Detecting the origin of information or infection spread in networks is a fundamental challenge with applications in misinformation tracking, epidemiology, and beyond. We study the multi-source detection problem: given snapshot observations of node infection status on a graph, estimate the set of source nodes that initiated the propagation. Existing methods either lack statistical guarantees or are limited to specific diffusion models and assumptions. We propose a novel conformal prediction framework that provides statistically valid recall guarantees for source set detection, independent of the underlying diffusion process or data distribution. Our approach introduces principled score functions to quantify the alignment between predicted probabilities and true sources, and leverages a calibration set to construct prediction sets with user-specified recall and coverage levels. The method is applicable to both single- and multi-source scenarios, supports general network diffusion dynamics, and is computationally efficient for large graphs. Empirical results demonstrate that our method achieves rigorous coverage with competitive accuracy, outperforming existing baselines in both reliability and scalability.

Given the significant potential of orthogonal time frequency space (OTFS) signals in advancing integrated sensing and communication (ISAC) systems, this paper investigates the target sensing enhancements in ISAC-OTFS systems. Specifically, a customized inter-range-cell interference (IRCI)-free range reconstruction method is proposed for high-quality target sensing. The resultant mainlobe signal-to-noise ratio (SNR) is derived as a function of OTFS waveform parameters, and the achievable maximum SNR is deduced to provide a benchmark for subsequent simulations. The optimization of OTFS waveforms for maximizing the target sensing SNR is formulated with hardware realization constraints (i.e., peak to average power ratio (PAPR) and energy constraints). The resultant non-convex multi-ratio fractional programming problem is solved using a hybrid algorithm named multi-ratio fractional programming and partial successive convex approximation (MR-FP-PSCA). Finally, the numerical results, including the IRCI-free target sensing method and the proposed optimization algorithm, demonstrate the effectiveness of the developed schemes.

Personalized federated learning (PFL) plays a pivotal role in ensuring efficient privacy preservation and secure collaborative learning. However, PFL faces significant challenges due to data heterogeneity and device diversity. To enhance personalization and robustness in PFL, we propose a novel model called FedNODE, which leverages hierarchical embeddings. FedNODE incorporates personalized, pseudo-generic, and fusion embeddings to facilitate hierarchical information representation. We utilize a hypernetwork based on neural ordinary differential equations (ODEs) within the server to generate backbone parameters for different clients, enabling the creation of personalized embeddings. Additionally, we introduce a pseudo-generic embedding based on a learnable vector to balance personalized and generic information. A neural ODE-based network follows the backbone module for each client, integrating personalized and pseudo-generic embeddings. To validate the efficacy of FedNODE, we conduct extensive evaluations across various classification datasets, encompassing diverse statistically heterogeneous settings and noisy scenarios. The results demonstrate that FedNODE achieves state-of-the-art performance.

The use of neural differential equation models in machine learning applications has gained significant traction in recent years. In particular, fractional differential equations (FDEs) have emerged as a powerful tool for capturing complex dynamics in various domains. While existing models have primarily focused on constant-order fractional derivatives, variable-order fractional operators offer a more flexible and expressive framework for modeling complex memory patterns. In this work, we introduce the Neural Variable-Order Fractional Differential Equation network (NvoFDE), a novel neural network framework that integrates variable-order fractional derivatives with learnable neural networks. Our framework allows for the modeling of adaptive derivative orders dependent on hidden features, capturing more complex feature-updating dynamics and providing enhanced flexibility. We conduct extensive experiments across multiple graph datasets to validate the effectiveness of our approach. Our results demonstrate that NvoFDE outperforms traditional constant-order fractional and integer models across a range of tasks, showcasing its superior adaptability and performance.

