Search

Convolutional Neural Networks (CNNs) have seen significant performance improvements in recent years. However, due to their size and complexity, they function as black-boxes, leading to transparency concerns. State-of-the-art saliency methods generate local explanations that highlight the area in the input image where a class is identified but cannot explain how a concept of interest contributes to the prediction, which is essential for bias mitigation. On the other hand, concept-based methods, such as TCAV (Testing with Concept Activation Vectors), provide insights into how sensitive is the network to a concept, but cannot compute its attribution in a specific prediction nor show its location within the input image. This paper introduces a novel post-hoc explainability framework, Visual-TCAV, which aims to bridge the gap between these methods by providing both local and global explanations for CNN-based image classification. Visual-TCAV uses Concept Activation Vectors (CAVs) to generate saliency maps that show where concepts are recognized by the network. Moreover, it can estimate the attribution of these concepts to the output of any class using a generalization of Integrated Gradients. This framework is evaluated on popular CNN architectures, with its validity further confirmed via experiments where ground truth for explanations is known, and a comparison with TCAV. Our code will be made available soon., Comment: Preprint currently under review

Chaos characterized by its irregularity and high sensitivity to initial conditions finds various applications in secure optical communications, random number generations, light detection and ranging systems, etc. Semiconductor lasers serve as ideal light platforms for chaos generations owing to the advantages in on-chip integration and complex nonlinear effects. In near-infrared wavelengths, semiconductor laser based chaotic light sources have been extensively studied and experimentally demonstrated. However, in the terahertz (THz) spectral range, due to the lack of effective THz light sources and high-speed detectors, chaos generation in THz semiconductor lasers, e.g., quantum cascade lasers (QCLs), is particularly challenging. Due to the fast intersubband carrier transitions, single mode THz QCLs resemble Class A lasers, where chaos can be hardly excited, even with external perturbations. In this work, we experimentally show a THz chaos source based on a sole multimode THz QCL without any external perturbations. Such a dynamical regime is characterized by the largest Lyapunov exponent associated to the temporal traces of the measured radio frequency (intermode beatnote) signal of the laser. The experimental results and chaos validation are confirmed by simulations of our model based on effective semiconductor Maxwell-Bloch Equations. To further understand the physical mechanism of the chaos generation in THz QCLs, a reduced model based on two coupled complex Ginzburg-Landau equations is derived from the full model cited above to systematically investigate the effects of the linewidth enhancement factor and group velocity dispersion on the chaotic regime. This model allows us to show that the chaos generation in the THz QCL can be ascribed to the system attaining the defect mediated turbulence regime., Comment: 30 pages, 6 figures

In this paper we study the birational geometry of $X$, a projective space $\mathbb{P}^n$ blown up at $s$ general points. We obtain a characterization of a special class of subvarieties, which we call Weyl $r$-planes, each of them being swept out by one $(n-r)$-moving curve class. For Mori dream spaces $X$, we unify different previous approaches, by reinterpreting the description of the Mori chamber decomposition of the effective cone in terms of these curve classes; in this case all relevant orbits are finite. For general $s$, we prove that Weyl $r$-planes are infinite and continue to appear as stable base locus. This yields the definition of the Weyl chamber decomposition of the pseudoeffective cone of divisors, that is a coarsening of its Mori chamber decomposition, when the latter can be defined. We prove that the combinatorics of such chamber decomposition obeys the Weyl group action on curves. We conjecture that for $s=n+4$ the Weyl and the Mori chamber decompositions coincide., Comment: 35 pages

This work provides a mathematical derivation of a quasi-stationary model for multimode parametric down-conversion, which was presented in [Gatti et al., Sci. Rep. 13, 16786] with heuristic arguments. The model is here derived from the 3D+1 propagation equation of the quantum fields in a nonlinear crystal, and its approximations discussed thoroughly. Thanks to its relative simplicity, and to the fact that it is valid in any gain regime, both at a quantum and classical level, it allows a unified description of disparate experimental observations conducted over the last 20 years, often described in the past by means of limited ad hoc models., Comment: Accepted for publication on Phyl.Trans. of The Royal Soc. A, will appear with DOI 10.1098/rsta.2023.0334. Main ms: 14 pages, 6 figures. Supplementary: 9 pages, 6 figures

