Search

In this work, we propose a large-graph limit estimate of the matching coverage for several matching algorithms, on general graphs generated by the configuration model. For a wide class of {\em local} matching algorithms, namely, algorithms that only use information on the immediate neighborhood of the explored nodes, we propose a joint construction of the graph by the configuration model, and of the resulting matching on the latter graph. This leads to a generalization in infinite dimension of the differential equation method of Wormald: We keep track of the matching algorithm over time by a measure-valued CTMC, for which we prove the convergence, to the large-graph limit, to a deterministic hydrodynamic limit, identified as the unique solution of a system of ODE's in the space of integer measures. Then, the asymptotic proportion of nodes covered by the matching appears as a simple function of that solution. We then make this solution explicit for three particular local algorithms: the classical {\sc greedy} algorithm, and then the {\sc uni-min} and {\sc uni-max} algorithms, two variants of the greedy algorithm that select, as neighbor of any explored node, its neighbor having the least (respectively largest) residual degree.

A four dimensional pseudo-Riemannian manifold of signature (2, 2) is called a Walker manifold if it admits a parallel degenerate plane field. In this paper, we study the curvature properties of such a class of four dimensional Walker manifolds. In particular, we characterize Walker metrics of a given form which are Einstein-like.

Despite their success and widespread adoption, the opaque nature of deep neural networks (DNNs) continues to hinder trust, especially in critical applications. Current interpretability solutions often yield inconsistent or oversimplified explanations, or require model changes that compromise performance. In this work, we introduce TRACER, a novel method grounded in causal inference theory designed to estimate the causal dynamics underpinning DNN decisions without altering their architecture or compromising their performance. Our approach systematically intervenes on input features to observe how specific changes propagate through the network, affecting internal activations and final outputs. Based on this analysis, we determine the importance of individual features, and construct a high-level causal map by grouping functionally similar layers into cohesive causal nodes, providing a structured and interpretable view of how different parts of the network influence the decisions. TRACER further enhances explainability by generating counterfactuals that reveal possible model biases and offer contrastive explanations for misclassifications. Through comprehensive evaluations across diverse datasets, we demonstrate TRACER's effectiveness over existing methods and show its potential for creating highly compressed yet accurate models, illustrating its dual versatility in both understanding and optimizing DNNs.

Asymmetries in deuterium-tritium (D-T) particle transport and divertor pumping speeds are shown to enhance tritium self-sufficiency in fusion power plants. Using a diffusive particle transport model that links the plasma core, separatrix, and divertor regions, it is shown that reducing tritium transport while increasing deuterium transport improves both tritium burn efficiency and overall fusion power. By selectively increasing deuterium transport, tritium burn efficiency can be further optimized, assuming the availability of asymmetric D-T fueling and advanced divertor technologies. These asymmetries become especially beneficial at high tritium burn efficiencies. In one example, by increasing the D-T particle diffusivity ratio and decreasing the D-T divertor pump speed ratio, each by a factor of five, the tritium burn efficiency increases eleven-fold from 0.026 to 0.29 at fixed fusion power. We propose a novel approach to achieve asymmetric D-T pumping with a divertor re-injection system utilizing a partial ionization plasma centrifuge. In an ARC-class power plant, this approach could yield an order-of-magnitude improvement in tritium burn efficiency. These findings motivate the development of techniques and technologies to reduce core tritium transport and increase tritium divertor pumping speeds., Comment: 26 pages, 18 figures

