Search

Here we investigate a general class of massless local theories of spin-2 and spin-3 both invariant under generalized transverse diffeomorphisms (TDiff). We identify the ghost free region in their parameters space and show the relationship of those models with the ``doublet'' action stemming from the tensionless limit of the open bosonic string field theory (for symmetric tensors). The connection is implemented via a non local field redefinition which introduces a Stuckelberg-like field of rank-0 (rank-1) for the spin-2 (spin-3) case. An apparent mismatch between most of the TDiff models and the ``doublet'' action has led us to prove a nontrivial duality within the TDiff models which restores the equivalence. Any point in the parameters subspace of ghost free TDiff models is equivalent to any other one within that subspace. In particular, they are all physically equivalent to their simplest version known as Maxwell-like models. So the physical TDiff models seem to differ from each other by a BRST cohomologically trivial term., Comment: 24 pages, 1 table

Here we suggest a new local action describing arbitrary integer spin-$s$ massless particles in terms of only two symmetric fields $\varphi $ and $\alpha$ of rank-$s$ and $(s-3)$ respectively. It is an unconstrained version of the Fronsdal theory where the double traceless constraint on the physical field is evaded via a rank-$(s-4)$ Weyl like symmetry. The constrained higher spin diffeomorphism is enlarged to full diffeomorphism via the Stueckelberg field $\alpha$ through an appropriate field redefinition. After a partial gauge fixing where the Weyl symmetry is broken while preserving diffeomorphisms, the field equations reproduce, for arbitrary integer spin-$s$, diffeomorphism invariant equations of motion previously obtained via a truncation of the spectrum of the open bosonic string field theory in the tensionless limit. In the $s=4$ case we show that the functional integration over $\alpha$ leads to a unique non local Weyl and diffeomorphism invariant action given only in terms of the physical field $\varphi$ whose spectrum is confirmed via an analysis of the analytic structure of the spin-4 propagator for which we introduce a complete basis of projection and transition non local differential operators. We also show that the elimination of $\alpha$ after the Weyl gauge fixing leads to a non local diffeomorphism invariant action previously obtained in the literature., Comment: 27 pages, no figures

This work describes our group's submission to the PROCESS Challenge 2024, with the goal of assessing cognitive decline through spontaneous speech, using three guided clinical tasks. This joint effort followed a holistic approach, encompassing both knowledge-based acoustic and text-based feature sets, as well as LLM-based macrolinguistic descriptors, pause-based acoustic biomarkers, and multiple neural representations (e.g., LongFormer, ECAPA-TDNN, and Trillson embeddings). Combining these feature sets with different classifiers resulted in a large pool of models, from which we selected those that provided the best balance between train, development, and individual class performance. Our results show that our best performing systems correspond to combinations of models that are complementary to each other, relying on acoustic and textual information from all three clinical tasks.

Stellar surface heterogeneities, such as spots and faculae, often contaminate exoplanet transit spectra, hindering precise atmospheric characterization. We demonstrate a novel, epoch-based, model-independent method to mitigate stellar contamination, applicable to multi-planet systems with at least one airless planet. We apply this method using quasi-simultaneous transits of TRAPPIST-1 b and TRAPPIST-1 c observed on July 9, 2024, with JWST NIRSpec PRISM. These two planets, with nearly identical radii and impact parameters, are likely either bare rocks or possess thin, low-pressure atmospheres, making them ideal candidates for this technique, as variations in their transit spectra would be primarily attributed to stellar activity. Our observations reveal their transit spectra exhibit consistent features, indicating similar levels of stellar contamination. We use TRAPPIST-1 b to correct the transit spectrum of TRAPPIST-1 c, achieving a 2.5x reduction in stellar contamination at shorter wavelengths. At longer wavelengths, lower SNR prevents clear detection of contamination or full assessment of mitigation. Still, out-of-transit analysis reveals variations across the spectrum, suggesting contamination extends into the longer wavelengths. Based on the success of the correction at shorter wavelengths, we argue that contamination is also reduced at longer wavelengths to a similar extent. This shifts the challenge of detecting atmospheric features to a predominantly white noise issue, which can be addressed by stacking observations. This method enables epoch-specific stellar contamination corrections, allowing co-addition of planetary spectra for reliable searches of secondary atmospheres with signals of 60-250 ppm. Additionally, we identify small-scale cold (2000 K) and warm (2600 K) regions almost uniformly distributed on TRAPPIST-1, with overall covering fractions varying by 0.1% per hour., Comment: Resubmitted to ApJL after addressing reviewer comments

The diagnosis of diseases in food crops based on machine learning seemed satisfactory and suitable for use on a large scale. The Convolutional Neural Networks (CNNs) perform accurately in the disease prediction considering the image capture of the crop leaf, being extensively enhanced in the literature. These machine learning techniques fall short in data privacy, as they require sharing the data in the training process with a central server, disregarding competitive or regulatory concerns. Thus, Federated Learning (FL) aims to support distributed training to address recognized gaps in centralized training. As far as we know, this paper inaugurates the use and evaluation of FL applied in maize leaf diseases. We evaluated the performance of five CNNs trained under the distributed paradigm and measured their training time compared to the classification performance. In addition, we consider the suitability of distributed training considering the volume of network traffic and the number of parameters of each CNN. Our results indicate that FL potentially enhances data privacy in heterogeneous domains.