Fractional-order differential equations (FDEs) enhance traditional differential equations by extending the order of differential operators from integers to real numbers, offering greater flexibility in modeling complex dynamic systems with nonlocal characteristics. Recent progress at the intersection of FDEs and deep learning has catalyzed a new wave of innovative models, demonstrating the potential to address challenges such as graph representation learning. However, training neural FDEs has primarily relied on direct differentiation through forward-pass operations in FDE numerical solvers, leading to increased memory usage and computational complexity, particularly in large-scale applications. To address these challenges, we propose a scalable adjoint backpropagation method for training neural FDEs by solving an augmented FDE backward in time, which substantially reduces memory requirements. This approach provides a practical neural FDE toolbox and holds considerable promise for diverse applications. We demonstrate the effectiveness of our method in several tasks, achieving performance comparable to baseline models while significantly reducing computational overhead.

Place recognition (PR) aims at retrieving the query place from a database and plays a crucial role in various applications, including navigation, autonomous driving, and augmented reality. While previous multi-modal PR works have mainly focused on the same-view scenario in which ground-view descriptors are matched with a database of ground-view descriptors during inference, the multi-modal cross-view scenario, in which ground-view descriptors are matched with aerial-view descriptors in a database, remains underexplored. We propose AGPlace, a model that effectively integrates information from multi-modal ground sensors (cameras and LiDARs) to achieve accurate aerial-ground PR. AGPlace achieves effective aerial-ground cross-view PR by leveraging a manifold-based neural ordinary differential equation (ODE) framework with a multi-domain alignment loss. It outperforms existing state-of-the-art cross-view PR models on large-scale datasets. As most existing PR models are designed for ground-ground PR, we adapt these baselines into our cross-view pipeline. Experiments demonstrate that this direct adaptation performs worse than our overall model architecture AGPlace. AGPlace represents a significant advancement in multi-modal aerial-ground PR, with promising implications for real-world applications.

Graph-structured data is prevalent in many applications. In subgraph federated learning (FL), this data is distributed across clients, each with a local subgraph. Personalized subgraph FL aims to develop a customized model for each client to handle diverse data distributions. However, performance variation across clients remains a key issue due to the heterogeneity of local subgraphs. To overcome the challenge, we propose FedSheafHN, a novel framework built on a sheaf collaboration mechanism to unify enhanced client descriptors with efficient personalized model generation. Specifically, FedSheafHN embeds each client’s local subgraph into a server-constructed collaboration graph by leveraging graph-level embeddings and employing sheaf diffusion within the collaboration graph to enrich client representations. Subsequently, FedSheafHN generates customized client models via a server-optimized hypernetwork. Empirical evaluations demonstrate that FedSheafHN outperforms existing personalized subgraph FL methods on various graph datasets. Additionally, it exhibits fast model convergence and effectively generalizes to new clients.

We consider the multiple hypothesis testing (MHT) problem over the joint domain formed by a graph and a measure space. On each sample point of this joint domain, we assign a hypothesis test and a corresponding p-value. The goal is to make decisions for all hypotheses simultaneously, using all available p-values. In practice, this problem resembles the detection problem over a sensor network during a period of time. To solve this problem, we extend the traditional two-groups model such that the prior probability of the null hypothesis and the alternative distribution of p-values can be inhomogeneous over the joint domain. We model the inhomogeneity via a generalized graph signal. This more flexible statistical model yields a more powerful detection strategy by leveraging the information from the joint domain.

We introduce a novel uncertainty principle for generalized graph signals that extends classical time-frequency and graph uncertainty principles into a unified framework. By defining joint vertex-time and spectral-frequency spreads, we quantify signal localization across these domains, revealing a trade-off between them. This framework allows us to identify a class of signals with maximal energy concentration in both domains, forming the fundamental atoms for a new joint vertex-time dictionary. This dictionary enhances signal reconstruction under practical constraints, such as intermittent data, commonly encountered in sensor and social networks. Numerical experiments on real-world datasets demonstrate the effectiveness of the proposed approach, showing improved reconstruction accuracy and noise robustness compared to existing methods.