Coupling is an essential mechanism that drives complexity in natural systems, transforming single, non-interacting elements into intricate networks with rich physical properties. Here, we demonstrate a chip-scale coupled laser system that exhibits complex optical states impossible to achieve in an uncoupled system. We show that a pair of coupled semiconductor ring lasers spontaneously forms a frequency comb consisting of the hybridized modes of its coupled cavity, exhibiting a large number of phase-locked tones that anticross with one another. Experimental coherent waveform reconstruction reveals that the hybridized frequency comb manifests itself as pairs of bright and dark picosecond-long solitons circulating simultaneously. The dark and bright solitons exit the coupled cavity at the same time, leading to breathing bright solitons temporally overlapped with their dark soliton counterparts - a state inaccessible for a single, free-running laser. Our results demonstrate that the rules that govern allowable states of light can be broken by simply coupling elements together, paving the way for the design of more complex networks of coupled on-chip lasers., Comment: 13 pages, 3 figures

The discovery of XYZ exotic states in the hadronic sector with two heavy quarks, represents a significant challenge in particle theory. Understanding and predicting their nature remains an open problem. In this work, we demonstrate how the Born$-$Oppenheimer (BO) effective field theory (BOEFT), derived from Quantum Chromodynamics (QCD) on the basis of scale separation and symmetries, can address XYZ exotics of any composition. We derive the Schr\"odinger coupled equations that describe hybrids, tetraquarks, pentaquarks, doubly heavy baryons, and quarkonia at leading order, incorporating nonadiabatic terms, and present the predicted multiples. We define the static potentials in terms of the QCD static energies for all relevant cases. We provide the precise form of the nonperturbative low-energy gauge-invariant correlators required for the BOEFT: static energies, generalized Wilson loops, gluelumps, and adjoint mesons. These are to be calculated on the lattice and we calculate here their short-distance behavior. Furthermore, we outline how spin-dependent corrections and mixing terms can be incorporated using matching computations. Lastly, we discuss how static energies with the same BO quantum numbers mix at large distances leading to the phenomenon of avoided level crossing. This effect is crucial to understand the emergence of exotics with molecular characteristics, such as the $\chi_{c1}(3872)$. With BOEFT both the tetraquark and the molecular picture appear as part of the same description., Comment: 115 pages, 13 figures, 13 tables

Aerospace manufacturing companies, such as Thales Alenia Space, design, develop, integrate, verify, and validate products characterized by high complexity and low volume. They carefully document all phases for each product but analyses across products are challenging due to the heterogeneity and unstructured nature of the data in documents. In this paper, we propose a hybrid methodology that leverages Knowledge Graphs (KGs) in conjunction with Large Language Models (LLMs) to extract and validate data contained in these documents. We consider a case study focused on test data related to electronic boards for satellites. To do so, we extend the Semantic Sensor Network ontology. We store the metadata of the reports in a KG, while the actual test results are stored in parquet accessible via a Virtual Knowledge Graph. The validation process is managed using an LLM-based approach. We also conduct a benchmarking study to evaluate the performance of state-of-the-art LLMs in executing this task. Finally, we analyze the costs and benefits of automating preexisting processes of manual data extraction and validation for subsequent cross-report analyses., Comment: Paper Accepted at ISWC 2024 In-Use Track

The Terracini locus $\mathbb{T}(n, d; x)$ is the locus of all finite subsets $S \subset \mathbb{P}^n$ of cardinality $x$ such that $\langle S \rangle = \mathbb{P}^n$, $h^0(\mathcal{I}_{2S}(d)) > 0$, and $h^1(\mathcal{I}_{2S}(d)) > 0$. The celebrated Alexander-Hirschowitz Theorem classifies the triples $(n,d,x)$ for which $\dim\mathbb{T}(n, d; x)=xn$. Here we fully characterize the next step in the case $n=2$, namely, we prove that $\mathbb{T}(2,d;x)$ has at least one irreducible component of dimension $2x-1$ if and only if either $(d,x)=(6,10)$ or $(d,x)=(4,4)$ or $d\equiv 1,2 \pmod{3}$ and $x=(d+2)(d+1)/6$., Comment: 9 pages, comments are welcome

