization and fits the observed distribution of the Lyman-$\alpha$ optical depths well. Our QSO-only model is inconsistent with the observed Lyman-$\alpha$ optical depths distribution. It also results in too high IGM temperatures at $z\lesssim 5$ due to an early onset of HeII reionization unless the escape fraction of HeII-ionizing photons is assumed to be low. Our results suggest that a modest contribution to reionization by faint AGN is in good agreement with the Lyman-$\alpha$ forest data. In contrast, a scenario dominated by faint AGN appears difficult to reconcile with these observations., Comment: 19 pages, 11 figures

We prove that the usual gradient flow in parameter space that underlies many training algorithms for neural networks in deep learning can be continuously deformed into an adapted gradient flow which yields (constrained) Euclidean gradient flow in output space. Moreover, if the Jacobian of the outputs with respect to the parameters is full rank (for fixed training data), then the time variable can be reparametrized so that the resulting flow is simply linear interpolation, and a global minimum can be achieved.

To autonomously collect soil in uncultivated terrain, robotic arms must distinguish between different amorphous materials and submerge themselves into the correct material. We develop a prototype that collects soil in heterogeneous terrain. If mounted to a mobile robot, it can be used to perform soil collection and analysis without human intervention. Unique among soil sampling robots, we use a general-purpose robotic arm rather than a soil core sampler.

Recent evidence supporting reionisation ending at redshift z<6 includes the rapid redshift evolution of the mean free path for Lyman-limit photons through the intergalactic medium at 510 discovered with JWST., Comment: 17 pages, 14 figures, accepted to MNRAS

Donor-Acceptor (D-A) type thermally activated delayed fluorescence (TADF) emitters which constitute the key functional units in proposed Gen3 Organic Light Emitting Diodes (OLEDs), are sensitive to the rigidity and polarity of their local environment. In particular, the torsional freedom of the D-A dihedral angle and the excited state dipole moments of the occurring charge transfer states, condition a distribution of TADF dynamics over the emitter ensemble, concealed in standard optical spectroscopy. Here we apply spectroscopy on the single molecule level to directly access individual emitter properties, and, thus, bypass the downside of ensemble averaging. By photon correlation data and locally resolved spectral information on single D-A type TADF molecules embedded in technological relevant host materials of different polarity and rigidity, we derive host-dependent characteristics and distributions in the TADF dynamics. Those can be related to the conformational freedom and the dielectric environment imposed by the specific local host rigidity and polarity, thereby pointing out new selection criteria for host-emitter combinations in OLEDs.

Recent quasar absorption line observations suggest that reionization may end as late as $z \approx 5.3$. As a means to search for large neutral hydrogen islands at $z<6$, we revisit long dark gaps in the Ly$\beta$ forest in VLT/X-Shooter and Keck/ESI quasar spectra. We stack the Ly$\alpha$ forest corresponding to both edges of these Ly$\beta$ dark gaps and identify a damping wing-like extended absorption profile. The average redshift of the stacked forest is $z=5.8$. By comparing these observations with reionization simulations, we infer that such a damping wing-like feature can be naturally explained if these gaps are at least partially created by neutral islands. Conversely, simulated dark gaps lacking neutral hydrogen struggle to replicate the observed damping wing features. Furthermore, this damping wing-like profile implies that the volume-averaged neutral hydrogen fraction must be $\langle x_{\rm HI} \rangle \geq 6.1 \pm 3.9\%$ at $z = 5.8$. Our results offer robust evidence that reionization extends below $z=6$., Comment: 8 pages, 5 figures, 1 table; accepted for publication in MNRAS Letters

We explicitly construct zero loss neural network classifiers. We write the weight matrices and bias vectors in terms of cumulative parameters, which determine truncation maps acting recursively on input space. The configurations for the training data considered are (i) sufficiently small, well separated clusters corresponding to each class, and (ii) equivalence classes which are sequentially linearly separable. In the best case, for $Q$ classes of data in $\mathbb{R}^M$, global minimizers can be described with $Q(M+2)$ parameters., Comment: AMS Latex, 22 pages, 3 figures

Simulations and observations find long tails in jellyfish galaxies, which are commonly thought to originate from ram-pressure stripped gas of the interstellar medium (ISM) in the immediate galactic wake. While at larger distances from the galaxy, they have been claimed to form in situ owing to thermal instability and fast radiative cooling of mixed ISM and intracluster medium (ICM). In this paper, we use magneto-hydrodynamical windtunnel simulations of a galaxy with the arepo code to study the origin of gas in the tails of jellyfish galaxies. To this end, we model the galaxy orbit in a cluster by accounting for a time-varying galaxy velocity, ICM density and turbulent magnetic field. Tracking gas flows between the ISM, the circumgalactic medium (CGM) and the ICM, we find contrary to popular opinion that the majority of gas in the tail originated in the CGM. Prior to the central passage of the jellyfish galaxy in the cluster, the CGM is directly transported to the clumpy jellyfish tail that has been shattered into small cloudlets. After the central cluster passage, gas in the tail originates both from the initial ISM and the CGM, but that from the latter was accreted to the galactic ISM before being ram-pressure stripped to form filamentary tentacles in the tail. Our simulation shows a declining gas metallicity in the tail as a function of downstream distance from the galaxy. We conclude that the CGM plays an important role in shaping the tails of jellyfish galaxies., Comment: 9 pages, 5 figures, accepted for publication in A&A (Sect. 4. Extragalactic astronomy)

