Neuro-Symbolic Artificial Intelligence (NeSy AI) has emerged as a promising direction for integrating neural learning with symbolic reasoning. Typically, in the probabilistic variant of such systems, a neural network first extracts a set of symbols from sub-symbolic input, which are then used by a symbolic component to reason in a probabilistic manner towards answering a query. In this work, we address the problem of formally verifying the robustness of such NeSy probabilistic reasoning systems, therefore paving the way for their safe deployment in critical domains. We analyze the complexity of solving this problem exactly, and show that a decision version of the core computation is $\mathrm{NP}^{\mathrm{PP}}$-complete. In the face of this result, we propose the first approach for approximate, relaxation-based verification of probabilistic NeSy systems. We demonstrate experimentally on a standard NeSy benchmark that the proposed method scales exponentially better than solver-based solutions and apply our technique to a real-world autonomous driving domain, where we verify a safety property under large input dimensionalities.

Vacuum polarisation effects induced by the sourceless solutions of the Yang-Mills equation are investigated. These solutions represent oppositely oriented chromomagnetic flux tubes permeating the space in all directions. We review and prove the gauge invariance of the effective Lagrangian on sourceless gauge fields and present a number of alternative methods allowing to compute the effective Lagrangian. We compute the effective Lagrangian on chromomagnetic flux tube solutions and demonstrate that the effective Lagrangian has a universal form that supports the stability of the chromomagnetic flux tubes condensation and indicates that the Yang-Mills vacuum is a highly degenerate state. The stability is a result of the quartic nonlinear self-interaction of the negative mode completely eliminating the instability and the imaginary term from the effective Lagrangian in chromomagnetic field. It is suggested that the condensate of chromomagnetic flux tubes represents a dual analog of the Cooper pairs condensate in a superconductor.

In this paper, we study the factors that contribute to the effect of oversmoothing in deep Graph Neural Networks (GNNs). Specifically, our analysis is based on a new metric (Mean Average Squared Distance - $MASED$) to quantify the extent of oversmoothing. We derive layer-wise bounds on $MASED$, which aggregate to yield global upper and lower distance bounds. Based on this quantification of oversmoothing, we further analyze the importance of two different properties of the model; namely the norms of the generated node embeddings, along with the largest and smallest singular values of the weight matrices. Building on the insights drawn from the theoretical analysis, we show that oversmoothing increases as the number of trainable weight matrices and the number of adjacency matrices increases. We also use the derived layer-wise bounds on $MASED$ to form a proposal for decoupling the number of hops (i.e., adjacency depth) from the number of weight matrices. In particular, we introduce G-Reg, a regularization scheme that increases the bounds, and demonstrate through extensive experiments that by doing so node classification accuracy increases, achieving robustness at large depths. We further show that by reducing oversmoothing in deep networks, we can achieve better results in some tasks than using shallow ones. Specifically, we experiment with a ``cold start" scenario, i.e., when there is no feature information for the unlabeled nodes. Finally, we show empirically the trade-off between receptive field size (i.e., number of weight matrices) and performance, using the $MASED$ bounds. This is achieved by distributing adjacency hops across a small number of trainable layers, avoiding the extremes of under- or over-parameterization of the GNN.

Jet flavour classification is of paramount importance for a broad range of applications in modern-day high-energy-physics experiments, particularly at the LHC. In this paper we propose a novel architecture for this task that exploits modern deep learning techniques. This new model, called DeepJet, overcomes the limitations in input size that affected previous approaches. As a result, the heavy flavour classification performance improves, and the model is extended to also perform quark-gluon tagging.

Practical parameter identifiability in ODE-based epidemiological models is a known issue, yet one that merits further study. It is essentially ubiquitous due to noise and errors in real data. In this study, to avoid uncertainty stemming from data of unknown quality, simulated data with added noise are used to investigate practical identifiability in two distinct epidemiological models. Particular emphasis is placed on the role of initial conditions, which are assumed unknown, except those that are directly measured. Instead of just focusing on one method of estimation, we use and compare results from various broadly used methods, including maximum likelihood and Markov Chain Monte Carlo (MCMC) estimation. Among other findings, our analysis revealed that the MCMC estimator is overall more robust than the point estimators considered. Its estimates and predictions are improved when the initial conditions of certain compartments are fixed so that the model becomes globally identifiable. For the point estimators, whether fixing or fitting the that are not directly measured improves parameter estimates is model-dependent. Specifically, in the standard SEIR model, fixing the initial condition for the susceptible population S(0) improved parameter estimates, while this was not true when fixing the initial condition of the asymptomatic population in a more involved model. Our study corroborates the change in quality of parameter estimates upon usage of pre-peak or post-peak time-series under consideration. Finally, our examples suggest that in the presence of significantly noisy data, the value of structural identifiability is moot.