We prove that Segre-Veronese varieties are never secant defective if each degree is at least three. The proof is by induction on the number of factors, degree and dimension. As a corollary, we give an almost optimal non-defectivity result for Segre-Veronese varieties with one degree equal to one and all the others at least three., Comment: 12 pages

Despite significant advances in the field of phononic crystals, the development of acoustic metafluids that replicate the behaviour of liquids in three dimensions remains elusive. For instance, water - the quintessential pentamode (PM) material - has a bulk modulus two orders of magnitude higher than current state-of-the-art PMs. The need for a low shear modulus inherently conflicts with the desire of high bulk modulus and density. In this letter, we shed light on the limitations of existing PM geometries and propose an innovative shape for the links that constitute the network. Inspired by the kinematics of ropes, these links are constructed from thin fibres and demonstrate the potential to create PMs with properties akin to those of liquids. As a prime example, we present the design of the first metamaterial that fully deserves the name 3D metal water, since its acoustic properties in the low frequency regime are indistinguishable with water. Additionally, we highlight a shear band gap in the lattice dispersion diagram, and illustrate the influence of geometric parameters on the dynamic properties at higher frequencies. This novel design of metafluids holds promise for applications requiring anisotropic materials such as acoustic lenses, waveguides, and cloaks., Comment: 13 pages, 6 figures

Let $\overline{\mathrm{Mov}}^k(X)$ be the closure of the cone $\mathrm{Mov}^k(X)$ generated by classes of effective divisors on a projective variety $X$ with stable base locus of codimension at least $k+1$. We propose a generalized version of the Log Nonvanishing Conjecture and of the Log Abundance Conjecture for a klt pair $(X,\Delta)$, that is: if $K_X+\Delta \in \overline{\mathrm{Mov}}^{k}(X)$, then $K_X+\Delta \in \mathrm{Mov}^{k}(X)$. Moreover, we prove that if the Log Minimal Model Program, the Log Nonvanishing, and the Log Abundance hold, then so does our conjecture., Comment: 4 pages, comments welcome

Transparency and explainability in image classification are essential for establishing trust in machine learning models and detecting biases and errors. State-of-the-art explainability methods generate saliency maps to show where a specific class is identified, without providing a detailed explanation of the model's decision process. Striving to address such a need, we introduce a post-hoc method that explains the entire feature extraction process of a Convolutional Neural Network. These explanations include a layer-wise representation of the features the model extracts from the input. Such features are represented as saliency maps generated by clustering and merging similar feature maps, to which we associate a weight derived by generalizing Grad-CAM for the proposed methodology. To further enhance these explanations, we include a set of textual labels collected through a gamified crowdsourcing activity and processed using NLP techniques and Sentence-BERT. Finally, we show an approach to generate global explanations by aggregating labels across multiple images., Comment: International Joint Conference on Artificial Intelligence 2024 (to be published)

We compute the suppression of bottomonium in the quark-gluon plasma using the three-loop QCD static potential. The potential describes the spin-averaged bottomonium spectrum below threshold with a less than 1% error. Within potential nonrelativistic quantum chromodynamics and an open quantum systems framework, we compute the evolution of the bottomonium density matrix. The values of the quarkonium transport coefficients are obtained from lattice QCD measurements of the bottomonium in-medium width and thermal mass shift; we additionally include for the first time a vacuum contribution to the dispersive coefficient ${\gamma}$. Using the three-loop potential and the values of the heavy quarkonium transport coefficients, we find that the resulting bottomonium nuclear modification factor is consistent with experimental observations, while at the same time reproducing the lattice measurements of the in-medium width., Comment: 15 pages, 3 figures, 1 table

