Search

Vehicular Edge Computing (VEC) is a key research area in autonomous driving. As Intelligent Transportation Systems (ITSs) continue to expand, ground vehicles (GVs) face the challenge of handling huge amounts of sensor data to drive safely. Specifically, due to energy and capacity limitations, GVs will need to offload resource-hungry tasks to external (cloud) computing units for faster processing. In 6th generation (6G) wireless systems, the research community is exploring the concept of Non-Terrestrial Networks (NTNs), where satellites can serve as space edge computing nodes to aggregate, store, and process data from GVs. In this paper we propose new data offloading strategies between a cluster of GVs and satellites in the Low Earth Orbits (LEOs), to optimize the trade-off between coverage and end-to-end delay. For the accuracy of the simulations, we consider real data and orbits from the Starlink constellation, one of the most representative and popular examples of commercial satellite deployments for communication. Our results demonstrate that Starlink satellites can support real-time offloading under certain conditions that depend on the onboard computational capacity of the satellites, the frame rate of the sensors, and the number of GVs., Comment: 7 pages, 6 figures, 1 table. This paper has been accepted for the 2025 IEEE Wireless Communications and Networking Conference (WCNC). \c{opyright} 2025 IEEE. Cite as: A. Bonora, A. Traspadini, M. Giordani, and M. Zorzi, "Performance Evaluation of Satellite-Based Data Offloading on Starlink Constellations," in Proc. IEEE Wireless Communications and Networking Conference (WCNC), 2025

Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind with broad subject coverage. HLE consists of 3,000 questions across dozens of subjects, including mathematics, humanities, and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable, but cannot be quickly answered via internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a significant gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai., Comment: 25 pages, 6 figures

The susceptibility to biases and discrimination is a pressing issue in today's labor markets. Though digital recruitment systems play an increasingly significant role in human resources management, thus far we lack a systematic understanding of human-centered design principles for fair online hiring. This work proposes a fair recruitment framework based on job-seekers' fairness concerns shared in an online forum. Through qualitative analysis, we uncover four overarching themes of job-seekers' fairness concerns, including discrimination against sensitive attributes, interaction biases, improper interpretations of qualifications, and power imbalance. Based on these findings, we derive design implications for algorithms and interfaces in recruitment systems, integrating them into a fair recruitment framework spanning different hiring stages and fairness considerations., Comment: 24 pages, 3 figures

Some previous studies based on IceCube neutrinos had found intriguing preliminary evidence that some of them might be GRB neutrinos with travel times affected by quantum properties of spacetime delaying them proportionally to their energy, an effect often labeled as "quantum-spacetime-induced in-vacuo dispersion". Those previous studies looked for candidate GRB neutrinos in a fixed (neutrino-energy-independent) time window after the GRB onset and relied rather crucially on crude estimates of the redshift of GRBs whose redshift has not been measured. We here introduce a complementary approach to the search of quantum-spacetime-affected GRB neutrinos which restricts the analysis to GRBs of sharply known redshift, and, in a way that we argue is synergistic with having sharp information on redshift, adopts a neutrino-energy-dependent time window. We find that knowing the redshift of the GRBs strengthens the analysis enough to compensate for the fact that of course the restriction to GRBs of known redshift reduces the number of candidate GRB neutrinos. And rather remarkably our estimate of the magnitude of the in-vacuo-dispersion effects is fully consistent with what had been found using the previous approach. Our findings are still inconclusive, since their significance is quantified by a $p$-value of little less than $0.01$, but provide motivation for monitoring the accrual of neutrino observations by IceCube and KM3NeT as well as for further refinements of the strategy of analysis here proposed.

Our concern is the data complexity of answering linear monadic datalog queries whose atoms in the rule bodies can be prefixed by operators of linear temporal logic LTL. We first observe that, for data complexity, answering any connected query with operators $\bigcirc/\bigcirc^-$ (at the next/previous moment) is either in AC0, or in $ACC0\!\setminus\!AC0$, or $NC^1$-complete, or LogSpace-hard and in NLogSpace. Then we show that the problem of deciding LogSpace-hardness of answering such queries is PSpace-complete, while checking membership in the classes AC0 and ACC0 as well as $NC^1$-completeness can be done in ExpSpace. Finally, we prove that membership in AC0 or in ACC0, $NC^1$-completeness, and LogSpace-hardness are undecidable for queries with operators $\Diamond_f/\Diamond_p$ (sometime in the future/past) provided that $NC^1 \ne NLogSpace$, and $LogSpace \ne NLogSpace$., Comment: Extended version of a paper accepted at ICDT'2025

Biological evolution continuously refines the design of species, resulting in highly optimised organisms over hundreds of millennia. Intuitively, we expect that random changes-evolution's primary mechanism-are more likely to be harmful than beneficial, leading to widespread detrimental effects in evolving species. The Evolvable Soma Theory of Ageing (ESTA) suggests that ageing is the cumulative result of these harmful effects, which predominantly cause bodily damage, while a few may lead to beneficial adaptations that evolution can exploit. While the disposable soma theory views ageing as a consequence of limited evolutionary pressure, ESTA posits that ageing is essentially evolution in action. In this study, we gather evidence supporting this theory through computer simulations. We conduct experiments using a platform where genes are linked to onset values that determine when they are expressed. Three scenarios are tested: one with single-point fitness evaluation, constant mutation rate and fixed gene onsets; one with single-point fitness evaluation, onset-dependent mutation rate and fixed gene onsets; and one with spread fitness evaluation, onset-dependent mutation rate and evolvable gene onsets. The last scenario, which embodies the evolvable soma hypothesis, demonstrates superior performance in both algorithmic efficiency and biological plausibility compared to the others.