Although people spend most of their time indoors, outdoor tracking systems, such as the Global Positioning System (GPS), are predominantly used for location-based services. These systems are accurate outdoors, easy to use, and operate autonomously on each mobile device. In contrast, Indoor Tracking Systems~(ITS) lack standardization and are often difficult to operate because they require costly infrastructure. In this paper, we propose an indoor tracking algorithm that uses collected data from inertial sensors embedded in most mobile devices. In this setting, mobile devices autonomously estimate their location, hence removing the burden of deploying and maintaining complex and scattered hardware infrastructure. In addition, these devices collaborate by anonymously exchanging data with other nearby devices, using wireless communication, such as Bluetooth, to correct errors in their location estimates. Our collaborative algorithm relies on low-complexity geometry operations and can be deployed on any recent mobile device with commercial-grade sensors. We evaluate our solution on real-life data collected by different devices. Experimentation with 16 simultaneously moving and collaborating devices shows an average accuracy improvement of 44% compared to the standalone Pedestrian Dead Reckoning algorithm., Comment: ACCEPTED FOR PUBLICATION AND PRESENTED IN EAI MOBIQUITOUS 2023

The design and analysis of high energy density (HED) laser experiments typically rely on radiation hydrodynamics simulations. However, some laser-plasma interaction regimes are not collisional and cannot be adequately modeled with hydrodynamics. For example, strongly driven magnetic reconnection and magnetized collisionless shock experiments possess extended hydrodynamic or even kinetic properties, necessitating first-principles kinetic simulations. In this paper, we present the benchmarking and first results obtained with a laser-ray-tracing and inverse Bremsstrahlung absorption module implemented in the particle-in-cell code PSC. The simulation results are compared to radiation hydrodynamic simulations using the FLASH code as well as analytical estimates. We successfully benchmark the energy deposition model and overall hydrodynamic evolution of the systems. We also consider possible kinetic effects that may be expected from laser-target ablation in the HED regime, including non-local transport and two-temperature effects., Comment: 17 pages, 12 figures

Federated Learning (FL) has gained attention across various industries for its capability to train machine learning models without centralizing sensitive data. While this approach offers significant benefits such as privacy preservation and decreased communication overhead, it presents several challenges, including deployment complexity and interoperability issues, particularly in heterogeneous scenarios or resource-constrained environments. Over-the-air (OTA) FL was introduced to tackle these challenges by disseminating model updates without necessitating direct device-to-device connections or centralized servers. However, OTA-FL brought forth limitations associated with heightened energy consumption and network latency. In this paper, we propose a multi-attribute client selection framework employing the grey wolf optimizer (GWO) to strategically control the number of participants in each round and optimize the OTA-FL process while considering accuracy, energy, delay, reliability, and fairness constraints of participating devices. We evaluate the performance of our multi-attribute client selection approach in terms of model loss minimization, convergence time reduction, and energy efficiency. In our experimental evaluation, we assessed and compared the performance of our approach against the existing state-of-the-art methods. Our results demonstrate that the proposed GWO-based client selection outperforms these baselines across various metrics. Specifically, our approach achieves a notable reduction in model loss, accelerates convergence time, and enhances energy efficiency while maintaining high fairness and reliability indicators.

We propose an alternative to the Bethe Ansatz method for strongly-interacting fermionic (or bosonic) mixtures on a ring. Starting from the knowledge of the solution for single-component non-interacting fermions (or strongly-interacting bosons), we explicitly impose periodic condition on the amplitudes of the spin configurations. This reduces drastically the number of independent complex amplitudes that we determine by constrained diagonalization of an effective Hamiltonian. This procedure allows us to obtain a complete basis for the exact low-energy many-body solutions for mixtures with a large number of particles, both for $SU(\kappa)$ and symmetry-breaking systems., Comment: 29 pages and 4 figures. Submission to SciPost

We describe the design of a Thomson scattering (TS) diagnostic to be used on the SMall Aspect Ratio Tokamak (SMART). SMART is a spherical tokamak being commissioned in Spain that aims to explore positive triangularity (PT) and negative triangularity (NT) plasma scenarios at a low aspect ratio. The SMART TS diagnostic is designed to enable a wide range of electron temperature (1 eV to 1 keV) and density (0.5 - 10 $\times 10^{19} ~\rm{m^{ - 3}}$) measurements. A 2Joule laser operating at 1064 nm will be used to probe the electron temperature and density of the plasma. The laser is capable of operating in the burst mode at 1kHz, 2kHz, and 4kHz to investigate fast phenomena or at $30$ Hz to study 1 sec (or more) long discharges. The scattered light will be collected over an angular range of 60-120 degrees at 28 spatial points in the midplane covering the core region and edge plasma on both the low-field side (LFS) and the high-field side (HFS). Simulation data is used to determine the optimum location of Thomson scattering measurement points to effectively resolve the edge pedestal in the LFS and HFS regions under different triangularity conditions. Each scattering signal will be spectrally resolved on five wavelength channels of a polychromator to obtain the electron temperature measurement. We will also present a method to monitor in-situ laser alignment in the core during calibrations and plasma operations., Comment: Five pages and six figures