We present a unified model to describe the dynamics of optical frequency combs (OFCs) in quantum cascade lasers (QCLs), incorporating both ring and Fabry-P\'erot (FP) cavity configurations. The model derives a modified complex Ginzburg-Landau equation (CGLE), leveraging an order parameter approach and is capable of capturing the dynamics of both configurations, thus enabling a comparative analysis. In the modified CGLE, a nonlinear integral term appears which is associated with the coupling between counterpropagating fields in the FP cavity and whose suppression yields the ring model, which is known to be properly described by a conventional CGLE. We show that this crucial term holds a key role in inhibiting the formation of harmonic frequency combs (HFCs), associated to multi-peaked localized structures, due to its anti-patterning effect. We provide in support a comprehensive campaign of numerical simulations, in which we observe a higher occurrence of HFCs in the ring configuration compared to the FP case. Furthermore, the simulations demonstrate the model's capability to reproduce experimental observations, including the coexistence of amplitude and frequency modulation, linear chirp, and typical dynamic scenarios observed in QCLs. Finally, we perform a linear stability analysis of the single-mode solution for the ring case, confirming its consistency with numerical simulations and highlighting its predictive power regarding the formation of harmonic combs.

Accurate positioning is known to be a fundamental requirement for the deployment of Connected Automated Vehicles (CAVs). To meet this need, a new emerging trend is represented by cooperative methods where vehicles fuse information coming from navigation and imaging sensors via Vehicle-to-Everything (V2X) communications for joint positioning and environmental perception. In line with this trend, this paper proposes a novel data-driven cooperative sensing framework, termed Cooperative LiDAR Sensing with Message Passing Neural Network (CLS-MPNN), where spatially-distributed vehicles collaborate in perceiving the environment via LiDAR sensors. Vehicles process their LiDAR point clouds using a Deep Neural Network (DNN), namely a 3D object detector, to identify and localize possible static objects present in the driving environment. Data are then aggregated by a centralized infrastructure that performs Data Association (DA) using a Message Passing Neural Network (MPNN) and runs the Implicit Cooperative Positioning (ICP) algorithm. The proposed approach is evaluated using two realistic driving scenarios generated by a high-fidelity automated driving simulator. The results show that CLS-MPNN outperforms a conventional non-cooperative localization algorithm based on Global Navigation Satellite System (GNSS) and a state-of-the-art cooperative Simultaneous Localization and Mapping (SLAM) method while approaching the performances of an oracle system with ideal sensing and perfect association., Comment: This work has been submitted to the IEEE Transactions on Signal Processing for possible publication

The fabrication and integration of high-quality structures of Yttrium Iron Garnet (YIG) is critical for magnonics.Films with excellent properties are obtained only on single crystal Gadolinium Gallium Garnet (GGG) substrates using high-temperature processes. The subsequent realization of magnonic structures via lithography and etching is not straightforward as it requires a tight control of the edge roughness, to avoid magnon scattering, and planarization in case of multilayer devices. In this work we describe a different approach based on local laser annealing of amorphous YIG films, avoiding the need for subjecting the entire sample to high thermal budgets and for physical etching. Starting from amorphous and paramagnetic YIG films grown by pulsed laser deposition at room temperature on GGG, a 405 nm laser is used for patterning arbitrary shaped ferrimagnetic structures by local crystallization. In thick films (160 nm) the laser induced surface corrugation prevents the propagation of spin-wave modes in patterned conduits. For thinner films (80 nm) coherent propagation is observed in 1.2 micron wide conduits displaying an attenuation length of 5 micron which is compatible with a damping coefficient of about 5e-3. Possible routes to achieve damping coefficients compatible with state-of-the art epitaxial YIG films are discussed., Comment: 29 pages, 6 figures, This is the pre-peer reviewed version of the following article: "Patterning magnonic structures via laser induced crystallization of Yittrium Iron Garnet", by A. Del Giacco et al. https://doi.org/10.1002/adfm.202401129, which has been accepted for publication in Advanced Functional Materials

