Search

One goal of Artificial Intelligence is to learn meaningful representations for natural language expressions, but what this entails is not always clear. A variety of new linguistic behaviours present themselves embodied as computers, enhanced humans, and collectives with various kinds of integration and communication. But to measure and understand the behaviours generated by such systems, we must clarify the language we use to talk about them. Computational models are often confused with the phenomena they try to model and shallow metaphors are used as justifications for (or to hype) the success of computational techniques on many tasks related to natural language; thus implying their progress toward human-level machine intelligence without ever clarifying what that means. This paper discusses the challenges in the specification of "machines of meaning", machines capable of acquiring meaningful semantics from natural language in order to achieve their goals. We characterize "meaning" in a computational setting, while highlighting the need for detachment from anthropocentrism in the study of the behaviour of machines of meaning. The pressing need to analyse AI risks and ethics requires a proper measurement of its capabilities which cannot be productively studied and explained while using ambiguous language. We propose a view of "meaning" to facilitate the discourse around approaches such as neural language models and help broaden the research perspectives for technology that facilitates dialogues between humans and machines.

We report the crystal structures, magnetic and thermodynamic properties of two magnetically frustrated $A$'$A$"$M_2$F$_7$-type antiferromagnets, NaCdNi$_2$F$_7$ and NaCdMn$_2$F$_7$. While NaCdNi$_2$F$_7$ forms a stable pyrochlore structure (SG: $Fd \overline{3} m$, #227) with magnetic $S$ = 1 Ni$^{2+}$ ions on the frustrated pyrochlore 16$c$ site and fully disordered non-magnetic Na$^+$/Cd$^{2+}$ ions on the pyrochlore 16$d$ site, NaCdMn$_2$F$_7$ favors the defect-fluorite structure (SG: $Fm \overline{3} m$, #225) with magnetic $S$ = 5/2 Mn$^{2+}$ and non-magnetic Na$^+$ and Cd$^{2+}$ ions fully disordered on the fluorite 4$a$ site. This is a result of the Mn$^{2+}$ ionic radius being too close to the average Na$^+$/Cd$^{2+}$ ionic radius, hindering the cationic site ordering towards the stable pyrochlore structure. In both cases, dominant antiferromagnetic interactions $\theta_{CW,Ni}$ = -91.2(5) K and $\theta_{CW,Mn}$ = -42.4(4) K are noted, with no magnetic transition until $T_{f,Ni}$ = 3.2 K and $T_{f,Mn}$ = 2.0 K, implying substantial frustration, with frustration indices $f_{Ni}$ = 28.5 and $f_{Mn}$ = 21. AC susceptibility measurements and bifurcation of ZFC/FC low-field magnetization indicate a glassy ground-state, precipitated by the magnetic-bond-disorder stemming from the inherent structural disorder.

Accurate object detection and prediction are critical to ensure the safety and efficiency of self-driving architectures. Predicting object trajectories and occupancy enables autonomous vehicles to anticipate movements and make decisions with future information, increasing their adaptability and reducing the risk of accidents. Current State-Of-The-Art (SOTA) approaches often isolate the detection, tracking, and prediction stages, which can lead to significant prediction errors due to accumulated inaccuracies between stages. Recent advances have improved the feature representation of multi-camera perception systems through Bird's-Eye View (BEV) transformations, boosting the development of end-to-end systems capable of predicting environmental elements directly from vehicle sensor data. These systems, however, often suffer from high processing times and number of parameters, creating challenges for real-world deployment. To address these issues, this paper introduces a novel BEV instance prediction architecture based on a simplified paradigm that relies only on instance segmentation and flow prediction. The proposed system prioritizes speed, aiming at reduced parameter counts and inference times compared to existing SOTA architectures, thanks to the incorporation of an efficient transformer-based architecture. Furthermore, the implementation of the proposed architecture is optimized for performance improvements in PyTorch version 2.1. Code and trained models are available at https://github.com/miguelag99/Efficient-Instance-Prediction, Comment: The article has been presented in the 27th IEEE International Conference on Intelligent Transportation Systems (IEEE ITSC 2024) on September, 2024. Number of pages: 6, Number of figures: 4

