Search

We present full 3-28 micron JWST MIRI/MRS and NIRSpec/IFU spectra of the western nucleus of Arp 220, the nearest ultraluminous infrared galaxy. This nucleus has long been suggested to possibly host an embedded Compton-thick AGN. Millimeter observations of the dust continuum suggest the presence of a distinct 20 pc core with a dust temperature of $T_\mathrm{d} \gtrsim 500~\mathrm{K}$, in addition to a 100 pc circumnuclear starburst disk. However, unambiguously identifying the nature of this core is challenging, due to the immense obscuration, the nuclear starburst activity, and the nearby eastern nucleus. With the JWST integral field spectrographs, we can, for the first time, separate the two nuclei across this full wavelength range, revealing a wealth of molecular absorption features towards the western nucleus. We analyse the rovibrational bands detected at 4-22 micron, deriving column densities and rotational temperatures for 10 distinct species. Optically thick features of C$_2$H$_2$, HCN and HNC suggest that this molecular gas is hidden behind a curtain of cooler dust, and indicate that the column densities of C$_2$H$_2$ and HCN are an order of magnitude higher than previously derived from Spitzer observations. We identify a warm HCN component with rotational temperature $T_\mathrm{rot} = 330~\mathrm{K}$, which we associate with radiative excitation by the hot inner nucleus. We propose a geometry where the detected molecular gas is located in the inner regions of the starburst disk, directly surrounding the hot 20 parsec core. The chemical footprint of the western nucleus is reminiscent of that of hot cores, with additional evidence for shocks. No evidence for the presence of an AGN in the form of X-ray-driven chemistry or extreme excitation is found., Comment: Submitted to A&A

We developed, characterised, and verified an alignment method for the DESHIMA 2.0 instrument, an ultra wide-band spectrometer operating between 200-400 GHz, at the ASTE telescope. Due to the ultra-wide bandwidth of DESHIMA 2.0, we require the alignment to work across the entire spectral range. Moreover, alignment should be possible remotely, as on-site access is impractical. To fulfill these requirements we mounted the warm optics, consisting of a modified Dragonian dual reflector system, on a motor controlled hexapod. Crucial in the alignment procedure is our sky chopper, which allows fast beam switching. Importantly, it has a small entrance and exit aperture coupling to (cold) sky, which creates a measurable signal with respect to the warm cabin environment. By scanning the instrument beam across the entrance aperture of the sky chopper using the hexapod, we found the hexapod configuration that produced the lowest signal on our detectors, implying the beam is coupled fully to cold sky and not the warm cabin. We first characterised the alignment procedure in the laboratory, where we used a vat containing liquid nitrogen as the cold source behind the sky chopper. Then, we applied the alignment procedure to DESHIMA 2.0 at ASTE, using the sky as the cold source. We used observations of Mars to calculate the aperture efficiency. We did not recover the aperture efficiencies up to design values, which we suspect to be due to a raised sidelobe level or error beam pattern. However, we did find that the alignment procedure significantly improved the aperture efficiency compared to previously reported values of the aperture efficiency of DESHIMA at ASTE, which indicates the veracity of the alignment procedure., Comment: 10 pages, 9 figures

As part of the guaranteed time observations program Mid-Infrared Characterization Of Nearby Iconic galaxy Centers (MICONIC), we used the medium-resolution spectrometer (MRS) of the Mid-Infrared Instrument (MIRI) on board of the JWST to study the nearby merger NGC6240. We aim to characterise the dual active galactic nuclei (AGN), the ionised gas outflows and the main properties of the interstellar medium over a mapped area of 6.6"x7.7". We obtained integral field spectroscopic mid-infrared data of NGC6240, resolving both nuclei for the first time in the full 5-28{\mu}m spectral range. We modelled the emission lines through a kinematic decomposition, finding that the fine-structure lines in the southern (S) nucleus are broader than for the northern (N) nucleus (full width at half maximum of $\geq$1500 vs ~700 km s^{-1} on average). High excitation lines, such as [NeV], [NeVI], and [MgV], are clearly detected in the N nucleus. In the S nucleus, the same lines can be detected but only after a decomposition of the PAH features in the integrated spectrum, due to a combination of a strong mid-IR continuum, the broad emission lines, and the intense star formation (SF). The SF is distributed all over the mapped FoV of 3.5kpc x 4.1kpc (projected), with the maximum located around the S nucleus. Both nuclear regions appear to be connected by a bridge region detected with all the emission lines. Based on the observed MRS line ratios and the high velocity dispersion ({\sigma}~600 km s^{-1}), shocks are also dominating the emission in this system. We detected the presence of outflows as a bubble north-west from the N nucleus and at the S nucleus. We estimated a ionised mass outflow rate of 1.4$\pm$0.3 M yr^{-1} and 1.8$\pm$0.2 M yr^{-1}, respectively. Given the derived kinetic power of these outflows, both the AGN and the starburst could have triggered them. [Abridged], Comment: 23 pages, 10 figures. Accepted for publication in A&A on 26/11/2024