Many works on graph neural networks (GNNs) focus on graph topologies and analyze graph-related operations to enhance performance on tasks such as node classification. In this paper, we propose to understand GNNs based on a feature-centric approach. Our main idea is to treat the features of nodes from each label class as a whole, from which we can identify the centroid. The convex hull of these centroids forms a simplex called the feature centroid simplex, where a simplex is a high-dimensional generalization of a triangle. We borrow ideas from coarse geometry to analyze the geometric properties of the feature centroid simplex by comparing them with basic geometric models, such as regular simplexes and degenerate simplexes. Such a simplex provides a simple platform to understand graph-based feature aggregation, including phenomena such as heterophily, oversmoothing, and feature re-shuffling. Based on the theory, we also identify simple and useful tricks for the node classification task.

Conventional image sensors have limited dynamic range, causing saturation in high-dynamic-range (HDR) scenes. Modulo cameras address this by folding incident irradiance into a bounded range, yet require specialized unwrapping algorithms to reconstruct the underlying signal. Unlike HDR recovery, which extends dynamic range from conventional sampling, modulo recovery restores actual values from folded samples. Despite being introduced over a decade ago, progress in modulo image recovery has been slow, especially in the use of modern deep learning techniques. In this work, we demonstrate that standard HDR methods are unsuitable for modulo recovery. Transformers, however, can capture global dependencies and spatial-temporal relationships crucial for resolving folded video frames. Still, adapting existing Transformer architectures for modulo recovery demands novel techniques. To this end, we present Selective Spatiotemporal Vision Transformer (SSViT), the first deep learning framework for modulo video reconstruction. SSViT employs a token selection strategy to improve efficiency and concentrate on the most critical regions. Experiments confirm that SSViT produces high-quality reconstructions from 8-bit folded videos and achieves state-of-the-art performance in modulo video recovery.

The integration of differential equations with neural networks has created powerful tools for modeling complex dynamics effectively across diverse machine learning applications. While standard integer-order neural ordinary differential equations (ODEs) have shown considerable success, they are limited in their capacity to model systems with memory effects and historical dependencies. Fractional calculus offers a mathematical framework capable of addressing this limitation, yet most current fractional neural networks use static memory weightings that cannot adapt to input-specific contextual requirements. This paper proposes a generalized neural Fractional Attention Differential Equation (FADE), which combines the memory-retention capabilities of fractional calculus with contextual learnable attention mechanisms. Our approach replaces fixed kernel functions in fractional operators with neural attention kernels that adaptively weight historical states based on their contextual relevance to current predictions. This allows our framework to selectively emphasize important temporal dependencies while filtering less relevant historical information. Our theoretical analysis establishes solution boundedness, problem well-posedness, and numerical equation solver convergence properties of the proposed model. Furthermore, through extensive evaluation on tasks such as fluid flow, graph learning problems and spatio-temporal traffic flow forecasting, we demonstrate that our adaptive attention-based fractional framework outperforms both integer-order neural ODE models and existing fractional approaches. The results confirm that our framework provides superior modeling capacity for complex dynamics with varying temporal dependencies. The code is available at \url{https://github.com/cuiwjTech/NeurIPS2025_FADE}.

Data is used widely by service providers as input to inference systems to perform decision making for authorized tasks. The raw data however allows a service provider to infer other sensitive information it has not been authorized for. We formulate a data-driven inference privacy preserving framework that sanitizes data to prevent leakage of sensitive information present in the raw data while ensuring that the sanitized data is still compatible with the service provider's legacy inference system and provides maximal utility. We propose to use maximal correlation as a privacy metric and show its advantage over the variational method for approximating mutual information. We develop a practical implementation of maximal correlation and derive sufficient conditions under which the data-driven implementation converges to the true privacy metric in the large training sample size regime. Furthermore, we adopt maximum mean discrepancy and adversarial domain discriminator as techniques to regularize the domain of the sanitized data to ensure its legacy compatibility. Finally, we develop a deep learning model as an example of the proposed inference privacy framework. Numerical experiments verify the feasibility of our approach.