For a quantitative investigation on the time evolution of heavy thermal dark matter at and after thermal freeze-out, near-threshold processes need to be taken into account which have a large impact on the observed dark matter relic abundance. In this conference paper, we study the recoil effect of heavy dark matter pairs in a thermal bath and compute the annihilation cross section and the decay width as well as the bound-state formation cross section of dark matter fermion-antifermion pairs in the laboratory frame within the framework of potential non-relativistic effective field theories at finite temperature. For the considered hierarchy of energy scales, we highlight the effect of the recoil corrections to the thermal rates., Comment: 7 pages, 1 figure; Talk given at EPS-HEP2023 and TAUP2023

For non-relativistic thermal dark matter, close-to-threshold effects largely dominate the evolution of the number density for most of the times after thermal freeze-out, and hence affect the cosmological relic density. A precise evaluation of the relevant interaction rates in a thermal medium representing the early universe includes accounting for the relative motion of the dark matter particles and the thermal medium. We consider a model of dark fermions interacting with a plasma of dark gauge bosons, which is equivalent to thermal QED. The temperature is taken to be smaller than the dark fermion mass and the inverse of the typical size of the dark fermion-antifermion bound states, which allows for the use of non-relativistic effective field theories. For the annihilation cross section, bound-state formation cross section, bound-state dissociation width and bound-state transition width of dark matter fermion-antifermion pairs, we compute the leading recoil effects in the reference frame of both the plasma and the center-of-mass of the fermion-antifermion pair. We explicitly verify the Lorentz transformations among these quantities. We evaluate the impact of the recoil corrections on the dark matter energy density. Our results can be directly applied to account for the relative motion of quarkonia in the quark-gluon plasma formed in heavy-ion collisions. They may be also used to precisely assess thermal effects in atomic clocks based on atomic transitions; the present work provides a first field theory derivation of time dilation for these processes in vacuum and in a medium., Comment: 52 pages, 9 figures, published version, improved discussion of the effective Lagrangian

This document summarizes the efforts of the EMMI Rapid Reaction Task Force on "Suppression and (re)generation of quarkonium in heavy-ion collisions at the LHC", centered around their 2019 and 2022 meetings. It provides a review of existing experimental results and theoretical approaches, including lattice QCD calculations and semiclassical and quantum approaches for the dynamical evolution of quarkonia in the quark-gluon plasma as probed in high-energy heavy-ion collisions. The key ingredients of the transport models are itemized to facilitate comparisons of calculated quantities such as reaction rates, binding energies, and nuclear modification factors. A diagnostic assessment of the various results is attempted and coupled with an outlook for the future., Comment: 54 pages, 19 figures

Despite the ongoing progress in integrated optical frequency comb technology, compact sources of short bright pulses in the mid-infrared wavelength range from 3 {\mu}m to 12 {\mu}m so far remained beyond reach. The state-of-the-art ultrafast pulse emitters in the mid-infrared are complex, bulky, and inefficient systems based on the downconversion of near-infrared or visible pulsed laser sources. Here we show a purely DC-driven semiconductor laser chip that generates one picosecond solitons at the center wavelength of 8.3 {\mu}m at GHz repetition rates. The soliton generation scheme is akin to that of passive nonlinear Kerr resonators. It relies on a fast bistability in active nonlinear laser resonators, unlike traditional passive mode-locking which relies on saturable absorbers or active mode-locking by gain modulation in semiconductor lasers. Monolithic integration of all components - drive laser, active ring resonator, coupler, and pump filter - enables turnkey generation of bright solitons that remain robust for hours of continuous operation without active stabilization. Such devices can be readily produced at industrial laser foundries using standard fabrication protocols. Our work unifies the physics of active and passive microresonator frequency combs, while simultaneously establishing a technology for nonlinear integrated photonics in the mid-infrared.

