Search

We consider the voter model with binary opinions on a random regular graph with $n$ vertices of degree $d \geq 3$, subject to a rewiring dynamics in which pairs of edges are rewired, i.e., broken into four half-edges and subsequently reconnected at random. A parameter $\nu \in (0,\infty)$ regulates the frequency at which the rewirings take place, in such a way that any given edge is rewired exponentially at a rate $\nu$ in the limit as $n\to\infty$. We show that, under the joint law of the random rewiring dynamics and the random opinion dynamics, the fraction of vertices with either one of the two opinions converges on time scale $n$ to the Fisher-Wright diffusion with an explicit diffusion constant $\vartheta_{d,\nu}$ in the limit as $n\to\infty$. In particular, we identify $\vartheta_{d,\nu}$ in terms of a continued-fraction expansion and analyse its dependence on $d$ and $\nu$. A key role in our analysis is played by the set of discordant edges, which constitutes the boundary between the sets of vertices carrying the two opinions., Comment: 53 pages

In this paper, we advance the network theory of aging and mortality by developing a causal mathematical model for the mortality rate. First, we show that in large networks, where health deficits accumulate at nodes representing health indicators, the modeling of network evolution with Poisson processes is universal and can be derived from fundamental principles. Second, with the help of two simplifying approximations, which we refer to as mean-field assumption and homogeneity assumption, we provide an analytical derivation of Gompertz law under generic and biologically relevant conditions. We identify the parameters in Gompertz law as a function of the parameters driving the evolution of the network, and illustrate our computations with simulations and analytic approximations., Comment: 28 pages, 10 figure

The present paper is a brief overview of random opinion dynamics on random graphs based on the Ising Lecture given by the author at the World Congress in Probability and Statistics, 12--16 August 2024, Bochum, Germany. The content is a snapshot of an interesting area of research that is developing rapidly., Comment: 13 pages, 11 figures

The present overview of interacting particle systems on random graphs collects the notes of a mini-course given by the authors at the Brazilian School of Probability, 5--9 August 2024, in Salvador, Bahia, Brazil. The content is a personal snapshot of an interesting area of research at the interface between probability theory, combinatorics, statistical physics and network science that is developing rapidly., Comment: 44 pages, 13 figures

We consider two-opinion voter models on dense dynamic random graphs. Our goal is to understand and describe the occurrence of consensus versus polarisation over long periods of time. The former means that all vertices have the same opinion, the latter means that the vertices split into two communities with different opinions and few disagreeing edges. We consider three models for the joint dynamics of opinions and graphs: one with a one-way feedback and two which are co-evolutionary, i.e., with a two-way feedback. In the first model only coexistence is attainable, meaning that both opinions survive, but with the presence of many disagreeing edges. In the second model only consensus prevails, while in the third model polarisation is possible. Our main results are functional laws of large numbers for the densities of the two opinions, functional laws of large numbers for the dynamic random graphs in the space of graphons, and a characterisation of the limiting densities in terms of Beta-distributions. Our results are supported by simulations. To prove our results we develop a novel method that involves coupling the co-evolutionary process to a mimicking process with one-way feedback. We expect that this method can be extended to other dense co-evolutionary models., Comment: 39 pages

Learning performance data describe correct and incorrect answers or problem-solving attempts in adaptive learning, such as in intelligent tutoring systems (ITSs). Learning performance data tend to be highly sparse (80\%\(\sim\)90\% missing observations) in most real-world applications due to adaptive item selection. This data sparsity presents challenges to using learner models to effectively predict future performance explore new hypotheses about learning. This article proposes a systematic framework for augmenting learner data to address data sparsity in learning performance data. First, learning performance is represented as a three-dimensional tensor of learners' questions, answers, and attempts, capturing longitudinal knowledge states during learning. Second, a tensor factorization method is used to impute missing values in sparse tensors of collected learner data, thereby grounding the imputation on knowledge tracing tasks that predict missing performance values based on real observations. Third, a module for generating patterns of learning is used. This study contrasts two forms of generative Artificial Intelligence (AI), including Generative Adversarial Networks (GANs) and Generate Pre-Trained Transformers (GPT) to generate data associated with different clusters of learner data. We tested this approach on an adult literacy dataset from AutoTutor lessons developed for Adult Reading Comprehension (ARC). We found that: (1) tensor factorization improved the performance in tracing and predicting knowledge mastery compared with other knowledge tracing techniques without data augmentation, showing higher relative fidelity for this imputation method, and (2) the GAN-based simulation showed greater overall stability and less statistical bias based on a divergence evaluation with varying simulation sample sizes compared to GPT.