The building sector in the European Union significantly impacts energy consumption and greenhouse gas emissions. The EU's Horizon 2050 initiative sets ambitious goals to reduce these impacts through enhanced building renovation rates. The CoE HiDALGO2 supports this initiative by developing high-performance computing solutions, specifically through the Urban Building pilot application, which utilizes advanced CI/CD methodologies to streamline simulation and deployment across various computational platforms, such as the EuroHPC JU supercomputers. The present work provides an overview of the Ktirio Urban Building framework (KUB), starting with an overview of the workflow and a description of some of the main ingredients of the software stack and discusses some current results performed on EuroHPC JU supercomputers using an innovative CI/CD pipeline.

The deployment of safe and trustworthy machine learning systems, and particularly complex black box neural networks, in real-world applications requires reliable and certified guarantees on their performance. The conformal prediction framework offers such formal guarantees by transforming any point into a set predictor with valid, finite-set, guarantees on the coverage of the true at a chosen level of confidence. Central to this methodology is the notion of the nonconformity score function that assigns to each example a measure of ''strangeness'' in comparison with the previously seen observations. While the coverage guarantees are maintained regardless of the nonconformity measure, the point predictor and the dataset, previous research has shown that the performance of a conformal model, as measured by its efficiency (the average size of the predicted sets) and its informativeness (the proportion of prediction sets that are singletons), is influenced by the choice of the nonconformity score function. The current work introduces the Penalized Inverse Probability (PIP) nonconformity score, and its regularized version RePIP, that allow the joint optimization of both efficiency and informativeness. Through toy examples and empirical results on the task of crop and weed image classification in agricultural robotics, the current work shows how PIP-based conformal classifiers exhibit precisely the desired behavior in comparison with other nonconformity measures and strike a good balance between informativeness and efficiency.

This study demonstrates that using spin-polarized deuterium-tritium (D-T) fuel with more deuterium than tritium can increase tritium burn efficiency (TBE) by at least an order of magnitude without compromising fusion power output, compared to unpolarized fuel. Although previous studies show that a low tritium fraction can enhance TBE, this strategy resulted in reduced fusion power density. The surprising improvement in TBE at fixed power reported here is due to the TBE increasing nonlinearly with decreasing tritium fraction but the fusion power density increasing roughly linearly with D-T cross section. A study is performed for an ARC-like tokamak producing 482 MW of fusion power with unpolarized 51:49 D-T fuel, finding the minimum startup tritium inventory ($I_{\mathrm{startup,min}}$) is 0.69 kg. By spin-polarizing half of the fuel and using a 57:43 D-T mix, $I_{\mathrm{startup,min}}$ is reduced to 0.08 kg, and fully spin-polarizing the fuel with a 61:39 D-T mix further reduces $I_{\mathrm{startup,min}}$ to 0.03 kg. Some ARC-like scenarios could achieve plasma ignition with relatively modest spin polarization. These findings indicate that, with advancements in helium divertor pumping efficiency, TBE values of approximately 10-40% could be achieved using low-tritium-fraction and spin-polarized fuel with minimal power loss. This would dramatically lower tritium startup inventory requirements and reduce the amount of on-site tritium. More generally than just for spin-polarized fuels, high plasma performance can be used to increase TBE. This strongly motivates the development of spin-polarized fuels and low-tritium-fraction operation for burning plasmas., Comment: 30 pages, 27 figures