Two-dimensional (2D) carbon materials have been intensively investigated because of their distinctive structural framework and electronic behaviors as alternatives in energy conversion and storage applications. This study proposes a novel 2D carbon allotrope, Polymerized Pyracyclene Graphyne (PolyPyGY), characterized by a multi-ringed structure with 4-, 5-, 6-, 8-, and 16-membered rings comprising a porous structure. Using first-principles calculations and machine-learning techniques, we explore its structural, electronic, mechanical, optical, and lithium-ion storing properties. The vibrational properties assessed through the density functional perturbation theory framework confirm its structural stability. Moreover, ab initio molecular dynamics simulations at 1000 K demonstrate its thermal resilience, with no bond breaking or reconfiguration observed. The electronic band structure reveals a metallic nature, and the material exhibits anisotropic elastic properties, with Young's modulus varying between 421 and 664 GPa, suggesting good mechanical stability. Furthermore, lithium diffusion studies indicate low energy barriers (0.05-0.9 eV) and a high diffusion coefficient ( $>$ 6 $\times$ 10$^{-6}$ cm$^{2}$/s), along with a stable open circuit voltage of 1.2 V. These results highlight PolyPyGY's potential as a highly effective and durable anode material for lithium-ion batteries, featuring rapid Li-ion diffusion, stable intercalation, and consistent performance during charge and discharge cycles., Comment: 18 pages

This paper builds upon J. Nilsson's classification of rank one $\mathcal{U}(\mathfrak{h})$-free modules by extending the analysis to modules without rank restrictions, focusing on the category $\mathfrak{A}$ of $\mathcal{U}(\mathfrak{h})$-finite $\mathfrak{g}$-modules. A deeper investigation of the weighting functor $\mathcal{W}$ and its left derived functors, $\mathcal{W}_*$, led to the proof that simple $\mathcal{U}(\mathfrak{h})$-finite modules of infinite dimension are $\mathcal{U}(\mathfrak{h})$-torsion free. Furthermore, it is shown that these modules are $\mathcal{U}(\mathfrak{h})$-free if they possess non-integral or singular central characters. It is concluded that the existence of $\mathcal{U}(\mathfrak{h})$-torsion-free $\mathfrak{g}$-modules is restricted to Lie algebras of types A and C. The concept of an almost-coherent family, which generalizes O. Mathieu's definition of coherent families, is introduced. It is proved that $\mathcal{W}(M)$, for a $\mathcal{U}(\mathfrak{h})$-torsion-free module $M$, falls within this class of weight modules. Furthermore, a notion of almost-equivalence is defined to establish a connection between irreducible semi-simple almost-coherent families and O. Mathieu's original classification. Progress is also made in classifying simple modules within the category $\mathfrak{A}^{\text{irr}}$, which consists of $\mathcal{U}(\mathfrak{h})$-finite modules $M$ with the property that $\mathcal{W}(M)$ is an irreducible almost-coherent family. A complete classification is achieved for type C, with partial classification for type A. Finally, a conjecture is presented asserting that all simple $\mathfrak{sl}(n+1)$-modules in $\mathfrak{A}^{\text{irr}}$ are isomorphic to simple subquotients of exponential tensor modules, and supporting results are proved., Comment: 43 pages

A group element is called generalized torsion if a finite product of its conjugates is equal to the identity. We show that in a finitely generated abelian-by-finite group, an element is generalized torsion if and only if its image in the abelianization is torsion. We also prove a quantitative version with sharp bounds to the generalized exponent of these groups. In particular, we provide many examples of finitely presentable torsion-free groups in which all elements are generalized torsion. We also discuss positive generalized identities in abelian-by-finite groups and related classes.