Previous Hubble Space Telescope (HST) observations of the star-forming cluster NGC 346 in the Small Magellanic Cloud (SMC) had revealed a large population of pre-main sequence (PMS) candidates, characterised by Halpha excess emission in their photometry. However, without access to spectroscopy, the nature of these objects remained unclear. Using the NIRSpec instrument on board JWST, we studied a sample of these stars, with masses in the range ~0.9-1.8 Msun, effective temperatures in the range 4,500-8,000 K, and PMS ages between ~0.1 and 30 Myr. Here we present the first spectra of solar-mass PMS stars in the metal-poor SMC (Z=1/8 Zsun) and discuss the physical properties of ten representative sources with good signal-to-noise ratio. The observations indicate that even the oldest of these PMS candidates are still accreting gas with typical rates of ~10^{-8} Msun/yr for stars older than ~10 Myr, confirming their PMS nature. The spectra also reveal near-infrared excess and molecular hydrogen excitation lines consistent with the presence of discs around these stars. These findings suggest that in a low-metallicity environment circumstellar discs can live longer than previously thought., Comment: 17 pages, 7 figures, to be published in The Astrophysical Journal, Vol 977, Article 214

Hydrogen emission lines are routinely used to estimate the mass accretion rate of pre-main-sequence stars. Despite the clear correlation between the accretion luminosity of a star and hydrogen line luminosities, the physical origin of these lines is still unclear, with magnetospheric accretion and magneto-centrifugal winds as the two most often invoked mechanisms. Using a combination of HST photometry and new JWST NIRSpec spectra, we have analysed the SED and emission line spectra of five sources in order to determine their underlying photospheric properties, and to attempt to reveal the physical origin of their hydrogen emission lines. These sources reside in NGC 3603, a Galactic massive star forming region. We have fitted the SED of the five sources employing a Markov Chain Monte Carlo exploration to estimate $T_{eff}$, $R_{*}$, $M_{*}$ and $A(V)$ for each source. We have performed a kinematic analysis across three spectral series of hydrogen lines, Paschen, Brackett, and Pfund, totalling $\ge 15$ lines. The FWHM and optical depth of the spectrally resolved lines have been studied in order to constrain the emission origin. The five sources all have SEDs consistent with young intermediate-mass stars. We have classified three of these sources as Herbig Ae type stars based on their effective temperature. Their hydrogen lines show broad profiles with FWHMs $\ge 200$ km s$^{-1}$. Hydrogen lines with high upper energy levels $n_{up}$ tend to be significantly broader than lines with lower $n_{up}$. The optical depth of the emission lines is also highest for the high velocity component of each line, becoming optically thin in the low velocity component. This is consistent with emission from a magnetospheric accretion flow, and cannot be explained as originating in a magneto-centrifugal wind, or other line emission mechanisms thought to be present in protoplanetary disks.

NGC 3603 is the optically brightest massive star forming region (SFR) in the Milky Way, representing a small scale starburst region. Studying young stars in regions like this allows us to assess how star and planet formation proceeds in a dense clustered environment with high levels of UV radiation. JWST provides the sensitivity, unbroken wavelength coverage, and spatial resolution required to study individual pre-main-sequence (PMS) stars in distant massive SFRs in detail for the first time. Using the Micro-Shutter Assembly (MSA) onboard the Near InfraRed Spectrograph (NIRSpec), multi-object spectroscopy was performed, yielding 100 stellar spectra. We fit the PMS spectra to derive their photospheric properties, extinction, and NIR veiling. From this, we determined the masses and ages of our sources by placing them on the Hertzsprung-Russel diagram (HRD). Their accretion rates were determined by converting the luminosity of hydrogen emission lines to an accretion luminosity. We have classified 42 as actively accreting. Our sources span a range of masses from 0.5 to 7 $M_{\odot}$. Twelve of these accreting sources have ages consistent with $\ge$ 10 Myrs, with four having ages of $\ge$ 15 Myrs. Their mass accretion rates span 5 orders of magnitude and are systematically higher for a given stellar mass than for a comparative sample taken from low-mass SFRs. We report an environmental relationship between $\dot{M}_{acc}$ and the density of ambient molecular gas as traced by nebular $H_2$ emission.

In widely used neural network-based collaborative filtering models, users' history logs are encoded into latent embeddings that represent the users' preferences. In this setting, the models are capable of mapping users' protected attributes (e.g., gender or ethnicity) from these user embeddings even without explicit access to them, resulting in models that may treat specific demographic user groups unfairly and raise privacy issues. While prior work has approached the removal of a single protected attribute of a user at a time, multiple attributes might come into play in real-world scenarios. In the work at hand, we present AdvXMultVAE which aims to unlearn multiple protected attributes (exemplified by gender and age) simultaneously to improve fairness across demographic user groups. For this purpose, we couple a variational autoencoder (VAE) architecture with adversarial training (AdvMultVAE) to support simultaneous removal of the users' protected attributes with continuous and/or categorical values. Our experiments on two datasets, LFM-2b-100k and Ml-1m, from the music and movie domains, respectively, show that our approach can yield better results than its singular removal counterparts (based on AdvMultVAE) in effectively mitigating demographic biases whilst improving the anonymity of latent embeddings.