Coarse graining (CG) enables the investigation of molecular properties for larger systems and at longer timescales than the ones attainable at the atomistic resolution. Machine learning techniques have been recently proposed to learn CG particle interactions, i.e. develop CG force fields. Graph representations of molecules and supervised training of a graph convolutional neural network architecture are used to learn the potential of mean force through a force matching scheme. In this work, the force acting on each CG particle is correlated to a learned representation of its local environment that goes under the name of SchNet, constructed via continuous filter convolutions. We explore the application of SchNet models to obtain a CG potential for liquid benzene, investigating the effect of model architecture and hyperparameters on the thermodynamic, dynamical, and structural properties of the simulated CG systems, reporting and discussing challenges encountered and future directions envisioned.

Reference-guided DNA sequencing and alignment is an important process in computational molecular biology. The amount of DNA data grows very fast, and many new genomes are waiting to be sequenced while millions of private genomes need to be re-sequenced. Each human genome has 3.2 B base pairs, and each one could be stored with 2 bits of information, so one human genome would take 6.4 B bits or about 760 MB of storage (National Institute of General Medical Sciences). Today most powerful tensor processing units cannot handle the volume of DNA data necessitating a major leap in computing power. It is, therefore, important to investigate the usefulness of quantum computers in genomic data analysis, especially in DNA sequence alignment. Quantum computers are expected to be involved in DNA sequencing, initially as parts of classical systems, acting as quantum accelerators. The number of available qubits is increasing annually, and future quantum computers could conduct DNA sequencing, taking the place of classical computing systems. We present a novel quantum algorithm for reference-guided DNA sequence alignment modeled with gate-based quantum computing. The algorithm is scalable, can be integrated into existing classical DNA sequencing systems and is intentionally structured to limit computational errors. The quantum algorithm has been tested using the quantum processing units and simulators provided by IBM Quantum, and its correctness has been confirmed.

In an era increasingly focused on green computing and explainable AI, revisiting traditional approaches in theoretical and phenomenological particle physics is paramount. This project evaluates various machine learning (ML) algorithms-including Nearest Neighbors, Decision Trees, Random Forest, AdaBoost, Naive Bayes, Quadratic Discriminant Analysis (QDA), and XGBoost-alongside standard neural networks and a novel Physics-Informed Neural Network (PINN) for physics data analysis. We apply these techniques to a binary classification task that distinguishes the experimental viability of simulated scenarios based on Higgs observables and essential parameters. Through this comprehensive analysis, we aim to showcase the capabilities and computational efficiency of each model in binary classification tasks, thereby contributing to the ongoing discourse on integrating ML and Deep Neural Networks (DNNs) into physics research. In this study, XGBoost emerged as the preferred choice among the evaluated machine learning algorithms for its speed and effectiveness, especially in the initial stages of computation with limited datasets. However, while standard Neural Networks and Physics-Informed Neural Networks (PINNs) demonstrated superior performance in terms of accuracy and adherence to physical laws, they require more computational time. These findings underscore the trade-offs between computational efficiency and model sophistication.

Parameter-free theoretical predictions based on a dual shell mechanism within the proxy-SU(3) symmetry of atomic nuclei, as well as covariant density functional theory calculations using the DDME2 functional indicate that shape coexistence (SC) based on the particle-hole excitation mechanism cannot occur everywhere on the nuclear chart, but is restricted on islands lying within regions of 7-8, 17-20, 34-40, 59-70, 96-112, 146-168 protons or neutrons. Systematics of data for even-even nuclei possessing K=0 (beta) and K=2 (gamma) bands support the existence of these islands, on which shape coexistence appears whenever the K=0 bandhead 0_2^+ and the first excited state of the ground state band 2_1^+ lie close in energy, with nuclei characterized by 0_2^+ lying below the 2_1^+ found in the center of these islands. In addition a simple theoretical mechanism leading to multiple shape coexistence is briefly discussed.

This paper presents a novel hybrid approach to solving real-world drone routing problems by leveraging the capabilities of quantum computing. The proposed method, coined Quantum for Drone Routing (Q4DR), integrates the two most prominent paradigms in the field: quantum gate-based computing, through the Eclipse Qrisp programming language; and quantum annealers, by means of D-Wave System's devices. The algorithm is divided into two different phases: an initial clustering phase executed using a Quantum Approximate Optimization Algorithm (QAOA), and a routing phase employing quantum annealers. The efficacy of Q4DR is demonstrated through three use cases of increasing complexity, each incorporating real-world constraints such as asymmetric costs, forbidden paths, and itinerant charging points. This research contributes to the growing body of work in quantum optimization, showcasing the practical applications of quantum computing in logistics and route planning.