Indium Arsenide is a III-V semiconductor with low electron effective mass, a small band gap, strong spin-orbit coupling, and a large g-factor. These properties and its surface Fermi level pinned in the conduction band make InAs a good candidate for developing superconducting solid-state quantum devices. Here, we report the epitaxial growth of very thin InAs layers with thicknesses ranging from 12.5 nm to 500 nm grown by Molecular Beam Epitaxy on InxAl1-xAs metamorphic buffers. Differently than InAs substrates, these buffers have the advantage of being insulating at cryogenic temperatures, which allows for multiple device operations on the same wafer and thus making the approach scalable. The structural properties of the InAs layers were investigated by high-resolution X-ray diffraction, demonstrating the high crystal quality of the InAs layers. Furthermore, their transport properties, such as total and sheet carrier concentration, sheet resistance, and carrier mobility, were measured in the van der Pauw configuration at room temperature. A simple conduction model was employed to quantify the surface, bulk, and interface contributions to the overall carrier concentration and mobility.

Introduced by Takagi and Watanabe, the F-pure threshold is an invariant defined in terms of the Frobenius homomorphism. While it finds applications in various settings, it is primarily used as a local invariant. The purpose of this note is to start filling this gap by opening the study of its behavior on a scheme. To this end, we define the defect of the F-pure threshold of a local ring $(R,\mathfrak{m})$ by setting ${\rm dfpt}(R)=\dim (R) - {\rm fpt}(\mathfrak{m})$. It turns out that this invariant defines an upper semi-continuous function on a scheme and satisfies Bertini-type theorems. We also study the behavior of the defect of the F-pure threshold under flat extensions and after blowing up the maximal ideal of a local ring.

We show that on quasi-smooth weighted complete intersections of codimension at most 3, any ample Cartier divisor $H$ such that $H-K_X$ is ample admits a nontrivial global section. This is done by proving a generalisation of a numerical conjecture formulated by Pizzato, Sano and Tasin, which relates the existence of global sections of $H$ to the Frobenius number of the numerical semigroup generated by the weights of the ambient projective space., Comment: 22 pages

We study a chemotaxis-consumption mechanism, in which some chemical signal and cells density interact each other. In order to control the concentration of such a population, sources involving gradient nonlinearities, which introduce a dampening effect on the model, are considered. Moreover, the system is characterized by nonlinear diffusion and sensitivity terms. We derive conditions on some data of the problem so to ensure the boundedness of related solutions.

In this work we present a novel linear and positivity preserving upwind discontinuous Galerkin (DG) approximation of a class of chemotaxis models with damping gradient nonlinearities. In particular, both a local and a nonlocal model including nonlinear diffusion, chemoattraction, chemorepulsion and logistic growth are considered. Some numerical experiments in the context of chemotactic collapse are presented, whose results are in accordance with the previous analysis of the approximation and show how the blow-up can be prevented by means of the damping gradient term., Comment: 20 pages, 7 figures

The generation of synthetic CT (sCT) images from cone-beam CT (CBCT) data using deep learning methodologies represents a significant advancement in radiation oncology. This systematic review, following PRISMA guidelines and using the PICO model, comprehensively evaluates the literature from 2014 to 2024 on the generation of sCT images for radiation therapy planning in oncology. A total of 35 relevant studies were identified and analyzed, revealing the prevalence of deep learning approaches in the generation of sCT. This review comprehensively covers synthetic CT generation based on CBCT and proton-based studies. Some of the commonly employed architectures explored are convolutional neural networks (CNNs), generative adversarial networks (GANs), transformers, and diffusion models. Evaluation metrics including mean absolute error (MAE), root mean square error (RMSE), peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) consistently demonstrate the comparability of sCT images with gold-standard planning CTs (pCT), indicating their potential to improve treatment precision and patient outcomes. Challenges such as field-of-view (FOV) disparities and integration into clinical workflows are discussed, along with recommendations for future research and standardization efforts. In general, the findings underscore the promising role of sCT-based approaches in personalized treatment planning and adaptive radiation therapy, with potential implications for improved oncology treatment delivery and patient care., Comment: 21 pages, 14 Figures, Accepted in the IEEE Transactions on Radiation and Plasma Medical Sciences

Recent observations from JWST have revealed an abundant population of active galactic nuclei (AGN) and so-called ``Little Red Dots'' (LRDs) at $2\lesssim z \lesssim 11$, many of which are characterized by V-shaped UV-to-optical continua with turnovers around the Balmer limit. The physical nature of these LRDs is unclear, and it remains debated whether the peculiar spectral shape originates from AGN, compact galaxies, or both. We present the analysis of new NIRSpec-IFU data from the BlackTHUNDER JWST Large Programme and archival NIRSpec-MSA data of a lensed LRD at $z=7.04$. The spectra confirm the presence of a smooth Balmer break and a broad H$\beta$ tracing the Broad Line Region (BLR) of an AGN. The small velocity dispersion of the H$\beta$ narrow component indicates a small dynamical mass of the host galaxy of $M_{\rm dyn}<4 \times 10^8~M_{\odot}$, which implies that the stellar population cannot contribute more than 10% to the optical continuum. We show that the Balmer break can be well described by an AGN continuum absorbed by very dense ($n_{\rm H}\sim 10^{10}~{\rm cm^{-3}}$) and nearly dust-free gas along our line-of-sight (possibly gas in the BLR or its surrounding). The same gas is expected to produce H$\beta$ absorption, at a level consistent with a tentative detection ($3\sigma$) in the high-resolution spectrum. Such a non-stellar origin of the Balmer break may apply to other LRDs, and would alleviate the issue of extremely high stellar mass surface densities inferred in the case of a stellar interpretation of the Balmer break. We note that this is a rare case of a black hole that is overmassive relative to both the host galaxy stellar and dynamical masses. We finally report indications of variability and the first attempt of AGN reverberation mapping at such an early epoch., Comment: 34 pages, 24 figures, submitted to MNRAS