We present Paicos, a new object-oriented Python package for analyzing simulations performed with Arepo. Paicos strives to reduce the learning curve for students and researchers getting started with Arepo simulations. As such, Paicos includes many examples in the form of Python scripts and Jupyter notebooks as well as an online documentation describing the installation procedure and recommended first steps. Paicos' main features are automatic handling of cosmological and physical units, computation of derived variables, 2D visualization (slices and projections), 1D and 2D histograms, and easy saving and loading of derived data including units and all the relevant metadata., Comment: Published in JOSS, code at https://github.com/tberlok/paicos

While machine learning (ML) models are increasingly used due to their high predictive power, their use in understanding the data-generating process (DGP) is limited. Understanding the DGP requires insights into feature-target associations, which many ML models cannot directly provide due to their opaque internal mechanisms. Feature importance (FI) methods provide useful insights into the DGP under certain conditions. Since the results of different FI methods have different interpretations, selecting the correct FI method for a concrete use case is crucial and still requires expert knowledge. This paper serves as a comprehensive guide to help understand the different interpretations of global FI methods. Through an extensive review of FI methods and providing new proofs regarding their interpretation, we facilitate a thorough understanding of these methods and formulate concrete recommendations for scientific inference. We conclude by discussing options for FI uncertainty estimation and point to directions for future research aiming at full statistical inference from black-box ML models.

Unveiling the thermal history of the intergalactic medium (IGM) at $4 \leq z \leq 5$ holds the potential to reveal early onset HeII reionization or lingering thermal fluctuations from HI reionization. We set out to reconstruct the IGM gas properties along simulated Lyman-alpha forest data on pixel-by-pixel basis, employing deep Bayesian neural networks. Our approach leverages the Sherwood-Relics simulation suite, consisting of diverse thermal histories, to generate mock spectra. Our convolutional and residual networks with likelihood metric predicts the Ly$\alpha$ optical depth-weighted density or temperature for each pixel in the Ly$\alpha$ forest skewer. We find that our network can successfully reproduce IGM conditions with high fidelity across range of instrumental signal-to-noise. These predictions are subsequently translated into the temperature-density plane, facilitating the derivation of reliable constraints on thermal parameters. This allows us to estimate temperature at mean cosmic density, $T_{\rm 0}$ with one sigma confidence $\delta T_{\rm 0} \sim 1000{\rm K}$ using only one $20$Mpc/h sightline ($\Delta z\simeq 0.04$) with a typical reionization history. Existing studies utilize redshift pathlength comparable to $\Delta z\simeq 4$ for similar constraints. We can also provide more stringent constraints on the slope ($1\sigma$ confidence interval $\delta {\rm \gamma} \lesssim 0.1$) of the IGM temperature-density relation as compared to other traditional approaches. We test the reconstruction on a single high signal-to-noise observed spectrum ($20$ Mpc/h segment), and recover thermal parameters consistent with current measurements. This machine learning approach has the potential to provide accurate yet robust measurements of IGM thermal history at the redshifts in question., Comment: 17 pages, Submitted to MNRAS

On-demand storage and retrieval of quantum information in coherent light-matter interfaces is a key requirement for future quantum networking and quantum communication applications. Alkali vapor memories offer scalable and robust high-bandwidth storage at high repetition rates which makes them a natural fit to interface with solid-state single-photon sources. Here, we experimentally realize a room-temperature ladder-type atomic vapor memory that operates on the Cs D1 line. We provide a detailed experimental characterization and demonstration of on-demand storage and retrieval of weak coherent laser pulses with 0.06 photons per pulse at a high signal-to-noise ratio of SNR$=830(80)$. The memory achieves a maximum internal storage efficiency of $\eta_{\text{int}}=15(1)\%$ and an estimated $1/e$-storage time of $\tau_{\mathrm{s}}\approx32\,$ns. Benchmark properties for the storage of single photons from inhomogeneously broadened state-of-the-art solid-state emitters are estimated from the performance of the memory. Together with the immediate availability of high-quality InGaAs quantum dots emitting at 895\,nm, these results provide clear prospects for the development of a heterogeneous on-demand quantum light interface.