We investigate perturbations in power-law monomial potentials within inflationary models driven by a single scalar field. By introducing a second term to the original potential, we study how this perturbation influences the slow-roll parameters and analyze the consequent changes in the spectral index, $n_s$, and the tensor-to-scalar ratio, $r$, treating the additional term as a correction to the monomial case. Comparing our numerical results with current cosmological data from the Planck satellite observations on $n_s$, $r$, and the clustering parameter, $\sigma_8$, we place significant constraints on the free parameter of our class of inflationary potentials. We found that the perturbative consistency method we analyze could be an interesting test to explore for more complex inflationary models, looking for features that better match the data that might highlight fundamental physics implications., Comment: 17 pages, 9 figures

Communication primitives play a central role in modern computing. They offer a panel of reliability and ordering guarantees for messages, enabling the implementation of complex distributed interactions. In particular, atomic broadcast is a pivotal abstraction for implementing fault-tolerant distributed services. This primitive allows disseminating messages across the system in a total order. There are two group communication primitives closely related to atomic broadcast. Atomic multicast permits targeting a subset of participants, possibly stricter than the whole system. Generic broadcast leverages the semantics of messages to order them only where necessary (that is when they conflict). In this paper, we propose to combine all these primitives into a single, more general one, called generic multicast. We formally specify the guarantees offered by generic multicast and present efficient algorithms. Compared to prior works, our solutions offer appealing properties in terms of time and space complexity. In particular, when a run is conflict-free, that is no two messages conflict, a message is delivered after at most three message delays., Comment: 12 pages, 2 figures

A cantilever beam under axial flow, confined or not, is known to develop self-sustained oscillations at sufficiently large flow velocities. In recent decades, the analysis of this archetypal system has been mostly pursued under linearized conditions, to calculate the critical boundaries separating stable from unstable behavior. However, nonlinear analysis of the self-sustained oscillations ensuing flutter instabilities are considerably rarer. Here we present a simplified one-dimensional nonlinear model describing a cantilever beam subjected to confined axial flow, for generic axial profiles of the fluid channels. In particular, we explore how the shape of the confinement walls affects the dynamics of the system. To simplify the problem, we consider symmetric channels with plane walls in either converging or diverging configurations. The beam is modeled in a modal framework, while bulk-flow equations, including singular head-loss terms, are used to model the flow-structure coupling forces. The dynamics of the system are first analyzed through linear stability analysis to assess the stabilizing/destabilizing effects of the channel walls configuration. Subsequently, we develop a systematic nonlinear analysis based on the continuation of periodic solutions. The harmonic balance method is used in conjunction with the asymptotic numerical method to calculate branches of periodic solutions. The continuation-based methods are used to investigate bifurcations with respect to both the reduced flow velocity and the channel slope parameter. From the results presented, we illustrate how continuationbased approaches and bifurcation analysis provide an efficient tool to analyze the nonlinear behavior of flow-induced vibration problems, particularly when reduced/simplified models are available., Comment: 22nd International Conference of Numerical Analysis and Applied Mathematics, Sep 2024, Heraklion, Greece

Reinforcement learning (RL) has seen significant research and application results but often requires large amounts of training data. This paper proposes two data-efficient off-policy RL methods that use parametrized Q-learning. In these methods, the Q-function is chosen to be linear in the parameters and quadratic in selected basis functions in the state and control deviations from a base policy. A cost penalizing the $\ell_1$-norm of Bellman errors is minimized. We propose two methods: Linear Matrix Inequality Q-Learning (LMI-QL) and its iterative variant (LMI-QLi), which solve the resulting episodic optimization problem through convex optimization. LMI-QL relies on a convex relaxation that yields a semidefinite programming (SDP) problem with linear matrix inequalities (LMIs). LMI-QLi entails solving sequential iterations of an SDP problem. Both methods combine convex optimization with direct Q-function learning, significantly improving learning speed. A numerical case study demonstrates their advantages over existing parametrized Q-learning methods., Comment: Accepted for Presentation at 63rd IEEE Conference on Decision and Control, CDC 2024, Milan, Italy, 2024

The Covid-19 pandemic has affected the world at multiple levels. Data sharing was pivotal for advancing research to understand the underlying causes and implement effective containment strategies. In response, many countries have promoted the availability of daily cases to support research initiatives, fostering collaboration between organisations and making such data available to the public through open data platforms. Despite the several advantages of data sharing, one of the major concerns before releasing health data is its impact on individuals' privacy. Such a sharing process should be based on state-of-the-art methods in Data Protection by Design and by Default. In this paper, we use a data set related to Covid-19 cases in the second largest hospital in Portugal to show how it is feasible to ensure data privacy while improving the quality and maintaining the utility of the data. Our goal is to demonstrate how knowledge exchange in multidisciplinary teams of healthcare practitioners, data privacy, and data science experts is crucial to co-developing strategies that ensure high utility of de-identified data., Comment: 7 figures and 8 tables