In many data-driven applications, higher-order relationships among multiple objects are essential in capturing complex interactions. Hypergraphs, which generalize graphs by allowing edges to connect any number of nodes, provide a flexible and powerful framework for modeling such higher-order relationships. In this work, we introduce hypergraph diffusion wavelets and describe their favorable spectral and spatial properties. We demonstrate their utility for biomedical discovery in spatially resolved transcriptomics by applying the method to represent disease-relevant cellular niches for Alzheimer's disease.

We consider a discrete-time binary branching random walk with independent standard normal increments subject to a penalty $\b$ for every pair of particles that get within distance $\e$ of each other at any time. We give a precise description of the most likely configurations of the particles under this law for $N$ large and $\b,\e$ fixed. Particles spread out over a distance $2^{2N/3}$, essentially in finite time, and subsequently arrange themselves so that at time $2N/3$ they cover a grid of width $\e$ with one particle per site. After time $2N/3$, the bulk of the particles and their descendants do not move anymore, while the particles in a boundary layer of width $2^{N/3}$ form a ``staircase" to the particles in the bulk. At time $N$, each site in the boundary layer is occupied by $2^{N/3}$ particles., Comment: 20 pages, 4 figures

In this paper, we obtain a precise estimate of the probability that the sparse binomial random graph contains a large number of vertices in a triangle. The estimate of log of this probability is correct up to second order, and enables us to propose an exponential random graph model based on the number of vertices in a triangle. Specifically, by tuning a single parameter, we can with high probability induce any given fraction of vertices in a triangle. Moreover, in the proposed exponential random graph model we derive the large deviation principle for the number of edges. As a byproduct, we propose a consistent estimator of the tuning parameter., Comment: Follow up of arXiv:2112.06526v2. Revised version

This is the third in a series of three papers in which we study a lattice gas subject to Kawasaki dynamics at inverse temperature $\beta>0$ in a large finite box $\Lambda_\beta \subset \mathbb{Z}^2$ whose volume depends on $\beta$. Each pair of neighbouring particles has a negative binding energy $-U<0$, while each particle has a positive activation energy $\Delta>0$. The initial configuration is drawn from the grand-canonical ensemble restricted to the set of configurations where all the droplets are subcritical. Our goal is to describe, in the metastable regime $\Delta \in (U,2U)$ and in the limit as $\beta\to\infty$, how and when the system nucleates, i.e., creates a critical droplet somewhere in $\Lambda_\beta$ that subsequently grows by absorbing particles from the surrounding gas. In the first paper we showed that subcritical droplets behave as quasi-random walks. In the second paper we used the results in the first paper to analyse how subcritical droplets form and dissolve on multiple space-time scales when the volume is moderately large, namely, $|\Lambda_\beta| = \mathrm{e}^{\theta\beta}$ with $\Delta < \theta < 2\Delta-U$. In the present paper we consider the setting where the volume is very large, namely, $|\Lambda_\beta| = \mathrm{e}^{\Theta\beta}$ with $\Delta < \Theta < \Gamma-(2\Delta-U)$, where $\Gamma$ is the energy of the critical droplet in the local model with fixed volume, and use the results in the first two papers to identify the nucleation time and the tube of typical trajectories towards nucleation. We will see that in a very large volume critical droplets appear more or less independently in boxes of moderate volume, a phenomenon referred to as homogeneous nucleation. One of the key ingredients in the proof is an estimate showing that no information can travel between these boxes on relevant time scales., Comment: 21 pages, 5 figures. arXiv admin note: text overlap with arXiv:2304.14099