Counterfactuals are a concept inherited from the field of logic and in general attain to the existence of causal relations between sentences or events. In particular, this concept has been introduced also in the context of interpretability in artificial intelligence, where counterfactuals refer to the minimum change to the feature values that changes the prediction of a classification model. The artificial intelligence framework of counterfactuals is mostly focused on machine learning approaches, typically neglecting the physics of the variables that determine a change in class. However, a theoretical formulation of counterfactuals in a control system framework - i.e., able to account for the mechanisms underlying a change in class - is lacking. To fill this gap, in this work we propose an original control system, physics-informed, theoretical foundation for counterfactuals, by means of the formulation of an optimal control problem. We apply the proposed methodology to a general glucose-insulin regulation model and results appear promising and pave the way to the possible integration with artificial intelligence techniques, with the aim of feeding machine learning models with the physics knowledge acquired through the system framework.

Large clinical evidence acknowledges the crucial role played by physical activity in delaying the progression of type-2 diabetes. However, the literature lacks control approaches that leverage exercise for type-2 diabetes control and more in general lacks a quantitative assessment of medical guidelines on the recommended amount of physical activity to be performed, mainly due to the absence of mathematical models that suitably estimate its benefits on diabetes progression. In this work, in order to provide a control-theoretical formulation of the exercise, we design a feedback law in terms of recommended physical activity, following a model predictive control approach, based on a widespread compact diabetes progression model, suitably modified to properly account for the long-term effect of the exercise. Moreover we illustrate how the proposed approach proves to show reliable robustness properties with respect to initial conditions and parameter perturbations, which may be used to reflect inter-patient variability. Results are encouraging in view of the validation of the control law on comprehensive high-dimensional models of diabetes progression, with the aim of translating the prediction of the controller into reasonable recommendations and to quantitatively support medical decision-making.

Galactic core-collapse supernovae (CCSNe) are a target for current generation gravitational wave detectors with an expected rate of 1-3 per century. The development of data analysis methods used for their detection relies deeply on the availability of waveform templates. However, realistic numerical simulations producing such waveforms are computationally expensive (millions of CPU hours and $10^2-10^3$~GB of memory), and only a few tens of them are available nowadays in the literature. We have developed a novel parametrized phenomenological waveform generator for CCSNe, ccphen v4, that reproduces the morphology of numerical simulation waveforms with low computational cost ($\sim 10$~ms CPU time and a few MB of memory use). For the first time, the phenomenological waveforms include polarization and the effect of several oscillation modes in the proto-neutron star. This is sufficient to describe the case of non-rotating progenitor cores, representing the vast majority of possible events. The waveforms include a stochastic component and are calibrated using numerical simulation data. The code is publicly available. Their main application is the training of neural networks used in detection pipelines, but other applications in this context are also discussed., Comment: 22 pages, 12 figures. Submitted to PRD

Faster-than-Nyquist (FTN) signaling aims at improving the spectral efficiency of wireless communication systems by exceeding the boundaries set by the Nyquist-Shannon sampling theorem. 50 years after its first introduction in the scientific literature, wireless communications have significantly changed, but spectral efficiency remains one of the key challenges. To adopt FTN signaling, inter-symbol interference (ISI) patterns need to be equalized at the receiver. Motivated by the pattern recognition capabilities of convolutional neural networks with skip connections, we propose such deep learning architecture for ISI equalization and symbol demodulation in FTN receivers. We investigate the performance of the proposed model considering quadrature phase shift keying modulation and low density parity check coding, and compare it to a set of benchmarks, including frequency-domain equalization, a quadratic-programming-based receiver, and an equalization scheme based on a deep neural network. We show that our receiver outperforms any benchmark, achieving error rates comparable to those in additive white Gaussian noise channel, and higher effective throughput, thanks to the increased spectral efficiency of FTN signaling. With a compression factor of 60% and code rate 3/4, the proposed model achieves a peak effective throughput of 2.5 Mbps at just 10dB of energy per bit over noise power spectral density ratio, with other receivers being limited by error floors due to the strong inter-symbol interference. To promote reproducibility in deep learning for wireless communications, our code is open source at the repository provided in the references., Comment: Submitted for publication

The introduction of quantum concepts is increasingly making its way into generative machine learning models. However, while there are various implementations of quantum Generative Adversarial Networks, the integration of quantum elements into diffusion models remains an open and challenging task. In this work, we propose a potential version of a quantum diffusion model that leverages the established idea of classical latent diffusion models. This involves using a traditional autoencoder to reduce images, followed by operations with variational circuits in the latent space. To effectively assess the benefits brought by quantum computing, the images generated by the quantum latent diffusion model have been compared to those generated by a classical model with a similar number of parameters, evaluated in terms of quantitative metrics. The results demonstrate an advantage in using a quantum version, as evidenced by obtaining better metrics for the images generated by the quantum version compared to those obtained by the classical version. Furthermore, quantum models continue to outperform even when considering small percentages of the dataset for training, demonstrating the quantum's ability to extract features more effectively even in a few shot learning scenario., Comment: 40 pages, 15 figures, 5 tables

Understanding which parts of a dynamical system cause each other is extremely relevant in fundamental and applied sciences. However, inferring causal links from observational data, namely without direct manipulations of the system, is still computationally challenging, especially if the data are high-dimensional. In this study we introduce a framework for constructing causal graphs from high-dimensional time series, whose computational cost scales linearly with the number of variables. The approach is based on the automatic identification of dynamical communities, groups of variables which mutually influence each other and can therefore be described as a single node in a causal graph. These communities are efficiently identified by optimizing the Information Imbalance, a statistical quantity that assigns a weight to each putative causal variable based on its information content relative to a target variable. The communities are then ordered starting from the fully autonomous ones, whose evolution is independent from all the others, to those that are progressively dependent on other communities, building in this manner a community causal graph. We demonstrate the computational efficiency and the accuracy of our approach on time-discrete and time-continuous dynamical systems including up to 80 variables.