Instruction-tuned LLMs are able to provide an explanation about their output to users by generating self-explanations. These do not require gradient computations or the application of possibly complex XAI methods. In this paper, we analyse whether this ability results in a good explanation. We evaluate self-explanations in the form of input rationales with respect to their plausibility to humans as well as their faithfulness to models. We study two text classification tasks: sentiment classification and forced labour detection, i.e., identifying pre-defined risk indicators of forced labour. In addition to English, we include Danish and Italian translations of the sentiment classification task and compare self-explanations to human annotations for all samples. To allow for direct comparisons, we also compute post-hoc feature attribution, i.e., layer-wise relevance propagation (LRP) and analyse 4 LLMs. We show that self-explanations align more closely with human annotations compared to LRP, while maintaining a comparable level of faithfulness. This finding suggests that self-explanations indeed provide good explanations for text classification., Comment: preprint

Cognitive biases have been studied in psychology, sociology, and behavioral economics for decades. Traditionally, they have been considered a negative human trait that leads to inferior decision-making, reinforcement of stereotypes, or can be exploited to manipulate consumers, respectively. We argue that cognitive biases also manifest in different parts of the recommendation ecosystem and at different stages of the recommendation process. More importantly, we contest this traditional detrimental perspective on cognitive biases and claim that certain cognitive biases can be beneficial when accounted for by recommender systems. Concretely, we provide empirical evidence that biases such as feature-positive effect, Ikea effect, and cultural homophily can be observed in various components of the recommendation pipeline, including input data (such as ratings or side information), recommendation algorithm or model (and consequently recommended items), and user interactions with the system. In three small experiments covering recruitment and entertainment domains, we study the pervasiveness of the aforementioned biases. We ultimately advocate for a prejudice-free consideration of cognitive biases to improve user and item models as well as recommendation algorithms.

The Mid-infrared ELT Imager and Spectrograph (METIS) is one of the first-generation scientific instruments for the ELT, built under the supervision of ESO by a consortium of research institutes across and beyond Europe. Designed to cover the 3 to 13 $\mu$m wavelength range, METIS had its final design reviewed in Fall 2022, and has then entered in earnest its manufacture, assembly, integration, and test (MAIT) phase. Here, we present the final design of the METIS high-contrast imaging (HCI) modes. We detail the implementation of the two main coronagraphic solutions selected for METIS, namely the vortex coronagraph and the apodizing phase plate, including their combination with the high-resolution integral field spectrograph of METIS, and briefly describe their respective backup plans (Lyot coronagraph and shaped pupil plate). We then describe the status of the MAIT phase for HCI modes, including a review of the final design of individual components such as the vortex phase masks, the grayscale ring apodizer, and the apodizing phase plates, as well as a description of their on-going performance tests and of our plans for system-level integration and tests. Using end-to-end simulations, we predict the performance that will be reached on sky by the METIS HCI modes in presence of various environmental and instrumental disturbances, including non-common path aberrations and water vapor seeing, and discuss our strategy to mitigate these various effects. We finally illustrate with mock observations and data processing that METIS should be capable of directly imaging temperate rocky planets around the nearest stars., Comment: 16 pages, paper presented at SPIE Astronomical Telescopes + Instrumentation 2024

Context. A necessary ingredient in understanding the star formation history of a young cluster is knowledge of the extinction towards the region. This has typically been done by making use of the colour-difference method with photometry, or similar methods utilising the colour-colour diagram. Aims. The colour-excess can be independently determined by studying the decrements of the recombination lines produced by the nebular gas. Having access to many recombination lines from the same spectral series removes the need of adopting an extinction curve. Methods. Using the Micro-Shutter Assembly (MSA) on board the Near InfraRed Spectrograph (NIRSpec), multi-object spectroscopy was performed, yielding 600 nebular spectra from the Galactic massive star formation region NGC 3603. The recombination line intensity ratios were used to determine independent values of colour-excess. Results. The extinction characteristics of NGC 3603 are similar to other Galactic HII regions like Orion, as well as starburst regions such 30 Doradus in the Large Magellanic Cloud, in that we find a relatively large value of total-to-selective extinction of 4.8 $\pm$ 1.06, larger than the Galactic average of 3.1. We find a typical value for the colour-excess of 0.64 $\pm$ 0.27, significantly lower than values determined in previous studies. We also present a stacked nebular spectrum with a typical continuum S/N = 70. This spectrum highlights the recombination lines of the HII region, several s-process elements such as Kr III and Se IV, and molecular H2 emission lines. Conclusions. Using ratios of hydrogen recombination lines, we calculated the total-to-selective extinction, colour-excess and visual extinction for > 200 lines of sight across NGC 3603. An extinction curve with a total-to-selective extinction of 4.8 $\pm$ 1.06 was found, corresponding to a colour-excess of 0.64 $\pm$ 0.27.