Collecting and annotating real-world data for the development of object detection models is a time-consuming and expensive process. In the military domain in particular, data collection can also be dangerous or infeasible. Training models on synthetic data may provide a solution for cases where access to real-world training data is restricted. However, bridging the reality gap between synthetic and real data remains a challenge. Existing methods usually build on top of baseline Convolutional Neural Network (CNN) models that have been shown to perform well when trained on real data, but have limited ability to perform well when trained on synthetic data. For example, some architectures allow for fine-tuning with the expectation of large quantities of training data and are prone to overfitting on synthetic data. Related work usually ignores various best practices from object detection on real data, e.g. by training on synthetic data from a single environment with relatively little variation. In this paper we propose a methodology for improving the performance of a pre-trained object detector when training on synthetic data. Our approach focuses on extracting the salient information from synthetic data without forgetting useful features learned from pre-training on real images. Based on the state of the art, we incorporate data augmentation methods and a Transformer backbone. Besides reaching relatively strong performance without any specialized synthetic data transfer methods, we show that our methods improve the state of the art on synthetic data trained object detection for the RarePlanes and DGTA-VisDrone datasets, and reach near-perfect performance on an in-house vehicle detection dataset., Comment: Submitted to and presented at SPIE Defense + Commercial Sensing 2024, 13 pages, 4 figures, 3 tables

Recent studies have investigated new approaches for communicating an autonomous vehicle's (AV) intent and awareness to pedestrians. This paper adds to this body of work by presenting the design and evaluation of in-situ projections on the road. Our design combines common traffic light patterns with aesthetic visual elements. We describe the iterative design process and the prototyping methods used in each stage. The final design concept was represented as a virtual reality simulation and evaluated with 18 participants in four different street crossing scenarios, which included three scenarios that simulated various degrees of system errors. We found that different design elements were able to support participants' confidence in their decision even when the AV failed to correctly detect their presence. We also identified elements in our design that needed to be more clearly communicated. Based on these findings, the paper presents a series of design recommendations for projection-based communication between AVs and pedestrians.

Intelligent Tutoring Systems (ITSs) have significantly enhanced adult literacy training, a key factor for societal participation, employment opportunities, and lifelong learning. Our study investigates the application of advanced AI models, including Large Language Models (LLMs) like GPT-4, for predicting learning performance in adult literacy programs in ITSs. This research is motivated by the potential of LLMs to predict learning performance based on its inherent reasoning and computational capabilities. By using reading comprehension datasets from the ITS, AutoTutor, we evaluate the predictive capabilities of GPT-4 versus traditional machine learning methods in predicting learning performance through five-fold cross-validation techniques. Our findings show that the GPT-4 presents the competitive predictive abilities with traditional machine learning methods such as Bayesian Knowledge Tracing, Performance Factor Analysis, Sparse Factor Analysis Lite (SPARFA-Lite), tensor factorization and eXtreme Gradient Boosting (XGBoost). While XGBoost (trained on local machine) outperforms GPT-4 in predictive accuracy, GPT-4-selected XGBoost and its subsequent tuning on the GPT-4 platform demonstrates superior performance compared to local machine execution. Moreover, our investigation into hyper-parameter tuning by GPT-4 versus grid-search suggests comparable performance, albeit with less stability in the automated approach, using XGBoost as the case study. Our study contributes to the field by highlighting the potential of integrating LLMs with traditional machine learning models to enhance predictive accuracy and personalize adult literacy education, setting a foundation for future research in applying LLMs within ITSs., Comment: 26TH International Conference on Human-Computer Interaction

We analyse the mixing profile of a random walk on a dynamic random permutation, focusing on the regime where the walk evolves much faster than the permutation. Two types of dynamics generated by random transpositions are considered: one allows for coagulation of permutation cycles only, the other allows for both coagulation and fragmentation. We show that for both types, after scaling time by the length of the permutation and letting this length tend to infinity, the total variation distance between the current distribution and the uniform distribution converges to a limit process that drops down in a single jump. This jump is similar to a one-sided cut-off, occurs after a random time whose law we identify, and goes from the value 1 to a value that is a strictly decreasing and deterministic function of the time of the jump, related to the size of the largest component in Erd\H{o}s-R\'enyi random graphs. After the jump, the total variation distance follows this function down to 0.