To train generalist robot policies, machine learning methods often require a substantial amount of expert human teleoperation data. An ideal robot for humans collecting data is one that closely mimics them: bimanual arms and dexterous hands. However, creating such a bimanual teleoperation system with over 50 DoF is a significant challenge. To address this, we introduce Bidex, an extremely dexterous, low-cost, low-latency and portable bimanual dexterous teleoperation system which relies on motion capture gloves and teacher arms. We compare Bidex to a Vision Pro teleoperation system and a SteamVR system and find Bidex to produce better quality data for more complex tasks at a faster rate. Additionally, we show Bidex operating a mobile bimanual robot for in the wild tasks. The robot hands (5k USD) and teleoperation system (7k USD) is readily reproducible and can be used on many robot arms including two xArms (16k USD). Website at https://bidex-teleop.github.io/, Comment: In CoRL 2024. Website at https://bidex-teleop.github.io/

For a uniformly locally finite metric space $(X, d)$, we investigate \emph{coarse} flows on its uniform Roe algebra $\mathrm{C}^*_u(X)$, defined as one-parameter groups of automorphisms whose differentiable elements include all partial isometries arising from partial translations on $X$. We first show that any flow $\sigma$ on $\mathrm{C}^*_u(X)$ corresponds to a (possibly unbounded) self-adjoint operator $h$ on $\ell_2(X)$ such that $\sigma_t(a) = e^{ith} a e^{-ith}$ for all $t \in \mathbb{R}$, allowing us to focus on operators $h$ that generate flows on $ \mathrm{C}^*_u (X)$. Assuming Yu's property A, we prove that a self-adjoint operator $h$ on $\ell_2(X)$ induces a coarse flow on $\mathrm{C}^*_u(X)$ if and only if $h$ can be expressed as $h = a + d$, where $a \in \mathrm{C}^*_u(X)$ and $d$ is a diagonal operator with entries forming a coarse function on $X$. We further study cocycle equivalence and cocycle perturbations of coarse flows, showing that, under property A, any coarse flow is a cocycle perturbation of a diagonal flow. Finally, for self-adjoint operators $h$ and $k$ that induce coarse flows on $\mathrm{C}^*_u(X)$, we characterize conditions under which the associated flows are either cocycle perturbations of each other or cocycle conjugate. In particular, if $h - k$ is bounded, then the flow induced by $h$ is a cocycle perturbation of the flow induced by $k$.

Using methods of associative algebras, Lie theory, group cohomology, and modular representation theory, we construct profinite $p$-adic analytic groups such that the centralizer of each of their non-trivial elements is abelian. The paper answers questions of P.~Shumyatsky, P.~Zalesskii, and T.~Zapata in the Israel J. Math., v.~230, 2019., Comment: 23 pages

[Context and Motivation] Several studies have investigated attributes of great software practitioners. However, the investigation of such attributes is still missing in Requirements Engineering (RE). The current knowledge on attributes of great software practitioners might not be easily translated to the context of RE because its activities are, usually, less technical and more human-centered than other software engineering activities. [Question/Problem] This work aims to investigate which are the attributes of great requirements engineers, the relationship between them, and strategies that can be employed to obtain these attributes. We follow a method composed of a survey with 18 practitioners and follow up interviews with 11 of them. [Principal Ideas/Results] Investigative ability in talking to stakeholders, judicious, and understand the business are the most commonly mentioned attributes amongst the set of 22 attributes identified, which were grouped into four categories. We also found 38 strategies to improve RE skills. Examples are training, talking to all stakeholders, and acquiring domain knowledge. [Contribution] The attributes, their categories, and relationships are organized into a map. The relations between attributes and strategies are represented in a Sankey diagram. Software practitioners can use our findings to improve their understanding about the role and responsibilities of requirements engineers.

M dwarf systems offer a unique opportunity to study terrestrial exoplanetary atmospheres due to their smaller size and cooler temperatures. However, due to the extreme conditions these host stars impose, it is unclear whether their small, close-in rocky planets are able to retain any atmosphere at all. The Hot Rocks Survey aims to answer this question by targeting nine different M dwarf rocky planets spanning a range of planetary and stellar properties. LHS 1478 b orbits an M3-type star, has an equilibrium temperature of Teq = 585 K and experiences an instellation 21 times greater than that of Earth. We observe two secondary eclipses using photometric imaging at 15 um using the Mid-Infrared Instrument on the James Webb Space Telescope (JWST MIRI) to measure thermal emission from the dayside of the planet. We then compare these values to different atmospheric scenarios to evaluate potential heat transport and CO2 absorption signatures. We find a secondary eclipse depth of 146 +/- 56 ppm based on the first observation, while the second observation results in a non-detection due to significantly larger unexplained systematics. Based on the first observation alone, we can reject the null hypothesis of the dark (zero Bond albedo) no atmosphere bare rock model with a confidence level of 3.4 sigma. For an airless body with a Bond albedo of A=0.2, the significance decreases to 2.9 sigma. The secondary eclipse depth is consistent with the majority of atmospheric scenarios we considered, which all involve atmospheres which include different concentrations of CO2, and surface pressures from 0.1 to 10 bar. However, we stress that the two observations from our program do not yield consistent results, and more observations are needed to verify our findings. The Hot Rocks Survey serves as a relevant primer for future endeavors such as the Director's Discretionary Time (DDT) Rocky Worlds program., Comment: Submitted to A&A