Quantifying operational risks of power systems under multiple uncertainties and determining necessary reserves present complex challenges. This article introduces a risk dispatch approach based on uncertain intervals to address the conservatism of existing interval optimization methods. We use an uncertain model based on an interval random variable (IRV), substituting the unknown probability distributions of random variables. We establish a framework combining IRV and probabilistic random variable for risk quantification, leading to a risk evaluation model centered on uncertain interval–probabilistic conditional value-at-risk. We establish the joint probability distribution of wind power and load. Following this, we introduce reserve models based on the relative positions of actual and prediction intervals. Subsequently, we present an enhanced economic dispatch based on uncertain intervals to achieve more accurate results. Our proposed method breaks down the uncertain interval-based dispatch problem into two suboptimization models. We demonstrate the effectiveness of this method by applying it to IEEE-39 and IEEE-118 bus systems.

Due to the challenges in achieving accurate probabilities, representing uncertainty as intervals helps mitigate issues arising from the lack of distribution information. However, current interval-based methods for optimization modeling in power grids fail to fully capture the uncertainties of interval variables, leading to higher reserve costs. To overcome the conservatism of existing dispatch models, this work develops a novel uncertain interval variable (UIV) for risk assessment, where the radius of an interval is treated as a random variable. Inspired by the Affine algorithm, we propose a novel uncertain interval-based conditional value-at-risk (CVaR) metric, called UP-CVaR, for multiple random variables. The dispatch results for the New England 39-bus and 118-bus systems show that the proposed method can obtain tighter interval dispatch results compared to existing economic dispatch (ED) models. Moreover, as the stochastic level of wind power (mean and standard deviation) increases, the range of scheduling results expands.

In this work, we rigorously investigate the robustness of graph neural fractional-order differential equation (FDE) models. This framework extends beyond traditional graph neural (integer-order) ordinary differential equation (ODE) models by implementing the time-fractional Caputo derivative. Utilizing fractional calculus allows our model to consider long-term memory during the feature updating process, diverging from the memoryless Markovian updates seen in traditional graph neural ODE models. The superiority of graph neural FDE models over graph neural ODE models has been established in environments free from attacks or perturbations. While traditional graph neural ODE models have been verified to possess a degree of stability and resilience in the presence of adversarial attacks in existing literature, the robustness of graph neural FDE models, especially under adversarial conditions, remains largely unexplored. This paper undertakes a detailed assessment of the robustness of graph neural FDE models. We establish a theoretical foundation outlining the robustness characteristics of graph neural FDE models, highlighting that they maintain more stringent output perturbation bounds in the face of input and graph topology disturbances, compared to their integer-order counterparts. Our empirical evaluations further confirm the enhanced robustness of graph neural FDE models, highlighting their potential in adversarially robust applications.

Point cloud registration is a crucial technique in 3D computer vision with a wide range of applications. However, this task can be challenging, particularly in large fields of view with dynamic objects, environmental noise, or other perturbations. To address this challenge, we propose a model called PosDiffNet. Our approach performs hierarchical registration based on window-level, patch-level, and point-level correspondence. We leverage a graph neural partial differential equation (PDE) based on Beltrami flow to obtain high-dimensional features and position embeddings for point clouds. We incorporate position embeddings into a Transformer module based on a neural ordinary differential equation (ODE) to efficiently represent patches within points. We employ the multi-level correspondence derived from the high feature similarity scores to facilitate alignment between point clouds. Subsequently, we use registration methods such as SVD-based algorithms to predict the transformation using corresponding point pairs. We evaluate PosDiffNet on several 3D point cloud datasets, verifying that it achieves state-of-the-art (SOTA) performance for point cloud registration in large fields of view with perturbations. The implementation code of experiments is available at https://github.com/AI-IT-AVs/PosDiffNet.