Machine learning on large-scale genomic or transcriptomic data is important for many novel health applications. For example, precision medicine tailors medical treatments to patients on the basis of individual biomarkers, cellular and molecular states, etc. However, the data required is sensitive, voluminous, heterogeneous, and typically distributed across locations where dedicated machine learning hardware is not available. Due to privacy and regulatory reasons, it is also problematic to aggregate all data at a trusted third party.Federated learning is a promising solution to this dilemma, because it enables decentralized, collaborative machine learning without exchanging raw data. In this paper, we perform comparative experiments with the federated learning frameworks TensorFlow Federated and Flower. Our test case is the training of disease prognosis and cell type classification models. We train the models with distributed transcriptomic data, considering both data heterogeneity and architectural heterogeneity. We measure model quality, robustness against privacy-enhancing noise, computational performance and resource overhead. Each of the federated learning frameworks has different strengths. However, our experiments confirm that both frameworks can readily build models on transcriptomic data, without transferring personal raw data to a third party with abundant computational resources.

This paper explores the impact of stochastic mortality and disease on animal-based commodities, with a specific emphasis on aquaculture, particularly in the context of salmon farming. The investigation delves into the stochastic nature of mortality and treatment plans based on historical data related to salmon lice. Given that salmon lice pose a significant challenge in salmon farming, with associated treatment costs estimated to be comparable to feeding expenses, their removal is imperative to ensure the survival of the salmon and comply with the Norwegian government's stipulation of maintaining 0.5 lice per fish. We propose a new model that considers the relationship between hosts and parasites to estimate the number of treatments required and their overall cost. An important aspect of this model is its incorporation of stochastic effectiveness for each removal. After calibrating the model to the available data, the study examines the stochastic behavior's impact on the optimal harvesting decision in comparison to deterministic mortality models. The results indicate an approximate $1.5\,\%$ increase in the value of the salmon farm when employing the harvesting rule based on the stochastic host-parasite model as opposed to a deterministic model.

We analyze geometric aspects of the gradient descent algorithm in Deep Learning (DL), and give a detailed discussion of the circumstance that in underparametrized DL networks, zero loss minimization can generically not be attained. As a consequence, we conclude that the distribution of training inputs must necessarily be non-generic in order to produce zero loss minimizers, both for the method constructed in [Chen-Munoz Ewald 2023, 2024], or for gradient descent [Chen 2025] (which assume clustering of training data)., Comment: AMS Latex, 7 pages. Title changed, statement of Corollary 1.6 corrected

Jellyfish galaxies are promising laboratories for studying radiative cooling and magnetic fields in multiphase gas flows. Their long, dense tails are observed to be magnetised, and they extend up to 100 kpc into the intracluster medium (ICM), suggesting that their gas is thermally unstable so that the cold gas mass grows with time rather than being fully dissolved in the hot wind as a result of hydrodynamical interface instabilities. In this paper we use the AREPO code to perform magnetohydrodynamical windtunnel simulations of a jellyfish galaxy experiencing ram-pressure stripping by interacting with an ICM wind. The ICM density, temperature and velocity that the galaxy encounters are time-dependent and comparable to what a real jellyfish galaxy experiences while orbiting the ICM. In simulations with a turbulent magnetised wind we reproduce observations, which show that the magnetic field is aligned with the jellyfish tails. During the galaxy infall into the cluster with a near edge-on geometry, the gas flow in the tail is fountain-like, implying preferential stripping of gas where the rotational velocity vectors add up with the ram pressure while fall-back occurs in the opposite case. Hence, the tail velocity shows a memory of the rotation pattern of the disc. At the time of the nearest cluster passage, ram-pressure stripping is so strong that the fountain flow is destroyed and instead the tail is dominated by removal of gas. We show that gas in the tail is very fragmentative, which is a prediction of shattering due to radiative cooling., Comment: 14 pages, 7 figures, Accepted for publication in MNRAS

Confronting measurements of the Lyman-$\alpha$ forest with cosmological hydrodynamical simulations has produced stringent constraints on models of particle dark matter and the thermal and ionization state of the intergalactic medium. We investigate the robustness of such models of the Lyman-$\alpha$ forest, focussing on the effect of particle initial conditions on the Lyman-$\alpha$ forest statistics in cosmological SPH simulations. We study multiple particle initialization algorithms in simulations that are designed to be identical in other respects. In agreement with the literature, we find that the correct linear theory evolution is obtained when a glass-like configuration is used for initial unperturbed gas particle positions alongside a regular grid configuration for dark matter particles and the use of non-identical initial density perturbations for gas and dark matter. However, we report that this introduces a large scale-dependent distortion in the one-dimensional Lyman-$\alpha$ transmission power spectrum at small scales ($k > 0.05$ s/km). The effect is close to $50\%$ at $k\sim 0.1$ s/km, and persists at higher resolution. This can severely bias inferences in parameters such as the dark matter particle mass. By considering multiple initial conditions codes and their variations, we also study the impact of a variety of other assumptions and algorithmic choices, such as adaptive softening, background radiation density, particle staggering, and perturbation theory accuracy, on the matter power spectrum, the Lyman-$\alpha$ flux power spectrum, and the Lyman-$\alpha$ flux PDF. This work reveals possible pathways towards more accurate theoretical models of the Lyman-$\alpha$ forest to match the quality of upcoming measurements., Comment: 18 pages, 16 figures, accepted for publication in MNRAS, comments welcome