We propose a novel framework where MeV-scale Dirac Dark Matter annihilates into axion-like particles, providing a natural explanation for the 511 keV gamma-ray line observed in the Galactic Center. The relic abundance is determined by p-wave annihilation into two axion-like particles, while s-wave annihilation into three axion-like particles, decaying into $e^+ e^-$ pairs, accounts for the line intensity. Remarkably, this model, assuming a standard Navarro-Frenk-White profile, reproduces the observed emission morphology, satisfies in-flight annihilation and cosmological bounds, and achieves the correct relic density, offering a compelling resolution to this longstanding anomaly., Comment: 12 pages, 6 figures

A breakdown of Lorentz Invariance is predicted by some theories as a result of quantum physics applied to gravity or for the presence of vector/tensor fields, beside the metric tensor of General Relativity, mediating the gravitational interaction. A violation of Lorentz Invariance would manifest itself, in the gravitational sector, mainly through a non-zero value of the post-Newtonian parameter $\alpha_1$, that is null in General Relativity. In this paper we describe a test of Local Lorentz Invariance considering the effect of a possible preferred frame, that of the cosmic microwave background, on the orbit of the LAGEOS II satellite tracked by the Satellite Laser Ranging technique. We constrain this parameter down to the level of $3\times10^{-5}$, improving a previous limit obtained through the Lunar Laser Ranging technique., Comment: 6 pages, 4 figures. Under review in Physical Review Letters

Response functions are a key quantity to describe the near-equilibrium dynamics of strongly-interacting many-body systems. Recent techniques that attempt to overcome the challenges of calculating these \emph{ab initio} have employed expansions in terms of orthogonal polynomials. We employ a neural network prediction algorithm to reconstruct a response function $S(\omega)$ defined over a range in frequencies $\omega$. We represent the calculated response function as a truncated Chebyshev series whose coefficients can be optimized to reduce the representation error. We compare the quality of response functions obtained using coefficients calculated using a neural network (NN) algorithm with those computed using the Gaussian Integral Transform (GIT) method. In the regime where only a small number of terms in the Chebyshev series are retained, we find that the NN scheme outperforms the GIT method.

Long-lived heavy particles present during the big bang could have a decay channel opened by gravitons. Such decays can produce gravitational waves with large enough abundance to be detectable, and a peculiar narrow spectrum peaked today around optical frequencies. We identify which particles can decay in one or two gravitons. The maximal gravitational wave abundance arises from theories with extra hidden strong gauge dynamics, such as a confining pure-glue group. An interesting abundance also arises in theories with perturbative couplings. Future observation might shed light on early cosmology and allow some spectroscopy of sub-Planckian gravitationally-decaying particles, plausibly present in a variety of theories such as gauge unification, supersymmetry, extra dimensions, strings., Comment: 22 pages. Webinar presentation at https://youtu.be/GUndxZChJY8

We study the error arising in the numerical approximation of FBSDEs and related PIDEs by means of a deep learning-based method. Our results focus on decoupled FBSDEs with jumps and extend the seminal work of HAn and Long (2020) analyzing the numerical error of the deep BSDE solver proposed in E et al. (2017). We provide a priori and a posteriori error estimates for the finite and infinite activity case., Comment: 33 pages

Online motion planning is a challenging problem for intelligent robots moving in dense environments with dynamic obstacles, e.g., crowds. In this work, we propose a novel approach for optimal and safe online motion planning with minimal information about dynamic obstacles. Specifically, our approach requires only the current position of the obstacles and their maximum speed, but it does not need any information about their exact trajectories or dynamic model. The proposed methodology combines Monte Carlo Tree Search (MCTS), for online optimal planning via model simulations, with Velocity Obstacles (VO), for obstacle avoidance. We perform experiments in a cluttered simulated environment with walls, and up to 40 dynamic obstacles moving with random velocities and directions. With an ablation study, we show the key contribution of VO in scaling up the efficiency of MCTS, selecting the safest and most rewarding actions in the tree of simulations. Moreover, we show the superiority of our methodology with respect to state-of-the-art planners, including Non-linear Model Predictive Control (NMPC), in terms of improved collision rate, computational and task performance.

Planning for autonomous systems typically requires reasoning with models at different levels of abstraction, and the harmonization of two competing sets of objectives: high-level mission goals that refer to an interaction of the system with the external environment, and low-level platform constraints that aim to preserve the integrity and the correct interaction of the subsystems. The complicated interplay between these two models makes it very hard to reason on the system as a whole, especially when the objective is to find plans with robustness guarantees, considering the non-deterministic behavior of the lower layers of the system. In this paper, we introduce the problem of Platform-Aware Mission Planning (PAMP), addressing it in the setting of temporal durative actions. The PAMP problem differs from standard temporal planning for its exists-forall nature: the high-level plan dealing with mission goals is required to satisfy safety and executability constraints, for all the possible non-deterministic executions of the low-level model of the platform and the environment. We propose two approaches for solving PAMP. The first baseline approach amalgamates the mission and platform levels, while the second is based on an abstraction-refinement loop that leverages the combination of a planner and a verification engine. We prove the soundness and completeness of the proposed approaches and validate them experimentally, demonstrating the importance of heterogeneous modeling and the superiority of the technique based on abstraction-refinement.