The utilization of multi-modal sensor data in visual place recognition (VPR) has demonstrated enhanced performance compared to single-modal counterparts. Nonetheless, integrating additional sensors comes with elevated costs and may not be feasible for systems that demand lightweight operation, thereby impacting the practical deployment of VPR. To address this issue, we resort to knowledge distillation, which empowers single-modal students to learn from cross-modal teachers without introducing additional sensors during inference. Despite the notable advancements achieved by current distillation approaches, the exploration of feature relationships remains an under-explored area. In order to tackle the challenge of cross-modal distillation in VPR, we present DistilVPR, a novel distillation pipeline for VPR. We propose leveraging feature relationships from multiple agents, including self-agents and cross-agents for teacher and student neural networks. Furthermore, we integrate various manifolds, characterized by different space curvatures for exploring feature relationships. This approach enhances the diversity of feature relationships, including Euclidean, spherical, and hyperbolic relationship modules, thereby enhancing the overall representational capacity. The experiments demonstrate that our proposed pipeline achieves state-of-the-art performance compared to other distillation baselines. We also conduct necessary ablation studies to show design effectiveness. The code is released at: https://github.com/sijieaaa/DistilVPR

We introduce the FRactional-Order graph Neural Dynamical network (FROND), a new continuous graph neural network (GNN) framework. Unlike traditional continuous GNNs that rely on integer-order differential equations, FROND employs the Caputo fractional derivative to leverage the non-local properties of fractional calculus. This approach enables the capture of long-term dependencies in feature updates, moving beyond the Markovian update mechanisms in conventional integer-order models and offering enhanced capabilities in graph representation learning. We offer an interpretation of the node feature updating process in FROND from a non-Markovian random walk perspective when the feature updating is particularly governed by a diffusion process. We demonstrate analytically that oversmoothing can be mitigated in this setting. Experimentally, we validate the FROND framework by comparing the fractional adaptations of various established integer-order continuous GNNs, demonstrating their consistently improved performance and underscoring the framework's potential as an effective extension to enhance traditional continuous GNNs. The code is available at \url{https://github.com/zknus/ICLR2024-FROND}.

Traditionally, graph neural networks have been trained using a single observed graph. However, the observed graph represents only one possible realization. In many applications, the graph may encounter uncertainties, such as having erroneous or missing edges, as well as edge weights that provide little informative value. To address these challenges and capture additional information previously absent in the observed graph, we introduce latent variables to parameterize and generate multiple graphs. The parameters follow an unknown distribution to be estimated. We propose a formulation in terms of maximum likelihood estimation of the network parameters. Therefore, it is possible to devise an algorithm based on Expectation-Maximization (EM). Specifically, we iteratively determine the distribution of the graphs using a Markov Chain Monte Carlo (MCMC) method, incorporating the principles of PAC-Bayesian theory. Numerical experiments demonstrate improvements in performance against baseline models on node classification for both heterogeneous and homogeneous graphs.

This work introduces two novel neural spike detection schemes intended for use in next-generation neuromorphic brain-machine interfaces (iBMIs). The first, an Event-based Spike Detector (Ev-SPD) which examines the temporal neighborhood of a neural event for spike detection, is designed for in-vivo processing and offers high sensitivity and decent accuracy (94-97%). The second, Neural Network-based Spike Detector (NN-SPD) which operates on hybrid temporal event frames, provides an off-implant solution using shallow neural networks with impressive detection accuracy (96-99%) and minimal false detections. These methods are evaluated using a synthetic dataset with varying noise levels and validated through comparison with ground truth data. The results highlight their potential in next-gen neuromorphic iBMI systems and emphasize the need to explore this direction further to understand their resource-efficient and high-performance capabilities for practical iBMI settings.

Topological signal processing (TSP) utilizes simplicial complexes to model structures with higher order than vertices and edges. In this paper, we study the transferability of TSP via a generalized higher-order version of graphon, known as complexon. We recall the notion of a complexon as the limit of a simplicial complex sequence [1]. Inspired by the graphon shift operator and message-passing neural network, we construct a marginal complexon and complexon shift operator (CSO) according to components of all possible dimensions from the complexon. We investigate the CSO's eigenvalues and eigenvectors and relate them to a new family of weighted adjacency matrices. We prove that when a simplicial complex signal sequence converges to a complexon signal, the eigenvalues, eigenspaces, and Fourier transform of the corresponding CSOs converge to that of the limit complexon signal. This conclusion is further verified by two numerical experiments. These results hint at learning transferability on large simplicial complexes or simplicial complex sequences, which generalize the graphon signal processing framework.