We investigate the large-N limit of the BMN matrix model with classical bosonic membranes which have spherical topologies and spin inside the 11-dimensional maximally supersymmetric plane-wave background. First we classify all possible M2-brane configurations based on the distribution of their components inside the SO(3)xSO(6) symmetric plane-wave spacetime. We then formulate a number of simple but very representative ansatze of dielectric tops that rotate in this space. We examine the leading-order radial and angular/multipole stability for a wide range of these configurations. By analyzing perturbations at the next-to-leading order, we find that they exhibit the phenomenon of turbulent cascading of instabilities. Thereby, long-wavelength perturbations generate higher-order multipole instabilities through their nonlinear couplings.

Airborne infection risk analysis is usually performed for enclosed spaces where susceptible individuals are exposed to infectious airborne respiratory droplets by inhalation. It is usually based on exponential, dose-response models of which a widely used variant is the Wells-Riley (WR) model. We revisit this infection-risk estimate and extend it to the population level. We use an epidemiological model where the mode of pathogen transmission, either airborne or contact, is explicitly considered. We illustrate the link between epidemiological models and the WR model. We argue that airborne infection quanta are, up to an overall density, airborne infectious respiratory droplets modified by a parameter that depends on biological properties of the pathogen, physical properties of the droplet, and behavioural parameters of the individual. We calculate the time-dependent risk to be infected during the epidemic for two scenarios. We show how the epidemic infection risk depends on the viral latent period and the event time, the time infection occurs. The infection risk follows the dynamics of the infected population. As the latency period decreases, infection risk increases. The longer a susceptible is present in the epidemic, the higher is its risk of infection by equal exposure time to the mode of transmission.

Prolate to oblate shape transitions have been predicted in an analytic way in the framework of the Interacting Boson Model (IBM), determining O(6) as the symmetry at the critical point. Parameter-independent predictions for prolate to oblate transitions in various regions on the nuclear chart have been made in the framework of the proxy-SU(3) and pseudo-SU(3) symmetries, corroborated by recent non-relativistic and relativistic mean field calculations along series of nuclear isotopes, with parameters fixed throughout, as well as by shell model calculations taking advantage of the quasi-SU(3) symmetry. Experimental evidence for regions of prolate to oblate shape transitions is in agreement with regions in which nuclei bearing the O(6) dynamical symmetry of the IBM have been identified, lying below major shell closures. In addition, gradual oblate to prolate transitions are seen when crossing major nuclear shell closures, in analogy to experimental observations in alkali clusters.

It has recently been argued that a Naive Bayesian classifier can be used to filter unsolicited bulk e-mail ("spam"). We conduct a thorough evaluation of this proposal on a corpus that we make publicly available, contributing towards standard benchmarks. At the same time we investigate the effect of attribute-set size, training-corpus size, lemmatization, and stop-lists on the filter's performance, issues that had not been previously explored. After introducing appropriate cost-sensitive evaluation measures, we reach the conclusion that additional safety nets are needed for the Naive Bayesian anti-spam filter to be viable in practice.

The proxy-SU(3) symmetry predicts, in a parameter-free way, based only on the Pauli principle and the short-range nature of the nucleon-nucleon interaction, non-vanishing values of the collective variable gamma almost everywhere across the nuclear chart. Substantial triaxiality with gamma between 15 and 45 degrees is proved to be expected along horizontal and vertical stripes on the nuclear chart, covering the nucleon numbers 22-26, 34-48, 74-80, 116-124, 172-182. Empirical support for these stripes is found by collecting all even-even nuclei for which the first two excited 2+ states are known, along with the B(E2)s connecting them, as well as the second 2+ state to the ground state. The stripes are related to regions in which oblate SU(3) irreducible representations appear, bearing similarity to the appearance of triaxiality within the SU(3)* dynamical symmetry of the interacting boson model-2. Detailed comparisons of the proxy-SU(3) predictions to the data and to predictions by state-of-the-art Monte Carlo shell model calculations for deformed N=94, 96, 98 isotones in the rare earth region show good overall agreement, with the exception of Z=70 and N=94, which correspond to fully symmetric proxy-SU(3) irreps, suggesting that the latter are an artifact of the method which can be amended by considering the influence of the neighboring irreps.