Effective fall risk assessment is critical for post-stroke patients. The present study proposes a novel, data-informed fall risk assessment method based on the instrumented Timed Up and Go (ITUG) test data, bringing in many mobility measures that traditional clinical scales fail to capture. IFRA, which stands for Instrumented Fall Risk Assessment, has been developed using a two-step process: first, features with the highest predictive power among those collected in a ITUG test have been identified using machine learning techniques; then, a strategy is proposed to stratify patients into low, medium, or high-risk strata. The dataset used in our analysis consists of 142 participants, out of which 93 were used for training (15 synthetically generated), 17 for validation and 32 to test the resulting IFRA scale (22 non-fallers and 10 fallers). Features considered in the IFRA scale include gait speed, vertical acceleration during sit-to-walk transition, and turning angular velocity, which align well with established literature on the risk of fall in neurological patients. In a comparison with traditional clinical scales such as the traditional Timed Up & Go and the Mini-BESTest, IFRA demonstrates competitive performance, being the only scale to correctly assign more than half of the fallers to the high-risk stratum (Fischer's Exact test p = 0.004). Despite the dataset's limited size, this is the first proof-of-concept study to pave the way for future evidence regarding the use of IFRA tool for continuous patient monitoring and fall prevention both in clinical stroke rehabilitation and at home post-discharge., Comment: 26 pages, 2 figures, submitted for review dec 2024

Dark Matter can interact with electroweak gauge bosons via higher-dimensional operators, in spite of being neutral under gauge interactions, much like neutral atoms interact with photons through Rayleigh scattering. This study explores effective interactions between a real scalar Dark Matter particle, singlet under the SM gauge group, and electroweak gauge bosons. We present a comprehensive analysis of current constraints and projected sensitivities from both lepton and hadron colliders as well as direct and indirect detection experiments in testing Rayleigh Dark Matter interactions. We find that, thanks to the complementarity between collider experiments and cosmological probes, thermally produced Rayleigh Dark Matter at the hundreds of GeV scale can be thoroughly tested with the next generation of experiments. For lighter candidates, upcoming forecasts will explore uncharted parameter space, significantly surpassing the thermal Dark Matter benchmark., Comment: 39 pages, 14 figures

Free-space quantum key distribution (QKD) represents a groundbreaking advancement in secure communication, enabling secure key exchange over vast distances and offering robust encryption for the future quantum internet. However, the compatibility between fiber and free-space infrastructures continues to pose challenges for QKD protocols. Indeed, free-space and fiber-based networks commonly use different wavelengths and qubits encoding schemes. On the one hand, free-space QKD typically exploits visible light for its beneficial beam divergence compared to longer wavelengths, and polarization encoding for its robustness against turbulence. On the other hand, fiber-based QKD employs infrared light, particularly the C-band, because it shows the minimum losses with silica fibers, and time-bin encoding, due to polarization instability in optical fibers. In our study, we demonstrate the viability of a time-bin encoded QKD protocol operating in the C-band through horizontal turbulent free-space channels. We test the setup into a 50 m and a 500 m long links, achieving an average secure key rate of, respectively, 793 kbps and 40 kbps over several hours of measurements. The results encourage further exploration of the interoperability between free-space and fiber-based infrastructures, opening new possibilities for connecting terminal users with satellites in hybrid infrastructures., Comment: 10 pages, 6 figures

Since 2018, there has been a consistent decline in the distance traveled by U.S. manufacturing imports, reaching a level not observed since 2008. This trend is the result of the substitution away from imports from China and towards imports from closer countries. At the same time, U.S. manufacturing inventory-to-sales ratio has continued to rise. These trends are at odds with the literature, which finds that reductions in the distance of imports are associated with a decline in inventories. We argue that a rise in delivery time risk, driven by longer and more frequent delays and supply disruptions, can reconcile these trends. We do so in the context of a model of global sourcing with stochastic delivery times and inventories. Firms trade off the lower price of farther inputs with the increase in exposure to demand volatility and longer delays. In response, firms increase their inventories. Yet, as delivery delays rise, firms need to carry more inventories per unit of the input used. We calibrate the model for the period from 2018 to 2024 using data on the increase in tariffs for inputs from China, and the rise in inventories over sales. We find an increase in delivery delays for foreign inputs of 21 days across the period. The rise in delays and tariffs had an output loss of 7.3% and a price increase of 1.8%. Of these, the rise in delivery delays alone generated a 2.6% drop in output and a 0.4% increase in prices.

A key challenge in Deep Reinforcement Learning is sample efficiency, especially in real-world applications where collecting environment interactions is expensive or risky. Recent off-policy algorithms improve sample efficiency by increasing the Update-To-Data (UTD) ratio and performing more gradient updates per environment interaction. While this improves sample efficiency, it significantly increases computational cost due to the higher number of gradient updates required. In this paper we propose a sample-efficient method to improve computational efficiency by separating training into distinct learning phases in order to exploit gradient updates more effectively. Our approach builds on top of the Dropout Q-Functions (DroQ) algorithm and alternates between an online, low UTD ratio training phase, and an offline stabilization phase. During the stabilization phase, we fine-tune the Q-functions without collecting new environment interactions. This process improves the effectiveness of the replay buffer and reduces computational overhead. Our experimental results on continuous control problems show that our method achieves results comparable to state-of-the-art, high UTD ratio algorithms while requiring 56\% fewer gradient updates and 50\% less training time than DroQ. Our approach offers an effective and computationally economical solution while maintaining the same sample efficiency as the more costly, high UTD ratio state-of-the-art.