We present a JWST imaging survey with NIRCam and MIRI of NGC 346, the brightest star-forming region in the Small Magellanic Cloud (SMC). By combining aperture and point spread function (PSF) photometry of eleven wavelength bands across these two instruments, we have detected more than 200,000 unique sources. Using near-infrared (IR) color analysis, we observe various evolved and young populations, including 196 young stellar objects (YSOs) and pre-main sequence stars suitable for forthcoming spectroscopic studies. We expand upon this work, creating mid-IR color-magnitude diagrams and determining color cuts to identify 833 reddened sources which are YSO candidates. We observe that these candidate sources are spatially associated with regions of dusty, filamentary nebulosity. Furthermore, we fit model YSO spectral energy distributions (SEDs) to a selection of sources with detections across all of our MIRI bands. We classify with a high degree of confidence 23 YSOs in this sample and estimate their radii, bolometric temperatures, luminosities, and masses. We detect YSOs approaching 1 solar mass, the lowest-mass extragalactic YSOs confirmed to date.

The JWST has captured the sharpest IR images ever taken of the Horsehead nebula, a prototypical moderately irradiated PDR that is fully representative of most of the UV-illuminated molecular gas in the Milky Way and star-forming galaxies. We investigate the impact of FUV radiation of a molecular cloud and constrain the structure of the edge of the PDR and its illumination conditions. We used NIRCam and MIRI to obtain 17 broadband and 6 narrowband maps from 0.7 to 28 $\mu$m. We mapped the dust emission, scattered light, and several gas phase lines. We also used HST-WFC3 maps at 1.1 and 1. 6 $\mu$m, along with HST-STIS spectroscopic observations of the H$\alpha$ line. We probed the structure of the edge of the Horsehead and resolved its spatial complexity. We detected a network of faint striated features extending perpendicularly to the PDR front into the H\,II region in filters sensitive to nano-grain emission and light scattered by larger grains. This may indeed figure as the first detection of the entrainment of dust particles in the evaporative flow. The map of the 1-0 S(1) line of H$_2$ presents sharp sub-structures on scales as small as 1.5 arcsec. The ionization and dissociation fronts appear at distances 1-2 arcsec behind the edge of the PDR and seem to spatially coincide, indicating a thickness of the neutral atomic layer below 100 au. All broadband maps present strong color variations which can be explained by dust attenuation. Deviations of the emissions in the H$\alpha$, Pa$\alpha,$ and Br$\alpha$ lines also indicate dust attenuation. With a very simple model, we derive the main features of the extinction curve. A small excess of extinction at 3 $\mu$m may be attributed to icy H$_2$O mantles onto grains. In all lines of sight crossing the inner regions of the Horsehead, it appears that dust attenuation is non-negligible over the entire spectral range., Comment: 25 pages, 17 figures

Power system operation is of vital importance and must be developed far beyond today's practice to meet future needs. Almost all European countries are facing an abrupt and very important increase of renewables with intrinsically varying yields which are difficult to predict. In addition, an increase of new types of electric loads and a reduction of traditional production from bulk generation can be observed as well. Hence, the level of complexity of system operation steadily increases. Because of these developments, the traditional power system is being transformed into a smart grid. Previous and ongoing research has tended to focus on how specific aspects of smart grids can be developed and validated, but until now there exists no integrated approach for analysing and evaluating complex smart grid configurations. To tackle these research and development needs, a pan-European research infrastructure is realized in the ERIGrid project that supports the technology development as well as the roll out of smart grid technologies and solutions. This paper provides an overview of the main results of ERIGrid which have been achieved during the last four years. Also, experiences and lessons learned are discussed and an outlook to future research needs is provided.

Mid-infrared emission features probe the properties of ionized gas, and hot or warm molecular gas. The Orion Bar is a frequently studied photodissociation region (PDR) containing large amounts of gas under these conditions, and was observed with the MIRI IFU aboard JWST as part of the "PDRs4All" program. The resulting IR spectroscopic images of high angular resolution (0.2") reveal a rich observational inventory of mid-IR emission lines, and spatially resolve the substructure of the PDR, with a mosaic cutting perpendicularly across the ionization front and three dissociation fronts. We extracted five spectra that represent the ionized, atomic, and molecular gas layers, and measured the most prominent gas emission lines. An initial analysis summarizes the physical conditions of the gas and the potential of these data. We identified around 100 lines, report an additional 18 lines that remain unidentified, and measured the line intensities and central wavelengths. The H I recombination lines originating from the ionized gas layer bordering the PDR, have intensity ratios that are well matched by emissivity coefficients from H recombination theory, but deviate up to 10% due contamination by He I lines. We report the observed emission lines of various ionization stages of Ne, P, S, Cl, Ar, Fe, and Ni, and show how certain line ratios vary between the five regions. We observe the pure-rotational H$_2$ lines in the vibrational ground state from 0-0 S(1) to 0-0 S(8), and in the first vibrationally excited state from 1-1 S(5) to 1-1 S(9). We derive H$_2$ excitation diagrams, and approximate the excitation with one thermal (~700 K) component representative of an average gas temperature, and one non-thermal component (~2700 K) probing the effect of UV pumping. We compare these results to an existing model for the Orion Bar PDR and highlight the differences with the observations., Comment: 26 pages, 12 figures, 3 tables. Submitted to A&A, under review (1st revision)