The static QCD force from the lattice can be used to extract $\Lambda_{\overline{\textrm{MS}}}$, which determines the running of the strong coupling. Usually, this is done with a numerical derivative of the static potential. However, this introduces additional systematic uncertainties; thus, we use another observable to measure the static force directly. This observable consists of a Wilson loop with a chromoelectric field insertion. We work in the pure SU(3) gauge theory. We use gradient flow to improve the signal-to-noise ratio and to address the field insertion. We extract $\Lambda_{\overline{\textrm{MS}}}^{n_f=0}$ from the data by exploring different methods to perform the zero-flow-time limit. We obtain the value $\sqrt{8t_0} \Lambda_{\overline{\textrm{MS}}}^{n_f=0} =0.629^{+22}_{-26}$, where $t_0$ is a flow-time reference scale. We also obtain precise determinations of several scales: $r_0/r_1$, $\sqrt{8 t_0}/r_0$, $\sqrt{8 t_0}/r_1$ and we compare these to the literature. The gradient flow appears to be a promising method for calculations of Wilson loops with chromoelectric and chromomagnetic insertions in quenched and unquenched configurations, Comment: 27 pages, 19 figures The latest version matches the reviewed and accepted version

Off-lightcone Wilson-line operators are constructed using local operators connected by time-like or space-like Wilson lines, which ensure gauge invariance. Off-lightcone Wilson-line operators have broad applications in various contexts. For instance, space-like Wilson-line operators play a crucial role in determining quasi-distribution functions (quasi-PDFs), while time-like Wilson-line operators are essential for understanding quarkonium decay and production within the potential non-relativistic QCD (pNRQCD) framework. In this work, we establish a systematic approach for calculating the matching from the gradient-flow scheme to the $\overline{\rm MS}$ scheme in the limit of small flow time for off-lightcone Wilson-line operators. By employing the one-dimensional auxiliary-field formalism, we simplify the matching procedure, reducing it to the matching of local current operators. We provide one-loop level matching coefficients for these local current operators. For the case of hadronic matrix element related to the quark quasi-PDFs, we show at one-loop level that the finite flow time effect is very small as long as the flow radius is smaller than the physical distance $z$, which is usually satisfied in lattice gradient flow computations. Applications include lattice gradient flow computations of quark/gluon quasi-PDFs, gluonic correlators related to quarkonium decay and production in pNRQCD, and spin-dependent potentials in terms of chromoelectric and chromomagnetic field insertions into a Wilson loop., Comment: Published version, including updating footnote 5 and Note Added for comparison with existing calculation, and correcting several text typos

The widespread adoption of the fifth generation (5G) of cellular networks has brought new opportunities for the development of localization-based services. High-accuracy positioning use cases and functionalities defined by the standards are drawing the interest of vertical industries. In the transition towards the deployment, this paper aims to provide an in-depth tutorial on 5G positioning, summarizing the evolutionary path that led to the standardization of cellular-based positioning, describing the localization elements in current and forthcoming releases of the Third Generation Partnership Project (3GPP) standard, and the major research trends. By providing fundamental notions on wireless localization, comprehensive definitions of measurements and architectures, examples of algorithms, and details on simulation approaches, this paper is intended to represent an exhaustive guide for researchers and practitioners. Our approach aims to merge practical aspects of enabled use cases and related requirements with theoretical methodologies and fundamental bounds, allowing to understand the trade-off between system complexity and achievable, i.e., tangible, benefits of 5G positioning services. We analyze the performance of 3GPP Rel-16 positioning by standard-compliant simulations in realistic outdoor and indoor propagation environments, investigating the impact of the system configuration and the limitations to be resolved for delivering accurate positioning solutions., Comment: This work has been accepted in IEEE Communications Surveys & Tutorials

During the Covid-19 pandemic, research communities focused on collecting and understanding people's behaviours and feelings to study and tackle the pandemic indirect effects. Despite its consequences are slowly starting to fade away, such an interest is still alive. In this article, we propose a hybrid, gamified, story-driven data collection approach to spark self-empathy, hence resurfacing people's past feelings. The game is designed to include a physical board, decks of cards, and a digital application. As the player plays through the game, they customize and escape from their lockdown room by completing statements and answering a series of questions that define their story. The decoration of the lockdown room and the storytelling-driven approach are targeted at sparking people's emotions and self-empathy towards their past selves. Ultimately, the proposed approach was proven effective in sparking and collecting feelings, while a few improvements are still necessary.