In this work, we design a generative artificial intelligence (GAI) -based framework for joint resource allocation, beamforming, and power allocation in multi-cell multi-carrier non-orthogonal multiple access (NOMA) networks. We formulate the proposed problem as sum rate maximization problem. Next, we design a novel multi-task transformer (MTT) framework to handle the problem in real-time. To provide the necessary training set, we consider simplified but powerful mathematical techniques from the literature. Then, we train and test the proposed MTT. We perform simulation to evaluate the efficiency of the proposed MTT and compare its performance with the mathematical baseline., Comment: 5 pages, 3 figures, arXiv:submit/5582145 [cs.NI] 7 May 2024

In developing countries, several key sectors, including education, finance, agriculture, and healthcare, mainly deliver their services via mobile app technology on handheld devices. As a result, mobile app security has emerged as a paramount issue in developing countries. In this paper, we investigate the state of research on mobile app security, focusing on developing countries. More specifically, we performed a systematic literature review exploring the research directions taken by existing works, the different security concerns addressed, and the techniques used by researchers to highlight or address app security issues. Our main findings are: (1) the literature includes only a few studies on mobile app security in the context of developing countries ; (2) among the different security concerns that researchers study, vulnerability detection appears to be the leading research topic; (3) FinTech apps are revealed as the main target in the relevant literature. Overall, our work highlights that there is largely room for developing further specialized techniques addressing mobile app security in the context of developing countries., Comment: 38 pages, 7 figures, submitted to a well-known journal

Protein-protein interactions (PPIs) play a crucial role in numerous biological processes. Developing methods that predict binding affinity changes under substitution mutations is fundamental for modelling and re-engineering biological systems. Deep learning is increasingly recognized as a powerful tool capable of bridging the gap between in-silico predictions and in-vitro observations. With this contribution, we propose eGRAL, a novel SE(3) equivariant graph neural network (eGNN) architecture designed for predicting binding affinity changes from multiple amino acid substitutions in protein complexes. eGRAL leverages residue, atomic and evolutionary scales, thanks to features extracted from protein large language models. To address the limited availability of large-scale affinity assays with structural information, we generate a simulated dataset comprising approximately 500,000 data points. Our model is pre-trained on this dataset, then fine-tuned and tested on experimental data.

Function-as-a-Service (FaaS) offers a streamlined cloud computing paradigm, but existing centralized systems suffer from vendor lock-in and single points of failure. We propose DeFaaS, a decentralized FaaS system leveraging blockchain technology and decentralized API management. DeFaaS addresses these limitations by establishing a secure, transparent registry of functions on a blockchain and enabling applications to discover and invoke them. This approach fosters scalability, flexibility, enhanced security, and improved reliability. Furthermore, DeFaaS's architecture extends beyond decentralized FaaS, supporting other distributed computing scenarios like dApps, volunteer computing, and multi-cloud service meshes. DeFaaS represents a significant advancement in decentralized computing with the potential to unlock a multitude of novel applications and use cases., Comment: The 39th ACM/SIGAPP Symposium On Applied Computing

A new burn control scheme for tokamaks is described where the total fusion power is controlled by adjusting the plasma volume fraction that is packed into power dense regions. In an example spherical tokamak burning plasma, by modifying the plasma edge squareness the total fusion power is doubled at almost constant total plasma volume and fusion power density. Therefore, increased plasma squareness could be extremely beneficial to a fusion reactor and squareness control could be desirable for power load balancing. Experiments have observed the impact of increased edge squareness on modified core plasma volume, highlighting the practical relevance of this approach., Comment: 7 pages, 7 figures