Instruction-finetuned Large Language Models inherit clear political leanings that have been shown to influence downstream task performance. We expand this line of research beyond the two-party system in the US and audit Llama Chat in the context of EU politics in various settings to analyze the model's political knowledge and its ability to reason in context. We adapt, i.e., further fine-tune, Llama Chat on speeches of individual euro-parties from debates in the European Parliament to reevaluate its political leaning based on the EUandI questionnaire. Llama Chat shows considerable knowledge of national parties' positions and is capable of reasoning in context. The adapted, party-specific, models are substantially re-aligned towards respective positions which we see as a starting point for using chat-based LLMs as data-driven conversational engines to assist research in political science., Comment: accepted to NAACL 2024 as a short paper

In this letter we present the first systematic spectroscopic measurements of the Near-InfraRed (NIR) hydrogen recombination lines Paschen alpha ($Pa_{\alpha}$ $\lambda = 1.875 \mu m$) and Brackett beta ($Br_{\beta}$ $\lambda = 2.626 \mu m$), produced by pre-main-sequence (PMS) stars. These stars, consisting of T Tauri and Herbig AeBe stars are located in the massive Galactic star-formation region NGC 3603. The measurements were obtained from JWST NIRSpec, using multi-object spectroscopy (MOS) mode. Utilising existing empirical relations between $L_{acc}$ and $L_{Br_{\gamma}}$, we have used our new measurements to formulate, for the first time, an empirical relationship between the accretion luminosity $L_{acc}$ of the stars, and the line luminosities $L_{line}$ of both $Pa_{\alpha}$ and $Br_{\beta}$. These relationships are: $\log_{10}( \frac{L_{acc}}{L_{\odot}}) = 1.42 (\pm 0.18) \times \log_{10}( \frac{L_{Pa_{\alpha}}}{L_{\odot}}) + 3.33 (\pm 0.42)$ and $\log_{10}( \frac{L_{acc}}{L_{\odot}}) = 1.47 (\pm 0.18) \times \log_{10}( \frac{L_{Br_{\beta}}}{L_{\odot}}) + 4.60 (\pm 0.57)$. These new relationships are key for roughly estimating the accretion rates for large samples of PMS stars with JWST.

We present a JWST imaging survey of I Zw 18, the archetypal extremely metal-poor, star-forming (SF), blue compact dwarf galaxy. With an oxygen abundance of only $\sim$3% $Z_{\odot}$, it is among the lowest-metallicity systems known in the local Universe, and is, therefore, an excellent accessible analog for the galactic building blocks which existed at early epochs of ionization and star formation. These JWST data provide a comprehensive infrared (IR) view of I Zw 18 with eight filters utilizing both Near Infrared Camera (F115W, F200W, F356W, and F444W) and Mid-Infrared Instrument (F770W, F1000W, F1500W, and F1800W) photometry, which we have used to identify key stellar populations that are bright in the near- and mid-IR. These data allow for a better understanding of the origins of dust and dust-production mechanisms in metal-poor environments by characterizing the population of massive, evolved stars in the red supergiant (RSG) and asymptotic giant branch (AGB) phases. In addition, it enables the identification of the brightest dust-enshrouded young stellar objects (YSOs), which provide insight into the formation of massive stars at extremely low metallicities typical of the very early Universe. This paper provides an overview of the observational strategy and data processing, and presents first science results, including identifications of dusty AGB, RSG, and bright YSO candidates. These first results assess the scientific quality of JWST data and provide a guide for obtaining and interpreting future observations of the dusty and evolved stars inhabiting compact dwarf SF galaxies in the local Universe., Comment: Accepted for publication in The Astronomical Journal; 25 pages, 14 figures, 4 tables

Most low-mass stars form in stellar clusters that also contain massive stars, which are sources of far-ultraviolet (FUV) radiation. Theoretical models predict that this FUV radiation produces photo-dissociation regions (PDRs) on the surfaces of protoplanetary disks around low-mass stars, impacting planet formation within the disks. We report JWST and Atacama Large Millimetere Array observations of a FUV-irradiated protoplanetary disk in the Orion Nebula. Emission lines are detected from the PDR; modelling their kinematics and excitation allows us to constrain the physical conditions within the gas. We quantify the mass-loss rate induced by the FUV irradiation, finding it is sufficient to remove gas from the disk in less than a million years. This is rapid enough to affect giant planet formation in the disk.