As REST APIs have become widespread in modern web services, comprehensive testing of these APIs has become increasingly crucial. Due to the vast search space consisting of operations, parameters, and parameter values along with their complex dependencies and constraints, current testing tools suffer from low code coverage, leading to suboptimal fault detection. To address this limitation, we present a novel tool, AutoRestTest, which integrates the Semantic Operation Dependency Graph (SODG) with Multi-Agent Reinforcement Learning (MARL) and large language models (LLMs) for effective REST API testing. AutoRestTest determines operation-dependent parameters using the SODG and employs five specialized agents (operation, parameter, value, dependency, and header) to identify dependencies of operations and generate operation sequences, parameter combinations, and values. AutoRestTest provides a command-line interface and continuous telemetry on successful operation count, unique server errors detected, and time elapsed. Upon completion, AutoRestTest generates a detailed report highlighting errors detected and operations exercised. In this paper, we introduce our tool and present preliminary results., Comment: To be published in the 47th IEEE/ACM International Conference on Software Engineering - Demonstration Track (ICSE-Demo 2025)

Modern web services rely heavily on REST APIs, typically documented using the OpenAPI specification. The widespread adoption of this standard has resulted in the development of many black-box testing tools that generate tests based on these specifications. Recent advancements in Natural Language Processing (NLP), particularly with Large Language Models (LLMs), have enhanced REST API testing by extracting actionable rules and generating input values from the human-readable portions of the specification. However, these advancements overlook the potential of continuously refining the identified rules and test inputs based on server responses. To address this limitation, we present LlamaRestTest, a novel approach that employs two custom LLMs to generate realistic test inputs and uncover parameter dependencies during the testing process by incorporating server responses. These LLMs are created by fine-tuning the Llama3-8b model, using mined datasets of REST API example values and inter-parameter dependencies. We evaluated LlamaRestTest on 12 real-world services (including popular services such as Spotify), comparing it against RESTGPT, a GPT-powered specification-enhancement tool, as well as several state-of-the-art REST API testing tools, including RESTler, MoRest, EvoMaster, and ARAT-RL. Our results show that fine-tuning enables smaller LLMs to outperform larger models in detecting actionable rules and generating inputs for REST API testing. We evaluated configurations from the base Llama3-8B to fine-tuned versions and explored 2-bit, 4-bit, and 8-bit quantization for efficiency. LlamaRestTest surpasses state-of-the-art tools in code coverage and error detection, even with RESTGPT-enhanced specifications, and an ablation study highlights the impact of its novel components., Comment: To be published in the ACM International Conference on the Foundations of Software Engineering (FSE 2025)

An experimental study was conducted in the CICLoPE long-pipe facility to investigate the correlation between wall-pressure and turbulent velocity fluctuations in the logarithmic region, at high friction Reynolds numbers ($4\,794 \lesssim Re_\tau \lesssim 47\,015$). Hereby we explore the scalability of employing wall-pressure to effectively estimate off-the-wall velocity states (e.g., to be of use in real-time control of wall-turbulence). Coherence spectra for wall-pressure and streamwise (or wall-normal) velocity fluctuations collapse when plotted against $\lambda_x/y$ and thus reveals a Reynolds-number-independent scaling with distance-from-the-wall. When the squared wall-pressure fluctuations are considered instead of the linear wall-pressure term, the coherence spectra for the wall-pressure--squared and velocity are higher in amplitude at wavelengths corresponding to large-scale streamwise velocity fluctuations (e.g., at $\lambda_x/y = 60$ the coherence value increases from roughly 0.1 up to 0.3). This higher coherence typifies a modulation effect, because low-frequency content is introduced when squaring the wall-pressure time series. Finally, quadratic stochastic estimation is employed to estimate turbulent velocity fluctuations from the wall-pressure time series only. For each $Re_\tau$ investigated, the estimated time series and a true temporal measurement of velocity inside the turbulent pipe flow, yield a normalized correlation coefficient of $\rho \approx 0.6$ for all cases. This suggests that wall-pressure sensing can be employed for meaningful estimation of off-the-wall velocity fluctuations, and thus for real-time control of energetic turbulent velocity fluctuations at high $Re_\tau$ applications.

This roadmap consolidates recent advances while exploring emerging applications, reflecting the remarkable diversity of hardware platforms, neuromorphic concepts, and implementation philosophies reported in the field. It emphasizes the critical role of cross-disciplinary collaboration in this rapidly evolving field.

Human pose estimation, a vital task in computer vision, involves detecting and localising human joints in images and videos. While single-frame pose estimation has seen significant progress, it often fails to capture the temporal dynamics for understanding complex, continuous movements. We propose Poseidon, a novel multi-frame pose estimation architecture that extends the ViTPose model by integrating temporal information for enhanced accuracy and robustness to address these limitations. Poseidon introduces key innovations: (1) an Adaptive Frame Weighting (AFW) mechanism that dynamically prioritises frames based on their relevance, ensuring that the model focuses on the most informative data; (2) a Multi-Scale Feature Fusion (MSFF) module that aggregates features from different backbone layers to capture both fine-grained details and high-level semantics; and (3) a Cross-Attention module for effective information exchange between central and contextual frames, enhancing the model's temporal coherence. The proposed architecture improves performance in complex video scenarios and offers scalability and computational efficiency suitable for real-world applications. Our approach achieves state-of-the-art performance on the PoseTrack21 and PoseTrack18 datasets, achieving mAP scores of 88.3 and 87.8, respectively, outperforming existing methods.