Federated Kolmogorov-Arnold Networks (F-KANs) have already been proposed, but their assessment is at an initial stage. We present a comparison between KANs (using B-splines and Radial Basis Functions as activation functions) and Multi- Layer Perceptrons (MLPs) with a similar number of parameters for 100 rounds of federated learning in the MNIST classification task using non-IID partitions with 100 clients. After 15 trials for each model, we show that the best accuracies achieved by MLPs can be achieved by Spline-KANs in half of the time (in rounds), with just a moderate increase in computing time., Comment: 10 pages, 5 figures, for associated code see https://github.com/artsasse/fedkan

In this study, we propose a novel approach to uncover subgroup-specific and subgroup-common latent factors addressing the challenges posed by the heterogeneity of neurological and mental disorders, which hinder disease understanding, treatment development, and outcome prediction. The proposed approach, sparse Group Factor Analysis (GFA) with regularised horseshoe priors, was implemented with probabilistic programming and can uncover associations (or latent factors) among multiple data modalities differentially expressed in sample subgroups. Synthetic data experiments showed the robustness of our sparse GFA by correctly inferring latent factors and model parameters. When applied to the Genetic Frontotemporal Dementia Initiative (GENFI) dataset, which comprises patients with frontotemporal dementia (FTD) with genetically defined subgroups, the sparse GFA identified latent disease factors differentially expressed across the subgroups, distinguishing between "subgroup-specific" latent factors within homogeneous groups and "subgroup common" latent factors shared across subgroups. The latent disease factors captured associations between brain structure and non-imaging variables (i.e., questionnaires assessing behaviour and disease severity) across the different genetic subgroups, offering insights into disease profiles. Importantly, two latent factors were more pronounced in the two more homogeneous FTD patient subgroups (progranulin (GRN) and microtubule-associated protein tau (MAPT) mutation), showcasing the method's ability to reveal subgroup-specific characteristics. These findings underscore the potential of sparse GFA for integrating multiple data modalities and identifying interpretable latent disease factors that can improve the characterization and stratification of patients with neurological and mental health disorders., Comment: 38 pages, 14 figures

A rigorous treatment of light-matter interactions typically requires an interacting quantum field theory. However, most applications of interest are handled using classical or semiclassical models, which are valid only when quantum-field fluctuations can be neglected. This approximation breaks down in scenarios involving large light intensities or degenerate matter, where additional quantum effects become significant. In this work, we address these limitations by developing a quantum kinetic framework that treats both light and matter fields on equal footing, naturally incorporating both linear and nonlinear interactions. To accurately account for light fluctuations, we introduce a photon distribution function that, together with the classical electromagnetic fields, provides a better description of the photon fluid. From this formalism, we derive kinetic equations from first principles that recover classical electrodynamical results while revealing couplings that are absent in the corresponding classical theory. Furthermore, by addressing the Coulomb interaction in the Hartree-Fock approximation, we include the role of fermionic exchange exactly in both kinetic and fluid regimes through a generalized Fock potential. The latter provides corrections not only to the electrostatic forces but also to the plasma velocity fields, which become significant in degenerate conditions., Comment: Quantum kinetic theory, light matter interaction, quantum degeneracy, warm dense matter, laser-plasma interaction

The classical theory for the meta-analysis of $p$-values is based on the assumption that if the overall null hypothesis is true, then all $p$-values used in a chosen combined test statistic are genuine, i.e., are observations from independent and identically distributed standard uniform random variables. However, the pressure felt by most researchers to publish, which is worsen by publication bias, can originate fake $p$-values to be reported, usually Beta(1,2) distributed. In general, the existence of fake $p$-values in a sample of $p$-values to be combined is unknown, and if, for some reason, there is information that they do exist, their number will most likely be unknown as well. Moreover, even if fake $p$-values are accounted for, the cumulative distribution function of classical combined test statistics does not have a closed-form expression that facilitates its practical usage. To overcome this problem, tables with estimated critical values are supplied for the commonly used combined tests for the meta-analysis of $p$-values when a few of them are fake ones, i.e., Beta(1,2) distributed., Comment: 68 pages, 6 figures

We present a fully autonomous real-world RL framework for mobile manipulation that can learn policies without extensive instrumentation or human supervision. This is enabled by 1) task-relevant autonomy, which guides exploration towards object interactions and prevents stagnation near goal states, 2) efficient policy learning by leveraging basic task knowledge in behavior priors, and 3) formulating generic rewards that combine human-interpretable semantic information with low-level, fine-grained observations. We demonstrate that our approach allows Spot robots to continually improve their performance on a set of four challenging mobile manipulation tasks, obtaining an average success rate of 80% across tasks, a 3-4 improvement over existing approaches. Videos can be found at https://continual-mobile-manip.github.io/, Comment: CoRL 2024. Website at https://continual-mobile-manip.github.io/