Rationales in the form of manually annotated input spans usually serve as ground truth when evaluating explainability methods in NLP. They are, however, time-consuming and often biased by the annotation process. In this paper, we debate whether human gaze, in the form of webcam-based eye-tracking recordings, poses a valid alternative when evaluating importance scores. We evaluate the additional information provided by gaze data, such as total reading times, gaze entropy, and decoding accuracy with respect to human rationale annotations. We compare WebQAmGaze, a multilingual dataset for information-seeking QA, with attention and explainability-based importance scores for 4 different multilingual Transformer-based language models (mBERT, distil-mBERT, XLMR, and XLMR-L) and 3 languages (English, Spanish, and German). Our pipeline can easily be applied to other tasks and languages. Our findings suggest that gaze data offers valuable linguistic insights that could be leveraged to infer task difficulty and further show a comparable ranking of explainability methods to that of human rationales., Comment: Accepted to LREC-COLING 2024

Integrated superconducting spectrometers (ISSs) for wideband sub-mm astronomy utilise quasi-optical systems for coupling radiation from the telescope to the instrument. Misalignment in these systems is detrimental to the system performance. The common method of using an optical laser to align the quasi-optical components requires accurate alignment of the laser to the sub-mm beam coming from the instrument, which is not always guaranteed to a sufficient accuracy. We develop an alignment strategy for wideband ISSs directly utilising the sub-mm beam of the wideband ISS. The strategy should be applicable in both telescope and laboratory environments. Moreover, the strategy should deliver similar quality of the alignment across the spectral range of the wideband ISS. We measure misalignment in a quasi-optical system operating at sub-mm wavelengths using a novel phase and amplitude measurement scheme, capable of simultaneously measuring the complex beam patterns of a direct-detecting ISS across a harmonic range of frequencies. The direct detection nature of the MKID detectors in our device-under-test, DESHIMA 2.0, necessitates the use of this measurement scheme. Using geometrical optics, the measured misalignment, a mechanical hexapod, and an optimisation algorithm, we follow a numerical approach to optimise the positioning of corrective optics with respect to a given cost function. Laboratory measurements of the complex beam patterns are taken across a harmonic range between 205 and 391 GHz and simulated through a model of the ASTE telescope in order to assess the performance of the optimisation at the ASTE telescope. Laboratory measurements show that the optimised optical setup corrects for tilts and offsets of the sub-mm beam. Moreover, we find that the simulated telescope aperture efficiency is increased across the frequency range of the ISS after the optimisation., Comment: 10 pages, 12 figures. Accepted for publication by A&A

Trapped-ion quantum computers exhibit promising potential to provide platforms for high-quality qubits and reliable quantum computation. The Quantum Charge Coupled Device (QCCD) architecture is a leading example that offers a modular solution to enable the realization of scalable quantum computers, paving the way for practical quantum algorithms with large qubit numbers. Within these devices, ions can be shuttled (moved) throughout the trap and through different dedicated zones, e.g., a memory zone for storage and a processing zone for the actual computation. However, due to decoherence of the ions' quantum states, the qubits lose their quantum information over time. Thus, the required time steps of shuttling operations should be minimized. In this work, we propose a heuristic approach to determining an efficient shuttling schedule, which orchestrates the movement operations within the device. Given a quantum algorithm and a device architecture, the proposed approach produces shuttling schedules with a close-to-minimal amount of time steps for small-size QCCD architectures. For large scale QCCD devices, empirical evaluations show promising results with respect to quality of the solution as well as performance. An implementation of the proposed approach is publicly available as part of the open-source Munich Quantum Toolkit (MQT) at https://github.com/cda-tum/mqt-ion-shuttler.

We revisit the problem of fairly allocating a sequence of time slots when agents may have different levels of patience (Mackenzie and Komornik 2023). For each number of agents, we provide a lower threshold and an upper threshold on the level of patience such that (i) if each agent is at least as patient as the lower threshold, then there is a proportional allocation, and (ii) if each agent is at least as patient as the upper threshold and moreover has weak preference for earlier time slots, then there is an envy-free allocation. In both cases, the proof is constructive.

We advance the theoretical study of $\{0, 1/2\}$-cuts for integer programming problems $\max\{c^T x \colon A x \leq b, x \text{ integer}\}$. Such cuts are Gomory-Chv\'atal cuts that only need multipliers of value $0$ or $1/2$ in their derivation. The intersection of all $\{0, 1/2\}$-cuts derived from $Ax \le b$ is denoted by $P_{1/2}$ and called the $\{0,1/2\}$-closure of $P = \{x : Ax \le b\}$. The primal separation problem for $\{0, 1/2\}$-cuts is: Given a vertex $\hat x$ of the integer hull of $P$ and some fractional point $x^* \in P$, does there exist a $\{0,1/2\}$-cut that is tight at $\hat x$ and violated by $x^*$? Primal separation is the key ingredient of primal cutting-plane approaches to integer programming. In general, primal separation for $\{0,1/2\}$-cuts is NP-hard. We present two cases for which primal separation is solvable in polynomial time. As an interesting side product, we obtain a(nother) simple proof that matching can be solved in polynomial time. Furthermore, since optimization over the Gomory-Chv\'atal closure is also NP-hard, there has been recent research on solving the optimization problem over the Gomory-Chv\'atal closure approximately. In a similar spirit, we show that the optimization problem over the $\{0,1/2\}$-closure can be solved in polynomial time up to a factor $(1 + \varepsilon)$, for any fixed $\varepsilon > 0$., Comment: 14 pages