We introduce the Distributed-order fRActional Graph Operating Network (DRAGON), a novel continuous Graph Neural Network (GNN) framework that incorporates distributed-order fractional calculus. Unlike traditional continuous GNNs that utilize integer-order or single fractional-order differential equations, DRAGON uses a learnable probability distribution over a range of real numbers for the derivative orders. By allowing a flexible and learnable superposition of multiple derivative orders, our framework captures complex graph feature updating dynamics beyond the reach of conventional models. We provide a comprehensive interpretation of our framework's capability to capture intricate dynamics through the lens of a non-Markovian graph random walk with node feature updating driven by an anomalous diffusion process over the graph. Furthermore, to highlight the versatility of the DRAGON framework, we conduct empirical evaluations across a range of graph learning tasks. The results consistently demonstrate superior performance when compared to traditional continuous GNN models. The implementation code is available at \url{https://github.com/zknus/NeurIPS-2024-DRAGON}.

In urban areas with tall buildings and narrow streets, signal distortions from multipath and non-line-of-sight (NLOS) conditions significantly affect the localization accuracy using Long-Term Evolution (LTE) signals. To address these limitations and improve localization accuracy, we propose a teacher-student transfer learning framework based on graph neural network (GNN), utilizing LTE networks and receiver arrays. For the challenges of limited real data, our proposed model can effectively improve performance through fine-tuning with a generated synthetic dataset. Experimental findings validate the efficacy of our method, showcasing significant accuracy improvements of 41.3% and 53.3% for synthetic and real data, respectively, compared to existing techniques. Our approach outperforms conventional localization methods and alternative machine learning models, emphasizing its superior performance.

Point cloud registration is a fundamental technique in 3-D computer vision with applications in graphics, autonomous driving, and robotics. However, registration tasks under challenging conditions, under which noise or perturbations are prevalent, can be difficult. We propose a robust point cloud registration approach that leverages graph neural partial differential equations (PDEs) and heat kernel signatures. Our method first uses graph neural PDE modules to extract high-dimensional features from point clouds by aggregating information from the 3-D point neighborhood, thereby enhancing the robustness of the feature representations. Then, we incorporate heat kernel signatures into an attention mechanism to efficiently obtain corresponding keypoints. Finally, a singular value decomposition (SVD) module with learnable weights is used to predict the transformation between two point clouds. Empirical experiments on a 3-D point cloud dataset demonstrate that our approach not only achieves state-of-the-art performance for point cloud registration but also exhibits better robustness to additive noise or 3-D shape perturbations.

Electric power systems have been increasingly subjected to false data injection attacks (FDIAs) and adversarial examples, which inject well-designed disturbance signals into the measurements, and thereby generate erroneous state estimation (SE) results. The present work addresses this issue by proposing a robust data-driven attack detection algorithm. We apply a novel metric denoted as Euclidian distance similarity ratio for detecting stealthy attack during the SE process. Second, two different deep Q networks are, respectively, employed for detecting FDIAs and adversarial examples based on their respective inflection points (IPs). We also propose sufficient and necessary conditions for the successful detection of adversarial examples based on the corresponding analyses of IPs. Finally, two networks are trained using deep reinforcement learning. The effectiveness of the proposed robust detection method is demonstrated based on simulations involving IEEE 14, 57, and 118 bus power systems.

Place recognition plays a crucial role in the fields of robotics and computer vision, finding applications in areas such as autonomous driving, mapping, and localization. Place recognition identifies a place using query sensor data and a known database. One of the main challenges is to develop a model that can deliver accurate results while being robust to environmental variations. We propose two multi-modal place recognition models, namely PRFusion and PRFusion++. PRFusion utilizes global fusion with manifold metric attention, enabling effective interaction between features without requiring camera-LiDAR extrinsic calibrations. In contrast, PRFusion++ assumes the availability of extrinsic calibrations and leverages pixel-point correspondences to enhance feature learning on local windows. Additionally, both models incorporate neural diffusion layers, which enable reliable operation even in challenging environments. We verify the state-of-the-art performance of both models on three large-scale benchmarks. Notably, they outperform existing models by a substantial margin of +3.0 AR@1 on the demanding Boreas dataset. Furthermore, we conduct ablation studies to validate the effectiveness of our proposed methods.