This study reports a comprehensive environmental scan of the generative AI (GenAI) infrastructure in the national network for clinical and translational science across 36 institutions supported by the Clinical and Translational Science Award (CTSA) Program led by the National Center for Advancing Translational Sciences (NCATS) of the National Institutes of Health (NIH) at the United States. With the rapid advancement of GenAI technologies, including large language models (LLMs), healthcare institutions face unprecedented opportunities and challenges. This research explores the current status of GenAI integration, focusing on stakeholder roles, governance structures, and ethical considerations by administering a survey among leaders of health institutions (i.e., representing academic medical centers and health systems) to assess the institutional readiness and approach towards GenAI adoption. Key findings indicate a diverse range of institutional strategies, with most organizations in the experimental phase of GenAI deployment. The study highlights significant variations in governance models, with a strong preference for centralized decision-making but notable gaps in workforce training and ethical oversight. Moreover, the results underscore the need for a more coordinated approach to GenAI governance, emphasizing collaboration among senior leaders, clinicians, information technology staff, and researchers. Our analysis also reveals concerns regarding GenAI bias, data security, and stakeholder trust, which must be addressed to ensure the ethical and effective implementation of GenAI technologies. This study offers valuable insights into the challenges and opportunities of GenAI integration in healthcare, providing a roadmap for institutions aiming to leverage GenAI for improved quality of care and operational efficiency.

Through Monte Carlo simulations and the Associating Lattice Gas Model, the phases of a two-dimensional fluid under hydrophilic confinement are evaluated. The model, in its unconfined version, reproduces the anomalous behavior of water regarding its density, diffusion, and solubility, among other dynamic and thermodynamic properties. Extreme confinements suppress phase transitions since fluctuations suppress ordering. The fluid under hydrophilic confinement forms a single wetting layer that gradually wets the wall. From the wetting layer, the low-density liquid structure is formed. The confined fluid presents a first-order liquid-liquid transition, but always at lower temperatures than that observed in the bulk.

The current paradigm for motion planning generates solutions from scratch for every new problem, which consumes significant amounts of time and computational resources. For complex, cluttered scenes, motion planning approaches can often take minutes to produce a solution, while humans are able to accurately and safely reach any goal in seconds by leveraging their prior experience. We seek to do the same by applying data-driven learning at scale to the problem of motion planning. Our approach builds a large number of complex scenes in simulation, collects expert data from a motion planner, then distills it into a reactive generalist policy. We then combine this with lightweight optimization to obtain a safe path for real world deployment. We perform a thorough evaluation of our method on 64 motion planning tasks across four diverse environments with randomized poses, scenes and obstacles, in the real world, demonstrating an improvement of 23%, 17% and 79% motion planning success rate over state of the art sampling, optimization and learning based planning methods. Video results available at mihdalal.github.io/neuralmotionplanner, Comment: Website at mihdalal.github.io/neuralmotionplanner. Main paper: 7 pages, 4 figures, 2 tables. Appendix: 9 pages, 5 figures, 6 tables

Although human evaluation remains the gold standard for open-domain dialogue evaluation, the growing popularity of automated evaluation using Large Language Models (LLMs) has also extended to dialogue. However, most frameworks leverage benchmarks that assess older chatbots on aspects such as fluency and relevance, which are not reflective of the challenges associated with contemporary models. In fact, a qualitative analysis on Soda, a GPT-3.5 generated dialogue dataset, suggests that current chatbots may exhibit several recurring issues related to coherence and commonsense knowledge, but generally produce highly fluent and relevant responses. Noting the aforementioned limitations, this paper introduces Soda-Eval, an annotated dataset based on Soda that covers over 120K turn-level assessments across 10K dialogues, where the annotations were generated by GPT-4. Using Soda-Eval as a benchmark, we then study the performance of several open-access instruction-tuned LLMs, finding that dialogue evaluation remains challenging. Fine-tuning these models improves performance over few-shot inferences, both in terms of correlation and explanation., Comment: Accepted to EMNLP2024 (findings)

We describe and analyze a quantum thermometer based on a multi-layered collisional model. The proposed architecture provides significant sensitivity even for short interaction times between the ancillae comprised in the thermometer and the system to be probed, and a small number of information-acquiring collisions. The assessment of the flow of information taking place within the layered thermometer and between system and thermometer reveals that the tuning of the mutual backflow of information has a positive influence on the precision of thermometry, and helps unveiling the information-theoretic mechanisms behind the working principles of the proposed architecture.