Trapped ions are a promising technology for building scalable quantum computers. Not only can they provide a high qubit quality, but they also enable modular architectures, referred to as Quantum Charge Coupled Device (QCCD) architecture. Within these devices, ions can be shuttled (moved) throughout the trap and through different dedicated zones, e.g., a memory zone for storage and a processing zone for the actual computation. However, this movement incurs a cost in terms of required time steps, which increases the probability of decoherence, and, thus, should be minimized. In this paper, we propose a formalization of the possible movements in ion traps via Boolean satisfiability. This formalization allows for determining the minimal number of time steps needed for a given quantum algorithm and device architecture, hence reducing the decoherence probability. An empirical evaluation confirms that -- using the proposed approach -- minimal results (i.e., the lower bound) can be determined for the first time. An open-source implementation of the proposed approach is publicly available at https://github.com/cda-tum/mqt-ion-shuttler., Comment: 7 pages, 6 figures, 1 table. Accepted for Asia and South Pacific Design Automation Conference (ASP-DAC) 2024

Pretrained machine learning models are known to perpetuate and even amplify existing biases in data, which can result in unfair outcomes that ultimately impact user experience. Therefore, it is crucial to understand the mechanisms behind those prejudicial biases to ensure that model performance does not result in discriminatory behaviour toward certain groups or populations. In this work, we define gender bias as our case study. We quantify bias amplification in pretraining and after fine-tuning on three families of vision-and-language models. We investigate the connection, if any, between the two learning stages, and evaluate how bias amplification reflects on model performance. Overall, we find that bias amplification in pretraining and after fine-tuning are independent. We then examine the effect of continued pretraining on gender-neutral data, finding that this reduces group disparities, i.e., promotes fairness, on VQAv2 and retrieval tasks without significantly compromising task performance., Comment: To appear in EMNLP 2024

In order to build reliable and trustworthy NLP applications, models need to be both fair across different demographics and explainable. Usually these two objectives, fairness and explainability, are optimized and/or examined independently of each other. Instead, we argue that forthcoming, trustworthy NLP systems should consider both. In this work, we perform a first study to understand how they influence each other: do fair(er) models rely on more plausible rationales? and vice versa. To this end, we conduct experiments on two English multi-class text classification datasets, BIOS and ECtHR, that provide information on gender and nationality, respectively, as well as human-annotated rationales. We fine-tune pre-trained language models with several methods for (i) bias mitigation, which aims to improve fairness; (ii) rationale extraction, which aims to produce plausible explanations. We find that bias mitigation algorithms do not always lead to fairer models. Moreover, we discover that empirical fairness and explainability are orthogonal., Comment: 15 pages (incl Appendix), 4 figures, 8 tables

Contrastive explanations, where one decision is explained in contrast to another, are supposed to be closer to how humans explain a decision than non-contrastive explanations, where the decision is not necessarily referenced to an alternative. This claim has never been empirically validated. We analyze four English text-classification datasets (SST2, DynaSent, BIOS and DBpedia-Animals). We fine-tune and extract explanations from three different models (RoBERTa, GTP-2, and T5), each in three different sizes and apply three post-hoc explainability methods (LRP, GradientxInput, GradNorm). We furthermore collect and release human rationale annotations for a subset of 100 samples from the BIOS dataset for contrastive and non-contrastive settings. A cross-comparison between model-based rationales and human annotations, both in contrastive and non-contrastive settings, yields a high agreement between the two settings for models as well as for humans. Moreover, model-based explanations computed in both settings align equally well with human rationales. Thus, we empirically find that humans do not necessarily explain in a contrastive manner.9 pages, long paper at ACL 2022 proceedings., Comment: 9 pages, long paper at EMNLP 2023 proceedings

(Abridged) We investigate the impact of radiative feedback from massive stars on their natal cloud and focus on the transition from the HII region to the atomic PDR (crossing the ionisation front (IF)), and the subsequent transition to the molecular PDR (crossing the dissociation front (DF)). We use high-resolution near-IR integral field spectroscopic data from NIRSpec on JWST to observe the Orion Bar PDR as part of the PDRs4All JWST Early Release Science Program. The NIRSpec data reveal a forest of lines including, but not limited to, HeI, HI, and CI recombination lines, ionic lines, OI and NI fluorescence lines, Aromatic Infrared Bands (AIBs including aromatic CH, aliphatic CH, and their CD counterparts), CO2 ice, pure rotational and ro-vibrational lines from H2, and ro-vibrational lines HD, CO, and CH+, most of them detected for the first time towards a PDR. Their spatial distribution resolves the H and He ionisation structure in the Huygens region, gives insight into the geometry of the Bar, and confirms the large-scale stratification of PDRs. We observe numerous smaller scale structures whose typical size decreases with distance from Ori C and IR lines from CI, if solely arising from radiative recombination and cascade, reveal very high gas temperatures consistent with the hot irradiated surface of small-scale dense clumps deep inside the PDR. The H2 lines reveal multiple, prominent filaments which exhibit different characteristics. This leaves the impression of a "terraced" transition from the predominantly atomic surface region to the CO-rich molecular zone deeper in. This study showcases the discovery space created by JWST to further our understanding of the impact radiation from young stars has on their natal molecular cloud and proto-planetary disk, which touches on star- and planet formation as well as galaxy evolution., Comment: 52 pages, 30 figures, submitted to A&A