In this paper we consider the two-type Moran model with $N$ individuals. Each individual is assigned a resampling rate, drawn independently from a probability distribution ${\mathbb P}$ on ${\mathbb R}_+$, and a type, either $1$ or $0$. Each individual resamples its type at its assigned rate, by adopting the type of an individual drawn uniformly at random. Let $Y^N(t)$ denote the empirical distribution of the resampling rates of the individuals with type $1$ at time $Nt$. We show that if ${\mathbb P}$ has countable support and satisfies certain tail and moment conditions, then in the limit as $N\to\infty$ the process $(Y^N(t))_{t \geq 0}$ converges in law to the process $(S(t)\,\P)_{t \geq 0}$, in the so-called Meyer-Zheng topology, where $(S(t))_{t \geq 0}$ is the Fisher-Wright diffusion with diffusion constant $D$ given by $1/D = \int_{{\mathbb R}_+} (1/r)\,{\mathbb P}(\mathrm{d} r)$.

Let $G_n$ be an undirected finite graph on $n\in\mathbb{N}$ vertices labelled by $[n] = \{1,\ldots,n\}$. For $i \in [n]$, let $\Delta_{i,n}$ be the friendship bias of vertex $i$, defined as the difference between the average degree of the neighbours of vertex $i$ and the degree of vertex $i$ itself when $i$ is not isolated, and zero when $i$ is isolated. Let $\mu_n$ denote the friendship-bias empirical distribution, i.e., the measure that puts mass $\frac{1}{n}$ at each $\Delta_{i,n}$, $i \in [n]$. The friendship paradox says that $\int_{\mathbb{R}} x\mu_n(\mathrm{d}x) \geq 0$, with equality if and only if in each connected component of $G_n$ all the degrees are the same. We show that if $(G_n)_{n\in\mathbb{N}}$ is a sequence of sparse random graphs that converges to a rooted random tree in the sense of convergence locally in probability, then $\mu_n$ converges weakly to a limiting measure $\mu$ that is expressible in terms of the law of the rooted random tree. We study $\mu$ for four classes of sparse random graphs: the homogeneous Erd\H{o}s-R\'enyi random graph, the inhomogeneous Erd\H{o}s-R\'enyi random graph, the configuration model and the preferential attachment model. In particular, we compute the first two moments of $\mu$, identify the right tail of $\mu$, and argue that $\mu([0,\infty))\geq\tfrac{1}{2}$, a property we refer to as friendship paradox significance., Comment: To appear in Probability Theory and Related Fields

We analyse the largest eigenvalue of the adjacency matrix of the configuration model with large degrees, where the latter are treated as hard constraints. In particular, we compute the expectation of the largest eigenvalue for degrees that diverge as the number of vertices $n$ tends to infinity, uniformly on a scale between $1$ and $\sqrt{n}$, and show that a weak law of large numbers holds. We compare with what was derived in our earlier paper "Central limit theorem for the principal eigenvalue and eigenvector of Chung-Lu random graphs" for the Chung-Lu model, which in the regime considered represents the corresponding configuration model with soft constraints, and show that the expectation is shifted down by $1$ asymptotically. This shift is a signature of breaking of ensemble equivalence between the hard and soft (also known as micro-canonical and canonical) versions of the configuration model. The latter result generalizes a previous finding in "A spectral signature of breaking of ensemble equivalence for constrained random graphs" obtained in the case when all degrees are equal., Comment: 21 pages, 3 figures