This paper investigates integrated waveform design for radar systems to simultaneously achieve a low probability of intercept (LPI) by an adversary electronic support measure (ESM) system while maintaining communications with other radar nodes. A reconfigurable intelligent surface (RIS) is exploited and jointly designed to enhance the achievable performance of the whole system. LPI detection based on the ESM’s feature analysis on the radar waveform is achieved under the constraints of a desired signal-to-interference-and-noise ratio for the radar detection and a desired signal-to-noise ratio-based quality of service for the communication channels. To deal with the resulting non-convex optimization problem, a suboptimal composite algorithm with polynomial complexity is proposed. The initial feasible solution of the algorithm is efficiently obtained through an asymptotic optimization problem. Finally, the complexity of the proposed algorithm is analyzed. Simulation results including comparisons with baselines highlight the effectiveness of the proposed scheme.

Camera relocalization has various applications in autonomous driving. Previous camera pose regression models consider only ideal scenarios where there is little environmental perturbation. To deal with challenging driving environments that may have changing seasons, weather, illumination, and the presence of unstable objects, we propose RobustLoc, which derives its robustness against perturbations from neural differential equations. Our model uses a convolutional neural network to extract feature maps from multi-view images, a robust neural differential equation diffusion block module to diffuse information interactively, and a branched pose decoder with multi-layer training to estimate the vehicle poses. Experiments demonstrate that RobustLoc surpasses current state-of-the-art camera pose regression models and achieves robust performance in various environments. Our code is released at: https://github.com/sijieaaa/RobustLoc

LiDAR relocalization plays a crucial role in many fields, including robotics, autonomous driving, and computer vision. LiDAR-based retrieval from a database typically incurs high computation storage costs and can lead to globally inaccurate pose estimations if the database is too sparse. On the other hand, pose regression methods take images or point clouds as inputs and directly regress global poses in an end-to-end manner. They do not perform database matching and are more computationally efficient than retrieval techniques. We propose HypLiLoc, a new model for LiDAR pose regression. We use two branched back-bones to extract 3D features and 2D projection features, respectively. We consider multi-modal feature fusion in both Euclidean and hyperbolic spaces to obtain more effective feature representations. Experimental results indicate that HypLiLoc achieves state-of-the-art performance in both outdoor and indoor datasets. We also conduct extensive ablation studies on the framework design, which demonstrate the effectiveness of multi-modal feature extraction and multi-space embedding. Our code is released at: https://github.com/sijieaaa/HypLiLoc

Time series signals such as EEG signals may have large variability across different individuals, making it difficult to sample without distortion or clipping using the same sensor for different individuals. Modulo sampling allows one to overcome the problem of signal clipping in the case where the signal has a very high dynamic range. This paper studies the problem of recovering a time series signal from under-determined modulo observations. We propose a deep learning method for modulo signal recovery, which can be applied to recover folded EEG signals. We make the first attempt to introduce the Transformer framework to modulo signal recovery. In addition, for efficiency and robustness, we introduce a modification of the Transformer module by inserting a learnable pre-estimation. The experiment on the real data demonstrates the superior performance of the proposed algorithm.

In generalized graph signal processing (GGSP), a function (an element from a separable Hilbert space) is associated with each vertex. To perform non-linear filtering and regression under the GGSP framework, we formulate an operator-valued kernel ridge regression (KRR) filtering approach. Under a specific choice of separable kernels, we show that this problem is equivalent to learning a nonlinear frequency response on each frequency band. We specify the choice of the reproducing kernel according to the signal’s spectral properties and discuss its effect on the learning result. The proposed approach is validated on a real dataset and demonstrated to outperform other competing methods.