Device driver bugs are the leading cause of OS compromises, and their formal verification is therefore highly desirable. To the best of our knowledge, no realistic and performant driver has been verified for a non-trivial device. We propose Pancake, an imperative language for systems programming that features a well-defined and verification-friendly semantics. Leveraging the verified compiler backend of the CakeML functional language, we develop a compiler for Pancake that guarantees that the binary retains the semantics of the source code. Usng automatic translation of Pancake to the Viper SMT front-end, we verify a performant driver for an Ethernet NIC., Comment: 15 pages, 5 figures

Core-collapse supernovae (SNe) provide a unique environment to study Feebly Interacting Particles (FIPs) such as Axion-Like Particles (ALPs), sterile neutrinos, and Dark Photons (DPs). This paper focuses on heavy FIPs produced in SNe, whose decay produces electrons and positrons, generating observable secondary signals during their propagation and annihilation. We focus on the In-flight Annihilation (IA) of positrons, which emerge as the most significant contribution to the resulting gamma-ray spectrum. Using data from COMPTEL and EGRET we derive the most stringent bounds on the FIP-electron couplings for heavy ALPs, sterile neutrinos, and DPs. These results strenghten existing bounds of one or two orders of magnitude, depending on the FIP model., Comment: 15 figures, 10 pages. Comments are welcome

When discussing urban life, pedestrian accessibility to all main services is crucial for fostering social interactions, promoting healthy lifestyles, and reducing pollution. This is especially relevant in coherent urban agglomerations like university campuses, which feature a high concentration of streets and social facilities. Using Wi-Fi data, we study pedestrian movements within a confined geometric network representing the pathways on a university campus. We estimate the level of crowding in each arc of the network and identify pedestrian flows along all possible paths, measuring the entropy and robustness of the network. In particular, we calculate the information gain achieved through the use of Wi-Fi data and we assess how pedestrian traffic redistributes within the network after the removal of individual arcs. Our results can be used to facilitate the investigation of the current state of walkability across the university campus while also testing a set of methods for analyzing urban complex networks, potentially allowing us to pinpoint areas in urgent need of road maintenance and enhancement., Comment: 25 pages, 16 figures

Even in a simple stochastic process, the study of the full distribution of time integrated observables can be a difficult task. This is the case of a much-studied process such as the Ornstein-Uhlenbeck process where, recently, anomalous dynamical scaling of large deviations of time integrated functionals has been highlighted. Using the mapping of a continuous stochastic process to a continuous time random walk via the "excursions technique'', we introduce a comprehensive formalism that enables the calculation of the complete distribution of the time-integrated observable $A = \int_0^T v^n(t) dt$, where $n$ is a positive integer and $v(t)$ is the random velocity of a particle following Ornstein-Uhlenbeck dynamics. We reveal an interesting connection between the anomalous rate function associated with the observable $A$ and the statistics of the area under the first-passage functional during an excursion. The rate function of the latter, analyzed here for the first time, exhibits anomalous scaling behavior and a dynamical phase transition, both of which are explored in detail. The case of the anomalous scaling of large deviations, originally associated to the presence of an instantonic solution in the weak noise regime of a path integral approach, is here produced by a so called "big jump effect'', in which the contribution to rare events is dominated by the largest excursion. Our approach, which is quite general for continuous stochastic processes, allows to associate a physical meaning to the anomalous scaling of large deviations, through the big jump principle., Comment: 26 pages, 6 figures

Music has always been central to human culture, reflecting and shaping traditions, emotions, and societal changes. Technological advancements have transformed how music is created and consumed, influencing tastes and the music itself. In this study, we use Network Science to analyze musical complexity. Drawing on $\approx20,000$ MIDI files across six macro-genres spanning nearly four centuries, we represent each composition as a weighted directed network to study its structural properties. Our results show that Classical and Jazz compositions have higher complexity and melodic diversity than recently developed genres. However, a temporal analysis reveals a trend toward simplification, with even Classical and Jazz nearing the complexity levels of modern genres. This study highlights how digital tools and streaming platforms shape musical evolution, fostering new genres while driving homogenization and simplicity.

A Vec-variety is a suitable functor from finite-dimensional vector spaces to finite-dimensional varieties. Most varieties in the geometry of tensors, e.g. the variety of d-way tensors of slice rank at most r, are of this form. We prove that the singular locus of a Vec-variety is a proper closed Vec-subvariety, analogously to the situation for ordinary finite-dimensional varieties. Via earlier work of the third author, this implies that these singular loci admit a description by finitely many polynomial equations. A natural follow-up question to our main result is whether a Vec-variety also admits a suitably functorial resolution of singularities. We establish some preliminary results in this direction in the regime where the dimension of evaluations of a Vec-variety grows linearly with that of the input vector space.

Spectral densities encode essential information about system-environment interactions in open-quantum systems, playing a pivotal role in shaping the system's dynamics. In this work, we leverage machine learning techniques to reconstruct key environmental features, going beyond the weak-coupling regime by simulating the system's dynamics using the reaction coordinate mapping. For a dissipative spin-boson model with a structured spectral density expressed as a sum of Lorentzian peaks, we demonstrate that the time evolution of a system observable can be used by a neural network to classify the spectral density as comprising one, two, or three Lorentzian peaks and accurately predict their central frequency., Comment: 15+6 pages, 11 figures. Comments welcome!

This paper presents a novel approach combining inductive logic programming with reinforcement learning to improve training performance and explainability. We exploit inductive learning of answer set programs from noisy examples to learn a set of logical rules representing an explainable approximation of the agent policy at each batch of experience. We then perform answer set reasoning on the learned rules to guide the exploration of the learning agent at the next batch, without requiring inefficient reward shaping and preserving optimality with soft bias. The entire procedure is conducted during the online execution of the reinforcement learning algorithm. We preliminarily validate the efficacy of our approach by integrating it into the Q-learning algorithm for the Pac-Man scenario in two maps of increasing complexity. Our methodology produces a significant boost in the discounted return achieved by the agent, even in the first batches of training. Moreover, inductive learning does not compromise the computational time required by Q-learning and learned rules quickly converge to an explanation of the agent policy.