We show how to capture both the non-unitary Page curve and replica wormhole-like contributions that restore unitarity in a toy quantum system with random dynamics. The motivation is to find the simplest dynamical model that captures this aspect of gravitational physics. In our model, we evolve with an ensemble of Hamiltonians with GUE statistics within microcanonical windows. The entropy of the averaged state gives the non-unitary curve, the averaged entropy gives the unitary curve, and the difference comes from matrix index contractions in the Haar averaging that connect the density matrices in a replica wormhole-like manner., Comment: 43 pages, 10 figures

In this paper we consider stochastic processes taking values in a set of continuum graphs we call graphemes, defined as equivalence classes of sequences of vertices labelled by N embedded in an uncountable Polish space (with the cardinality of the continuum), together with an N x N connection matrix with entries 0 or 1 specifying the absence or presence of edges between pairs of vertices. In particular, we construct a Markov process on a Polish state space G of graphemes suitable to describe the time-space path of countable graphs. The class of dynamics we propose arises by specifying simple rules for the evolution of finite graphs and passing to the limit of infinite graphs. The evolution of graphemes is characterised by well-posed martingale problems, and leads to strong Markov processes with the Feller property., Comment: 78 pages, 2 figures

Graph neural networks (GNNs) have emerged as a powerful tool for tasks such as node classification and graph classification. However, much less work has been done on signal classification, where the data consists of many functions (referred to as signals) defined on the vertices of a single graph. These tasks require networks designed differently from those designed for traditional GNN tasks. Indeed, traditional GNNs rely on localized low-pass filters, and signals of interest may have intricate multi-frequency behavior and exhibit long range interactions. This motivates us to introduce the BLIS-Net (Bi-Lipschitz Scattering Net), a novel GNN that builds on the previously introduced geometric scattering transform. Our network is able to capture both local and global signal structure and is able to capture both low-frequency and high-frequency information. We make several crucial changes to the original geometric scattering architecture which we prove increase the ability of our network to capture information about the input signal and show that BLIS-Net achieves superior performance on both synthetic and real-world data sets based on traffic flow and fMRI data.

Our research explores the use of natural language processing (NLP) methods to automatically classify entities for the purpose of knowledge graph population and integration with food system ontologies. We have created NLP models that can automatically classify organizations with respect to categories associated with environmental issues as well as Standard Industrial Classification (SIC) codes, which are used by the U.S. government to characterize business activities. As input, the NLP models are provided with text snippets retrieved by the Google search engine for each organization, which serves as a textual description of the organization that is used for learning. Our experimental results show that NLP models can achieve reasonably good performance for these two classification tasks, and they rely on a general framework that could be applied to many other classification problems as well. We believe that NLP models represent a promising approach for automatically harvesting information to populate knowledge graphs and aligning the information with existing ontologies through shared categories and concepts., Comment: Presented at IFOW 2023 Integrated Food Ontology Workshop at the Formal Ontology in Information Systems Conference (FOIS) 2023 in Sherbrooke, Quebec, Canada July 17-20th, 2023

In this paper we introduce and study renewal-reward processes in random environments where each renewal involves a reward taking values in a Banach space. We derive quenched large deviation principles and identify the associated rate functions in terms of variational formulas involving correctors. We illustrate the theory with three examples: compound Poisson processes in random environments, pinning of polymers at interfaces with disorder, and returns of Markov chains in dynamic random environments., Comment: Submitted to the special issue of Stochastic Processes and their Applications in honor of Francis Comets

The Brain Imaging Data Structure (BIDS) is a community-driven standard for the organization of data and metadata from a growing range of neuroscience modalities. This paper is meant as a history of how the standard has developed and grown over time. We outline the principles behind the project, the mechanisms by which it has been extended, and some of the challenges being addressed as it evolves. We also discuss the lessons learned through the project, with the aim of enabling researchers in other domains to learn from the success of BIDS.