The clouds have a great impact on Venus's energy budget and climate evolution, but its three-dimensional structure is still not well understood. Here we incorporate a simple Venus cloud physics scheme into a flexible GCM to investigate the three-dimensional cloud spatial variability. Our simulations show good agreement with observations in terms of the vertical profiles of clouds and H2SO4 vapor. H2O vapor is overestimated above the clouds due to efficient transport in the cloud region. The cloud top decreases as latitude increases, qualitatively consistent with Venus Express observations. The underlying mechanism is the combination of H2SO4 chemical production and meridional circulation. The mixing ratios of H2SO4 at 50-60 km and H2O vapors in the main cloud deck basically exhibit maxima around the equator, due to the effect of temperature's control on the saturation vapor mixing ratios of the two species. The cloud mass distribution is subject to both H2SO4 chemical production and dynamical transport and shows a pattern that peaks around the equator in the upper cloud while peaks at mid-high latitudes in the middle cloud. At low latitudes, H2SO4 and H2O vapors, cloud mass loading and acidity show semidiurnal variations at different altitude ranges, which can be validated against future missions. Our model emphasizes the complexity of the Venus climate system and the great need for more observations and simulations to unravel its spatial variability and underlying atmospheric and/or geological processes., Comment: 34 pages, 8 figures, 8 supporting figures. Accepted by JGR-Planets

Background: The use of gray literature (GL) has grown in software engineering research, especially in studies that consider Questions and Answers (Q&A) sites, since software development professionals widely use them. Though snowballing (SB) techniques are standard in systematic literature reviews, little is known about how to apply them to gray literature reviews. Aims: This paper investigates how to use SB approaches on Q&A sites during gray literature reviews to identify new valid discussions for analysis. Method: In previous studies, we compiled and analyzed a set of Stack Exchange Project Management (SEPM) discussions related to software engineering technical debt (TD). Those studies used a data set consisting of 108 valid discussions extracted from SEPM. Based on this start data set, we perform forward and backward SB using two different approaches: link-based and similarity-based SB. We then compare the precision and recall of those two SB approaches against the search-based approach of the original study. Results: In just one snowballing iteration, the approaches yielded 291 new discussions for analysis, 130 of which were considered valid for our study. That is an increase of about 120% over the original data set (recall). The SB process also yielded a similar rate of valid discussion retrieval when compared to the search-based approach (precision). Conclusion: This paper provides guidelines on how to apply two SB approaches to find new valid discussions for review. To our knowledge, this is the first study that analyzes the use of SB on Q&A websites. By applying SB, it was possible to identify new discussions, significantly increasing the relevant data set for a gray literature review., Comment: 10 pages, 6 figures

Despite being heralded as the new standard for dialogue evaluation, the closed-source nature of GPT-4 poses challenges for the community. Motivated by the need for lightweight, open source, and multilingual dialogue evaluators, this paper introduces GenResCoh (Generated Responses targeting Coherence). GenResCoh is a novel LLM generated dataset comprising over 130k negative and positive responses and accompanying explanations seeded from XDailyDialog and XPersona covering English, French, German, Italian, and Chinese. Leveraging GenResCoh, we propose ECoh (Evaluation of Coherence), a family of evaluators trained to assess response coherence across multiple languages. Experimental results demonstrate that ECoh achieves multilingual detection capabilities superior to the teacher model (GPT-3.5-Turbo) on GenResCoh, despite being based on a much smaller architecture. Furthermore, the explanations provided by ECoh closely align in terms of quality with those generated by the teacher model., Comment: Accepted to SIGDIAL 2024

In this work, we test a possible redshift variation of the electron-to-proton mass ratio, $\mu = m_e/m_p$, directly from galaxy cluster gas mass fraction measurements and type Ia Supernovae observations. Our analysis is completely independent of any cosmological model. Our result reveals no variation of $\mu$ within 1 $\sigma$ confidence level. From the point of view of Particle Physics, we can use the precision on these results to constrain the parameter space of models beyond the Standard Model of electroweak interactions. We exemplify this by focusing in a specific Two Higgs Doublet model (2HDM), where the second scalar doublet couples exclusively to leptons. An important parameter in the model concerns the ratio between its vacuum expectation values, defined by $\tan\beta$. In our approach we can constrain the inverse parameter (cot$\beta$) to an optimal value, (tan$\beta)^{-1}=$ 0.02127 $\pm$ 0.0029, with the largest vacuum expectation value for 2HDM, $v_2$, estimated at around 240.033 $\pm$ 0.21~GeV. Also, by taking into account the $(g-2)_\mu$ discrepancy found between theory and experiment, we can reduce the validity region for this model and establish bounds on the scalar masses, in the light of our findings from galaxy clusters data for $\mu$. This study contributes valuable insights to the understanding of Particle Physics and Astrophysics interface, establishing a new interplay between data from large scale structure of the Universe and subatomic Physics., Comment: 14 pages, 8 figures