As the landscape of devices that interact with the electrical grid expands, also the complexity of the scenarios that arise from these interactions increases. Validation methods and tools are typically domain specific and are designed to approach mainly component level testing. For this kind of applications, software and hardware-in-the-loop based simulations as well as lab experiments are all tools that allow testing with different degrees of accuracy at various stages in the development life-cycle. However, things are vastly different when analysing the tools and the methodology available for performing system-level validation. Until now there are no available well-defined approaches for testing complex use cases involving components from different domains. Smart grid applications would typically include a relatively large number of physical devices, software components, as well as communication technology, all working hand in hand. This paper explores the possibilities that are opened in terms of testing by the integration of a real-time simulator into co-simulation environments. Three practical implementations of such systems together with performance metrics are discussed. Two control-related examples are selected in order to show the capabilities of the proposed approach., Comment: IECON 2019 - 45th Annual Conference of the IEEE Industrial Electronics Society

The detection and characterization of Earth-like exoplanets around Sun-like stars is a primary science motivation for the Habitable Worlds Observatory. However, the current best technology is not yet advanced enough to reach the 10^-10 contrasts at close angular separations and at the same time remain insensitive to low-order aberrations, as would be required to achieve high-contrast imaging of exo-Earths. Photonic technologies could fill this gap, potentially doubling exo-Earth yield. We review current work on photonic coronagraphs and investigate the potential of hybridized designs which combine both classical coronagraph designs and photonic technologies into a single optical system. We present two possible systems. First, a hybrid solution which splits the field of view spatially such that the photonics handle light within the inner working angle and a conventional coronagraph that suppresses starlight outside it. Second, a hybrid solution where the conventional coronagraph and photonics operate in series, complementing each other and thereby loosening requirements on each subsystem. As photonic technologies continue to advance, a hybrid or fully photonic coronagraph holds great potential for future exoplanet imaging from space., Comment: Conference Proceedings of SPIE: Techniques and Instrumentation for Detection of Exoplanets XI, vol. 12680 (2023)

Looking to the future of exo-Earth imaging from the ground, core technology developments are required in visible extreme adaptive optics (ExAO) to enable the observation of atmospheric features such as oxygen on rocky planets in visible light. UNDERGROUND (Ultra-fast AO techNology Determination for Exoplanet imageRs from the GROUND), a collaboration built in Feb. 2023 at the Optimal Exoplanet Imagers Lorentz Workshop, aims to (1) motivate oxygen detection in Proxima Centauri b and analogs as an informative science case for high-contrast imaging and direct spectroscopy, (2) overview the state of the field with respect to visible exoplanet imagers, and (3) set the instrumental requirements to achieve this goal and identify what key technologies require further development., Comment: SPIE Proceeding: 2023 / 12680-67

The JWST has captured the most detailed and sharpest infrared images ever taken of the inner region of the Orion Nebula, the nearest massive star formation region, and a prototypical highly irradiated dense photo-dissociation region (PDR). We investigate the fundamental interaction of far-ultraviolet photons with molecular clouds. The transitions across the ionization front (IF), dissociation front (DF), and the molecular cloud are studied at high-angular resolution. These transitions are relevant to understanding the effects of radiative feedback from massive stars and the dominant physical and chemical processes that lead to the IR emission that JWST will detect in many Galactic and extragalactic environments. Due to the proximity of the Orion Nebula and the unprecedented angular resolution of JWST, these data reveal that the molecular cloud borders are hyper structured at small angular scales of 0.1-1" (0.0002-0.002 pc or 40-400 au at 414 pc). A diverse set of features are observed such as ridges, waves, globules and photoevaporated protoplanetary disks. At the PDR atomic to molecular transition, several bright features are detected that are associated with the highly irradiated surroundings of the dense molecular condensations and embedded young star. Toward the Orion Bar PDR, a highly sculpted interface is detected with sharp edges and density increases near the IF and DF. This was predicted by previous modeling studies, but the fronts were unresolved in most tracers. A complex, structured, and folded DF surface was traced by the H2 lines. This dataset was used to revisit the commonly adopted 2D PDR structure of the Orion Bar. JWST provides us with a complete view of the PDR, all the way from the PDR edge to the substructured dense region, and this allowed us to determine, in detail, where the emission of the atomic and molecular lines, aromatic bands, and dust originate.