Optimizing deep learning models requires large amounts of annotated images, a process that is both time-intensive and costly. Especially for semantic segmentation models in which every pixel must be annotated. A potential strategy to mitigate annotation effort is active learning. Active learning facilitates the identification and selection of the most informative images from a large unlabelled pool. The underlying premise is that these selected images can improve the model's performance faster than random selection to reduce annotation effort. While active learning has demonstrated promising results on benchmark datasets like Cityscapes, its performance in the agricultural domain remains largely unexplored. This study addresses this research gap by conducting a comparative study of three active learning-based acquisition functions: Bayesian Active Learning by Disagreement (BALD), stochastic-based BALD (PowerBALD), and Random. The acquisition functions were tested on two agricultural datasets: Sugarbeet and Corn-Weed, both containing three semantic classes: background, crop and weed. Our results indicated that active learning, especially PowerBALD, yields a higher performance than Random sampling on both datasets. But due to the relatively large standard deviations, the differences observed were minimal; this was partly caused by high image redundancy and imbalanced classes. Specifically, more than 89\% of the pixels belonged to the background class on both datasets. The absence of significant results on both datasets indicates that further research is required for applying active learning on agricultural datasets, especially if they contain a high-class imbalance and redundant images. Recommendations and insights are provided in this paper to potentially resolve such issues.

This paper delves into the dynamic landscape of artificial intelligence, specifically focusing on the burgeoning prominence of large language models (LLMs). We underscore the pivotal role of Reinforcement Learning from Human Feedback (RLHF) in augmenting LLMs' rationality and decision-making prowess. By meticulously examining the intricate relationship between human interaction and LLM behavior, we explore questions surrounding rationality and performance disparities between humans and LLMs, with particular attention to the Chat Generative Pre-trained Transformer. Our research employs comprehensive comparative analysis and delves into the inherent challenges of irrationality in LLMs, offering valuable insights and actionable strategies for enhancing their rationality. These findings hold significant implications for the widespread adoption of LLMs across diverse domains and applications, underscoring their potential to catalyze advancements in artificial intelligence.

Artificial Intelligence (AI) requires new ways of evaluating national technology use and strategy for African nations. We conduct a survey of existing 'readiness' assessments both for general digital adoption and for AI policy in particular. We conclude that existing global readiness assessments do not fully capture African states' progress in AI readiness and lay the groundwork for how assessments can be better used for the African context. We consider the extent to which these indicators map to the African context and what these indicators miss in capturing African states' on-the-ground work in meeting AI capability. Through case studies of four African nations of diverse geographic and economic dimensions, we identify nuances missed by global assessments and offer high-level policy considerations for how states can best improve their AI readiness standards and prepare their societies to capture the benefits of AI.

In this work, we conduct a comparative study on two deep unfolding mechanisms to efficiently perform power control in the next generation wireless networks. The power control problem is formulated as energy efficiency over multiple interference links. The problem is nonconvex. We employ fractional programming transformation to design two solutions for the problem. The first solution is a numerical solution while the second solution is a closed-form solution. Based on the first solution, we design a semi-unfolding deep learning model where we combine the domain knowledge of the wireless communications and the recent advances in the data-driven deep learning. Moreover, on the highlights of the closed-form solution, fully deep unfolded deep learning model is designed in which we fully leveraged the expressive closed-form power control solution and deep learning advances. In the simulation results, we compare the performance of the proposed deep learning models and the iterative solutions in terms of accuracy and inference speed to show their suitability for the real-time application in next generation networks., Comment: 11 pages, 13 figures

This research explores strategies for steering the output of large language models (LLMs) towards specific styles, such as sentiment, emotion, or writing style, by adding style vectors to the activations of hidden layers during text generation. We show that style vectors can be simply computed from recorded layer activations for input texts in a specific style in contrast to more complex training-based approaches. Through a series of experiments, we demonstrate the effectiveness of activation engineering using such style vectors to influence the style of generated text in a nuanced and parameterisable way, distinguishing it from prompt engineering. The presented research constitutes a significant step towards developing more adaptive and effective AI-empowered interactive systems., Comment: Will be published as findings paper at EACL2024 - 18th Conference of the European Chapter of the Association for Computational Linguistics

Technology is gaining more and more place in our classrooms today, making its investigation more relevant than ever. Therefore, the author of this paper aimed to explore the role and impact of technology on students reading. To do so, the author has examined several aspects related to the use of technology for teaching and learning, including how technology can promote independent learning, communication, collaboration, assist students with learning needs, and certain limits of technology, before conducting research using interviews and observations with 86 participants (students, teachers, and parents). The research focused particularly on what role technologies such as text-to-speech apps, translators, electronic dictionaries, online reading apps and websites, etc., can play on students' reading, how they can impact students reading, and how they are perceived by students, teachers, and parents. At the end, the author provides some suggestions to consider for an effective use of these technologies.