The technology of Quantum Computing (QC) is continuously evolving, as researchers explore new technologies and the public gains access to quantum computers with an increasing number of qubits. In addition, the research community and industry are increasingly interested in the potential use, application, and contribution of QCs to large-scale problems in the real world as a result of this technological enhancement. QCs operations are based on quantum mechanics, and their special properties are mainly exploited to solve computationally intensive problems in polynomial time, problems that are commonly unsolvable, even by High-Performance Computing systems (HPCs) in a feasible time. However, since QCs cannot perform as general-purpose computing machines, alternative computational approaches aiming to boost further their enormous computing abilities are requested, and their combination as an additional computing resource to HPC systems is considered as one of the most promising ones. In the proposed hybrid HPCs, the Quantum Processing Units (QPUs), similar to GPUs and CPUs, target specific problems that, through Quantum Algorithms, can exploit quantum properties like quantum entanglement and superposition to achieve substantial performance gains from the HPC point of view. This interconnection between classical HPC systems and QCs towards the creation of Hybrid Quantum-Classical computing systems is neither straightforward nor standardized while crucial for unlocking the real potential of QCs and achieving real performance improvements. The interconnection between the classical and quantum systems can be performed in the hardware, software (system), or application layer. In this study, a concise overview of the existing architectures for the interconnection interface between HPCs and QCs is provided, focusing on hardware approaches that enable effective hybrid quantum-classical operation.

Generative AI models offer powerful capabilities but often lack transparency, making it difficult to interpret their output. This is critical in cases involving artistic or copyrighted content. This work introduces a search-inspired approach to improve the interpretability of these models by analysing the influence of training data on their outputs. Our method provides observational interpretability by focusing on a model's output rather than on its internal state. We consider both raw data and latent-space embeddings when searching for the influence of data items in generated content. We evaluate our method by retraining models locally and by demonstrating the method's ability to uncover influential subsets in the training data. This work lays the groundwork for future extensions, including user-based evaluations with domain experts, which is expected to improve observational interpretability further.

Triaxial shapes in even-even nuclei have been considered since the early days of the nuclear collective model. Although many theoretical approaches have been used over the years for their description, no effort appears to have been made for grouping them together and identifying regions on the nuclear chart where the appearance of triaxiality might be favored. In addition, over the last few years, discussion has started on the appearance of small triaxiality in nuclei considered so far as purely axial rotors. In the present work we collect the predictions made by various theoretical approaches and show that pronounced triaxiality appears to be favored within specific stripes on the nuclear chart, with low triaxiality being present in the regions between these stripes, in agreement with parameter-free predictions made by the proxy-SU(3) approximation to the shell model, based on the Pauli principle and the short-range nature of the nucleon-nucleon interaction. The robustness of triaxiality within these stripes is supported by global calculations made in the framework of the Finite-Range Droplet Model (FRDM), which is based on completely different assumptions and possesses parameters fitted in order to reproduce fundamental nuclear properties.

Artificial Intelligence (AI) technology epitomizes the complex challenges posed by human-made artifacts, particularly those widely integrated into society and exert significant influence, highlighting potential benefits and their negative consequences. While other technologies may also pose substantial risks, AI's pervasive reach makes its societal effects especially profound. The complexity of AI systems, coupled with their remarkable capabilities, can lead to a reliance on technologies that operate beyond direct human oversight or understanding. To mitigate the risks that arise, several theoretical tools and guidelines have been developed, alongside efforts to create technological tools aimed at safeguarding Trustworthy AI. The guidelines take a more holistic view of the issue but fail to provide techniques for quantifying trustworthiness. Conversely, while technological tools are better at achieving such quantification, they lack a holistic perspective, focusing instead on specific aspects of Trustworthy AI. This paper aims to introduce an assessment method that combines the ethical components of Trustworthy AI with the algorithmic processes of PageRank and TrustRank. The goal is to establish an assessment framework that minimizes the subjectivity inherent in the self-assessment techniques prevalent in the field by introducing algorithmic criteria. The application of our approach indicates that a holistic assessment of an AI system's trustworthiness can be achieved by providing quantitative insights while considering the theoretical content of relevant guidelines.

The proxy-SU(3) symmetry was first presented in HINPw4 in Ioannina in May2017, justified within the Nilsson model and applied to parameter-free predictions of the collective variables beta and gamma in medium-mass and heavy nuclei. Major steps forward, including the connection of the proxy-SU(3) symmetry to the shell model, the justification of the dominance of highest weight states in terms of the short range nature of the nucleon-nucleon interaction, as well as the first proposal of appearance of islands of shape coexistence on the nuclear chart, have been presented in HINPw6 in Athens in May 2021. The recently hot topic of the prevalence of triaxial shapes in heavy nuclei will also be briefly outlined in the proxy-SU(3) framework.