[Context] The adoption of micro-frontends architectures has gained traction as a promising approach to enhance modularity, scalability, and maintainability of web applications. [Goal] The primary aim of this research is to investigate the benefits and limitations of migrating a real-world application to a micro-frontends architecture from the perspective of the developers. [Method] Based on the action research approach, after diagnosis and planning, we applied an intervention of migrating the target web application to a micro-frontends architecture. Thereafter, the migration was evaluated in a workshop involving the remaining developers responsible for maintaining the application. During the workshop, these developers were presented with the migrated architecture, conducted a simple maintenance task, discussed benefits and limitations in a focus group to gather insights, and answered a questionnaire on the acceptance of the technology. [Results] Developers' perceptions gathered during the focus group reinforce the benefits and limitations reported in the literature. Key benefits included enhanced flexibility in technology choices, scalability of development teams, and gradual migration of technologies. However, the increased complexity of the architecture raised concerns among developers, particularly in dependency and environment management, debugging, and integration testing. [Conclusions] While micro-frontends represent a promising technology, unresolved issues still limit their broader applicability. Developers generally perceived the architecture as useful and moderately easy to use but hesitated to adopt it.

Traditional Byzantine Fault Tolerance (BFT) state machine replication protocols assume a partial synchrony model, leading to a design where a leader replica drives the protocol and is replaced after a timeout. Recently, we witnessed a surge of asynchronous BFT protocols, which use randomization to remove the need for bounds on message delivery times, making them more resilient to adverse network conditions. However, existing research proposals still fall short of gaining practical adoption, plausibly because they are not able to combine good performance with a simple design that can be readily understood and adopted. In this paper, we present Alea-BFT, a simple and highly efficient asynchronous BFT protocol, which is gaining practical adoption, namely in Ethereum distributed validators. Alea-BFT brings the key design insight from classical protocols of concentrating part of the work on a single designated replica and incorporates this principle in a simple two-stage pipelined design, with an efficient broadcast led by the designated replica, followed by an inexpensive binary agreement. The evaluation of our research prototype implementation and two real-world integrations in cryptocurrency ecosystems shows excellent performance, improving on the fastest protocol (Dumbo-NG) in terms of latency and displaying good performance under faults., Comment: arXiv admin note: substantial text overlap with arXiv:2202.02071

We present solutions corresponding to rotational configurations in the recently proposed Geometric Scalar Gravity (GSG) theory. The solutions obtained here have the important property that the associated closed time-like curves are always restricted to a compact domain of space-time. We compare this property with those of analogous solutions in General Relativity (GR).

We develop a semi-analytical model for transport in structured catalytic microreactors, where both reactant and product are compressible fluids. Making use of the lubrication and Fick-Jacobs approximations, we reduce the three-dimensional governing equations to an effective one-dimensional set of equations. Our model captures the effect of compressibility, of corrugations in the shape of the reactor, as well as of an inhomogeneous catalytic coating of the reactor walls. We show that in the weakly compressible limit (e.g., liquid-phase reactors), the distribution of catalyst does not influence the reactor yield, which we verify experimentally. Beyond this limit, we show that introducing inhomogeneities in the catalytic coating and corrugations to the reactor walls can improve the yield.

Sharing private data for learning tasks is pivotal for transparent and secure machine learning applications. Many privacy-preserving techniques have been proposed for this task aiming to transform the data while ensuring the privacy of individuals. Some of these techniques have been incorporated into tools, whereas others are accessed through various online platforms. However, such tools require manual configuration, which can be complex and time-consuming. Moreover, they require substantial expertise, potentially restricting their use to those with advanced technical knowledge. In this paper, we propose AUTOPRIV, the first automated privacy-preservation method, that eliminates the need for any manual configuration. AUTOPRIV employs meta-learning to automate the de-identification process, facilitating the secure release of data for machine learning tasks. The main goal is to anticipate the predictive performance and privacy risk of a large set of privacy configurations. We provide a ranked list of the most promising solutions, which are likely to achieve an optimal approximation within a new domain. AUTOPRIV is highly effective as it reduces computational complexity and energy consumption considerably., Comment: 12 pages, 6 figures, 3 tables