We have made significant progress towards building foundational video diffusion models. As these models are trained using large-scale unsupervised data, it has become crucial to adapt these models to specific downstream tasks. Adapting these models via supervised fine-tuning requires collecting target datasets of videos, which is challenging and tedious. In this work, we utilize pre-trained reward models that are learned via preferences on top of powerful vision discriminative models to adapt video diffusion models. These models contain dense gradient information with respect to generated RGB pixels, which is critical to efficient learning in complex search spaces, such as videos. We show that backpropagating gradients from these reward models to a video diffusion model can allow for compute and sample efficient alignment of the video diffusion model. We show results across a variety of reward models and video diffusion models, demonstrating that our approach can learn much more efficiently in terms of reward queries and computation than prior gradient-free approaches. Our code, model weights,and more visualization are available at https://vader-vid.github.io., Comment: Project Webpage: https://vader-vid.github.io; Code available at: https://github.com/mihirp1998/VADER

This work introduces an approach to enhancing the computational efficiency of 3D atmospheric simulations by integrating a machine-learned surrogate model into the OASIS global circulation model (GCM). Traditional GCMs, which are based on repeatedly numerically integrating physical equations governing atmospheric processes across a series of time-steps, are time-intensive, leading to compromises in spatial and temporal resolution of simulations. This research improves upon this limitation, enabling higher resolution simulations within practical timeframes. Speeding up 3D simulations holds significant implications in multiple domains. Firstly, it facilitates the integration of 3D models into exoplanet inference pipelines, allowing for robust characterisation of exoplanets from a previously unseen wealth of data anticipated from JWST and post-JWST instruments. Secondly, acceleration of 3D models will enable higher resolution atmospheric simulations of Earth and Solar System planets, enabling more detailed insights into their atmospheric physics and chemistry. Our method replaces the radiative transfer module in OASIS with a recurrent neural network-based model trained on simulation inputs and outputs. Radiative transfer is typically one of the slowest components of a GCM, thus providing the largest scope for overall model speed-up. The surrogate model was trained and tested on the specific test case of the Venusian atmosphere, to benchmark the utility of this approach in the case of non-terrestrial atmospheres. This approach yields promising results, with the surrogate-integrated GCM demonstrating above 99.0% accuracy and 147 factor GPU speed-up of the entire simulation compared to using the matched original GCM under Venus-like conditions., Comment: 20 pages, 14 figures

Large Language Models (LLMs) have showcased remarkable capabilities in various Natural Language Processing tasks. For automatic open-domain dialogue evaluation in particular, LLMs have been seamlessly integrated into evaluation frameworks, and together with human evaluation, compose the backbone of most evaluations. However, existing evaluation benchmarks often rely on outdated datasets and evaluate aspects like Fluency and Relevance, which fail to adequately capture the capabilities and limitations of state-of-the-art chatbot models. This paper critically examines current evaluation benchmarks, highlighting that the use of older response generators and quality aspects fail to accurately reflect modern chatbot capabilities. A small annotation experiment on a recent LLM-generated dataset (SODA) reveals that LLM evaluators such as GPT-4 struggle to detect actual deficiencies in dialogues generated by current LLM chatbots., Comment: Accepted to the 6th NLP for Conversational AI workshop at ACL

The high-energy radiative output, from the X-ray to the ultraviolet, of exoplanet host stars drives photochemical reactions and mass loss in the upper regions of planetary atmospheres. In order to place constraints on the atmospheric properties of the three closest terrestrial exoplanets transiting M dwarfs, we observe the high-energy spectra of the host stars LTT1445A and GJ486 in the X-ray with XMM-Newton and Chandra and in the ultraviolet with HST/COS and STIS. We combine these observations with estimates of extreme ultraviolet flux, reconstructions of the Ly-a lines, and stellar models at optical and infrared wavelengths to produce panchromatic spectra from 1A--20um for each star. While LTT1445Ab, LTT1445Ac, and GJ486b do not possess primordial hydrogen-dominated atmospheres, we calculate that they are able to retain pure CO2 atmospheres if starting with 10, 15, and 50% of Earth's total CO2 budget, respectively, in the presence of their host stars' stellar wind. We use age-activity relationships to place lower limits of 2.2 and 6.6 Gyr on the ages of the host stars LTT1445A and GJ486. Despite both LTT1445A and GJ486 appearing inactive at optical wavelengths, we detect flares at ultraviolet and X-ray wavelengths for both stars. In particular, GJ486 exhibits two flares with absolute energies of 10^29.5 and 10^30.1 erg (equivalent durations of 4357+/-96 and 19724+/-169 s) occurring three hours apart, captured with HST/COS G130M. Based on the timing of the observations, we suggest that these high-energy flares are related and indicative of heightened flaring activity that lasts for a period of days, but our interpretations are limited by sparse time-sampling. Consistent high-energy monitoring is needed to determine the duration and extent of high-energy activity on individual M dwarfs, as well as the population as a whole., Comment: 21 pages, published in A&A