A gyrokinetic threshold model for pedestal width-height scaling prediction is applied to multiple devices and to a shaping and aspect-ratio scan giving $\Delta_{\mathrm{ped}} = 0.92 A^{1.04} \kappa^{-1.24} 0.38^{\delta} \beta_{\theta,\mathrm{ped}}^{1.05}$ for pedestal width $\Delta_{\mathrm{ped}}$, aspect-ratio $A$, elongation $\kappa$, triangularity $\delta$, and normalized pedestal height $\beta_{\theta,\mathrm{ped}}$. We also find a width-transport scaling $\Delta_{\mathrm{ped} } = 0.028 \left(q_e/\Gamma_e - 1.7 \right)^{1.5} \sim \eta_e ^{1.5}$ where $q_e$ and $\Gamma_e$ are turbulent electron heat and particle fluxes and $\eta_e = \nabla \ln T_e / \nabla \ln n_e$ for electron temperature $T_e$ and density $n_e$. Pedestals close to those limited by kinetic-ballooning-modes (KBMs) have modified turbulent transport properties compared to strongly driven KBMs. The role of flow shear is studied as a width-height scaling constraint and pedestal saturation mechanism for a standard and wide pedestal discharge., Comment: 34 pages, 16 figures

Cooking recipes are one of the most readily available kinds of procedural text. They consist of natural language instructions that can be challenging to interpret. In this paper, we propose a model to identify relevant information from recipes and generate a graph to represent the sequence of actions in the recipe. In contrast with other approaches, we use an unsupervised approach. We iteratively learn the graph structure and the parameters of a $\mathsf{GNN}$ encoding the texts (text-to-graph) one sequence at a time while providing the supervision by decoding the graph into text (graph-to-text) and comparing the generated text to the input. We evaluate the approach by comparing the identified entities with annotated datasets, comparing the difference between the input and output texts, and comparing our generated graphs with those generated by state of the art methods.

Understanding procedural texts, such as cooking recipes, is essential for enabling machines to follow instructions and reason about tasks, a key aspect of intelligent reasoning. In cooking, these instructions can be interpreted as a series of modifications to a food preparation. For a model to effectively reason about cooking recipes, it must accurately discern and understand the inputs and outputs of intermediate steps within the recipe. We present a new corpus of cooking recipes enriched with descriptions of intermediate steps that describe the input and output for each step. PizzaCommonsense serves as a benchmark for the reasoning capabilities of LLMs because it demands rigorous explicit input-output descriptions to demonstrate the acquisition of implicit commonsense knowledge, which is unlikely to be easily memorized. GPT-4 achieves only 26\% human-evaluated preference for generations, leaving room for future improvements., Comment: Findings of EMNLP 2024. The data is available at: https://github.com/adiallo07/PizzaCommonsense

We use a new gyrokinetic threshold model to predict a bifurcation in tokamak pedestal width-height scalings that depends strongly on plasma shaping and aspect-ratio. The bifurcation arises from the first and second stability properties of kinetic-ballooning-modes that yields wide and narrow pedestal branches, expanding the space of accessible pedestal widths and heights. The wide branch offers potential for edge-localized-mode-free pedestals with high core pressure. For negative triangularity, low-aspect-ratio configurations are predicted to give steeper pedestals than conventional-aspect-ratio. Both wide and narrow branches have been attained in tokamak experiments., Comment: 7 pages, 7 figures