The advantages of temporal networks in capturing complex dynamics, such as diffusion and contagion, has led to breakthroughs in real world systems across numerous fields. In the case of human behavior, face-to-face interaction networks enable us to understand the dynamics of how communities emerge and evolve in time through the interactions, which is crucial in fields like epidemics, sociological studies and urban science. However, state-of-the-art datasets suffer from a number of drawbacks, such as short time-span for data collection and a small number of participants. Moreover, concerns arise for the participants' privacy and the data collection costs. Over the past years, many successful algorithms for static networks generation have been proposed, but they often do not tackle the social structure of interactions or their temporal aspect. In this work, we extend a recent network generation approach to capture the evolution of interactions between different communities. Our method labels nodes based on their community affiliation and constructs surrogate networks that reflect the interactions of the original temporal networks between nodes with different labels. This enables the generation of synthetic networks that replicate realistic behaviors. We validate our approach by comparing structural measures between the original and generated networks across multiple face-to-face interaction datasets.

Entanglement distribution in quantum networks will enable next-generation technologies for quantum-secured communications, distributed quantum computing and sensing. Future quantum networks will require dense connectivity, allowing multiple users to share entanglement in a reconfigurable and multiplexed manner, while long-distance connections are established through the teleportation of entanglement, or entanglement swapping. While several recent works have demonstrated fully connected, local multi-user networks based on multiplexing, extending this to a global network architecture of interconnected local networks remains an outstanding challenge. Here we demonstrate the next stage in the evolution of multiplexed quantum networks: a prototype global reconfigurable network where entanglement is routed and teleported in a flexible and multiplexed manner between two local multi-user networks composed of four users each. At the heart of our network is a programmable 8x8-dimensional multi-port circuit that harnesses the natural mode-mixing process inside a multi-mode fibre to implement on-demand high-dimensional operations on two independent photons carrying eight transverse-spatial modes. Our circuit design allows us to break away from the limited planar geometry and bypass the control and fabrication challenges of conventional integrated photonic platforms. Our demonstration showcases the potential of this architecture for enabling large-scale, global quantum networks that offer versatile connectivity while being fully compatible with an existing communications infrastructure.

This experimental work demonstrates multipartite quantum correlation in bright frequency combs out of a microresonator integrated on silicon nitride working above its oscillation threshold. Multipartite features, going beyond so far reported two-mode correlation, naturally arise due to a cascade of non-linear optical processes, making a single-color laser pump sufficient to initiate their generation. Our results show the transition from two-mode to multipartite correlation, witnessed by noise reductions as low as $-2.5$\,dB and $-2$\,dB, respectively, compared to corresponding classical levels. A constant of the movement of the non-linear interaction Hamiltonian is identified and used to asses the multipartite behavior. Reported demonstrations pave the way to next generation on-chip multipartite sources for quantum technologies applications., Comment: 4 figures

We consider a dense assembly of repulsive particles whose fluctuating sizes are subject to a landscape that defines three species: particles with finite size, either of type $A$ or of type $B$, and point particles $\varnothing$. We show that the nonequilibrium synchronization of sizes systematically leads to a homogeneous configuration associated with the survival of a single species. Remarkably, the relaxation towards such a configuration features a transient phase separation. By delineating and analyzing the dominant kinetic factors at play during relaxation, we recapitulate the phase diagram of species survival in terms of the parameters of the size landscape. Finally, we obtain a hydrodynamic mapping to equilibrium by coarse-graining the microscopic dynamics, which leads to predicting the nature of the transitions between various regimes where distinct species survive., Comment: 17 pages, 7 figures

Threat analysis is continuously growing in importance due to the always-increasing complexity and frequency of cyber attacks. Analyzing threats demands significant effort from security experts, leading to delays in the security analysis process. Different cybersecurity knowledge bases are currently available to support this task but manual efforts are often required to correlate such heterogenous sources into a unified view that would enable a more comprehensive assessment. To address this gap, we propose a methodology leveraging Natural Language Processing (NLP) to effectively and efficiently associate Common Vulnerabilities and Exposure (CVE) vulnerabilities with Common Attack Pattern Enumeration and Classification (CAPEC) attack patterns. The proposed technique combines semantic similarity with keyword analysis to improve the accuracy of association estimations. Experimental evaluations demonstrate superior performance compared to state-of-the-art models, reducing manual effort and analysis time, and enabling cybersecurity professionals to prioritize critical tasks., Comment: 20 pages

Split Learning (SL) is a distributed deep learning approach enabling multiple clients and a server to collaboratively train and infer on a shared deep neural network (DNN) without requiring clients to share their private local data. The DNN is partitioned in SL, with most layers residing on the server and a few initial layers and inputs on the client side. This configuration allows resource-constrained clients to participate in training and inference. However, the distributed architecture exposes SL to backdoor attacks, where malicious clients can manipulate local datasets to alter the DNN's behavior. Existing defenses from other distributed frameworks like Federated Learning are not applicable, and there is a lack of effective backdoor defenses specifically designed for SL. We present SafeSplit, the first defense against client-side backdoor attacks in Split Learning (SL). SafeSplit enables the server to detect and filter out malicious client behavior by employing circular backward analysis after a client's training is completed, iteratively reverting to a trained checkpoint where the model under examination is found to be benign. It uses a two-fold analysis to identify client-induced changes and detect poisoned models. First, a static analysis in the frequency domain measures the differences in the layer's parameters at the server. Second, a dynamic analysis introduces a novel rotational distance metric that assesses the orientation shifts of the server's layer parameters during training. Our comprehensive evaluation across various data distributions, client counts, and attack scenarios demonstrates the high efficacy of this dual analysis in mitigating backdoor attacks while preserving model utility., Comment: To appear at NDSS 2025; 18 pages, 6 Tables, and 11 figures