Testing Highly Configurable Systems (HCSs) is a challenging task, especially in an evolution scenario where features are added, changed, or removed, which hampers test case selection and prioritization. Existing work is usually based on the variability model, which is not always available or updated. Yet, the few existing approaches rely on links between test cases and changed files (or lines of code), not considering how features are implemented, usually spread over several and unchanged files. To overcome these limitations, we introduce FeaTestSelPrio, a feature-oriented test case selection and prioritization approach for HCSs. The approach links test cases to feature implementations, using HCS pre-processor directives, to select test cases based on features affected by changes in each commit. After, the test cases are prioritized according to the number of features they cover. Our approach selects a greater number of tests and takes longer to execute than a changed-file-oriented approach, used as baseline, but FeaTestSelPrio performs better regarding detected failures. By adding the approach execution time to the execution time of the selected test cases, we reached a reduction of $\approx$50%, in comparison with retest-all. The prioritization step allows reducing the average test budget in 86% of the failed commits.

The recent integration of visual capabilities into Large Language Models (LLMs) has the potential to play a pivotal role in science and technology education, where visual elements such as diagrams, charts, and tables are commonly used to improve the learning experience. This study investigates the performance of ChatGPT-4 Vision, OpenAI's most advanced visual model at the time the study was conducted, on the Bachelor in Computer Science section of Brazil's 2021 National Undergraduate Exam (ENADE). By presenting the model with the exam's open and multiple-choice questions in their original image format and allowing for reassessment in response to differing answer keys, we were able to evaluate the model's reasoning and self-reflecting capabilities in a large-scale academic assessment involving textual and visual content. ChatGPT-4 Vision significantly outperformed the average exam participant, positioning itself within the top 10 best score percentile. While it excelled in questions that incorporated visual elements, it also encountered challenges with question interpretation, logical reasoning, and visual acuity. The involvement of an independent expert panel to review cases of disagreement between the model and the answer key revealed some poorly constructed questions containing vague or ambiguous statements, calling attention to the critical need for improved question design in future exams. Our findings suggest that while ChatGPT-4 Vision shows promise in multimodal academic evaluations, human oversight remains crucial for verifying the model's accuracy and ensuring the fairness of high-stakes educational exams. The paper's research materials are publicly available at https://github.com/nabormendonca/gpt-4v-enade-cs-2021., Comment: Accepted for publication

Cyber-Physical Systems (CPS) consist of software interacting with the physical world, such as robots, vehicles, and industrial processes. CPS are frequently responsible for the safety of lives, property, or the environment, and so software correctness must be determined with a high degree of certainty. To that end, simply testing a CPS is insufficient, as its interactions with the physical world may be difficult to predict, and unsafe conditions may not be immediately obvious. Formal verification can provide stronger safety guarantees but relies on the accuracy of the verified system in representing the real system. Bringing together verification and implementation can be challenging, as languages that are typically used to implement CPS are not easy to formally verify, and languages that lend themselves well to verification often abstract away low-level implementation details. Translation between verification and implementation languages is possible, but requires additional assurances in the translation process and increases software complexity; having both in a single language is desirable. This paper presents a formalization of MARVeLus, a CPS language which combines verification and implementation. We develop a metatheory for its synchronous refinement type system and demonstrate verified synchronous programs executing on real systems.

Spectroscopic phase curves of hot Jupiters measure their emission spectra at multiple orbital phases, thus enabling detailed characterisation of their atmospheres. Precise constraints on the atmospheric composition of these exoplanets offer insights into their formation and evolution. We analyse four phase-resolved emission spectra of the hot Jupiter WASP-43b, generated from a phase curve observed with the MIRI/LRS onboard the JWST, to retrieve its atmospheric properties. Using a parametric 2D temperature model and assuming a chemically homogeneous atmosphere within the observed pressure region, we simultaneously fit the four spectra to constrain the abundances of atmospheric constituents, thereby yielding more precise constraints than previous work that analysed each spectrum independently. Our analysis reveals statistically significant evidence of NH3 (4$\sigma$) in a hot Jupiter's emission spectra for the first time, along with evidence of H2O (6.5$\sigma$), CO (3.1$\sigma$), and a non-detection of CH4. With our abundance constraints, we tentatively estimate the metallicity of WASP-43b at 0.6-6.5$\times$solar and its C/O ratio at 0.6-0.9. Our findings offer vital insights into the atmospheric conditions and formation history of WASP-43b by simultaneously constraining the abundances of carbon, oxygen, and nitrogen-bearing species., Comment: 17 pages, 14 figures, accepted for publication in MNRAS. Comments welcome!