Adversarial examples, designed to trick Artificial Neural Networks (ANNs) into producing wrong outputs, highlight vulnerabilities in these models. Exploring these weaknesses is crucial for developing defenses, and so, we propose a method to assess the adversarial robustness of image-classifying ANNs. The t-distributed Stochastic Neighbor Embedding (t-SNE) technique is used for visual inspection, and a metric, which compares the clean and perturbed embeddings, helps pinpoint weak spots in the layers. Analyzing two ANNs on CIFAR-10, one designed by humans and another via NeuroEvolution, we found that differences between clean and perturbed representations emerge early on, in the feature extraction layers, affecting subsequent classification. The findings with our metric are supported by the visual analysis of the t-SNE maps.

Multiple synthetic data generation models have emerged, among which deep learning models have become the vanguard due to their ability to capture the underlying characteristics of the original data. However, the resemblance of the synthetic to the original data raises important questions on the protection of individuals' privacy. As synthetic data is perceived as a means to fully protect personal information, most current related work disregards the impact of re-identification risk. In particular, limited attention has been given to exploring outliers, despite their privacy relevance. In this work, we analyze the privacy of synthetic data w.r.t the outliers. Our main findings suggest that outliers re-identification via linkage attack is feasible and easily achieved. Furthermore, additional safeguards such as differential privacy can prevent re-identification, albeit at the expense of the data utility.

The main science aim of the BlackGEM array is to detect optical counterparts to gravitational wave mergers. Additionally, the array will perform a set of synoptic surveys to detect Local Universe transients and short time-scale variability in stars and binaries, as well as a six-filter all-sky survey down to ~22nd mag. The BlackGEM Phase-I array consists of three optical wide-field unit telescopes. Each unit uses an f/5.5 modified Dall-Kirkham (Harmer-Wynne) design with a triplet corrector lens, and a 65cm primary mirror, coupled with a 110Mpix CCD detector, that provides an instantaneous field-of-view of 2.7~square degrees, sampled at 0.564\arcsec/pixel. The total field-of-view for the array is 8.2 square degrees. Each telescope is equipped with a six-slot filter wheel containing an optimised Sloan set (BG-u, BG-g, BG-r, BG-i, BG-z) and a wider-band 440-720 nm (BG-q) filter. Each unit telescope is independent from the others. Cloud-based data processing is done in real time, and includes a transient-detection routine as well as a full-source optimal-photometry module. BlackGEM has been installed at the ESO La Silla observatory as of October 2019. After a prolonged COVID-19 hiatus, science operations started on April 1, 2023 and will run for five years. Aside from its core scientific program, BlackGEM will give rise to a multitude of additional science cases in multi-colour time-domain astronomy, to the benefit of a variety of topics in astrophysics, such as infant supernovae, luminous red novae, asteroseismology of post-main-sequence objects, (ultracompact) binary stars, and the relation between gravitational wave counterparts and other classes of transients, Comment: Published in PASP

We prove that local stable/unstable sets of homeomorphisms of an infinite compact metric space satisfying the gluing-orbit property always contain compact and perfect subsets of the space. As a consequence, we prove that if a positively countably expansive homeomorphism satisfies the gluing-orbit property, then the space is a single periodic orbit. We also prove that there are homeomorphisms with gluing-orbit such that its induced homeomorphism on the hyperspace of compact subsets does not have gluing-orbit, contrasting with the case of the shadowing and specification properties, proving that if the induced map has gluing-orbit, then the base map has gluing-orbit and is topologically mixing., Comment: 13 pages, 1 figure

Event-triggered control is often argued to lower the average triggering rate compared to time-triggered control while still achieving a desired control goal, e.g., the same performance level. However, this property, often called consistency, cannot be taken for granted and can be hard to analyze in many settings. In particular, although numerous decentralized event-triggered control schemes have been proposed in the past years, their performance properties with respect to time-triggered control remain mostly unexplored. In this paper, we therefore examine the performance properties of event-triggered control (relative to time-triggered control) for a single-integrator consensus problem with a level-triggering rule. We consider the long-term average quadratic deviation from consensus as a performance measure. For this setting, we show that enriching the information the local controllers use improves the performance of the consensus algorithm but renders a previously consistent event-triggered control scheme inconsistent. In addition, we do so while deploying optimal control inputs which we derive for both information cases and all triggering schemes. With this insight, we can furthermore explain the relationship between two contrasting consistency results from the literature on decentralized event-triggered control. We support our theoretical findings with simulation results.