Image matching identifies matching street landmark patches between the images captured by a vehicular camera and those stored in a database. Applications include autonomous driving perception and localization. However, in practical scenarios, challenging conditions such as changing weather, illumination, and dynamic objects result in perturbations of the captured images, leading to inaccurate matching. To achieve robust landmark patch matching, we present a method, named GRAND-Mat, which leverages a neural diffusion over graph embeddings to counteract perturbations. We first extract high-dimensional features of landmark patches using a ResNet. Then, we utilize graph neural diffusion models to aggregate the self and cross-graph information from these features. Furthermore, we apply feature similarity learning to acquire the final matching score. We evaluate the performance of our model on a street scene dataset, which demonstrates state-of-the-art matching performance under additive perturbations.

In node classification using graph neural networks (GNNs), a typical model generates logits for different class labels at each node. A softmax layer often outputs a label prediction based on the largest logit. We demonstrate that it is possible to infer hidden graph structural information from the dataset using these logits. We introduce the key notion of label non-uniformity, which is derived from the Wasserstein distance between the softmax distribution of the logits and the uniform distribution. We demonstrate that nodes with small label non-uniformity are harder to classify correctly. We theoretically analyze how the label non-uniformity varies across the graph, which provides insights into boosting the model performance: increasing training samples with high non-uniformity or dropping edges to reduce the maximal cut size of the node set of small non-uniformity. These mechanisms can be easily added to a base GNN model. Experimental results demonstrate that our approach improves the performance of many benchmark base models.

In the graph node embedding problem, embedding spaces can vary significantly for different data types, leading to the need for different GNN model types. In this paper, we model the embedding update of a node feature as a Hamiltonian orbit over time. Since the Hamiltonian orbits generalize the exponential maps, this approach allows us to learn the underlying manifold of the graph in training, in contrast to most of the existing literature that assumes a fixed graph embedding manifold with a closed exponential map solution. Our proposed node embedding strategy can automatically learn, without extensive tuning, the underlying geometry of any given graph dataset even if it has diverse geometries. We test Hamiltonian functions of different forms and verify the performance of our approach on two graph node embedding downstream tasks: node classification and link prediction. Numerical experiments demonstrate that our approach adapts better to different types of graph datasets than popular state-of-the-art graph node embedding GNNs. The code is available at https://github.com/zknus/Hamiltonian-GNN.

Graph neural networks (GNNs) have shown promising results across various graph learning tasks, but they often assume homophily, which can result in poor performance on heterophilic graphs. The connected nodes are likely to be from different classes or have dissimilar features on heterophilic graphs. In this paper, we propose a novel GNN that incorporates the principle of heterophily by modeling the flow of information on nodes using the convection-diffusion equation (CDE). This allows the CDE to take into account both the diffusion of information due to homophily and the ``convection'' of information due to heterophily. We conduct extensive experiments, which suggest that our framework can achieve competitive performance on node classification tasks for heterophilic graphs, compared to the state-of-the-art methods. The code is available at https://github.com/zknus/Graph-Diffusion-CDE.

This work explores the architectural trade-offs and implications of a neuromorphic compression based neural sensing architecture with address-event representation inspired readout protocol for massively parallel, next-gen wireless iBMI. We use quantitative metrics such as root-mean-square error and correlation coefficient between the original and recovered signal to assess the effect of neuromorphic compression on spike shape, and spike detection accuracy, sensitivity, and false detection rate to understand the effect of compression on downstream iBMI tasks. We demonstrate that a data compression ratio of$> 50$can be achieved by selective transmission of event pulses generated in different modes for large electrode arrays with a correlation coefficient of$\approx 0.9$and a spike detection accuracy of over 90%.

Graph neural networks (GNNs) are vulnerable to adversarial perturbations, including those that affect both node features and graph topology. This paper investigates GNNs derived from diverse neural flows, concentrating on their connection to various stability notions such as BIBO stability, Lyapunov stability, structural stability, and conservative stability. We argue that Lyapunov stability, despite its common use, does not necessarily ensure adversarial robustness. Inspired by physics principles, we advocate for the use of conservative Hamiltonian neural flows to construct GNNs that are robust to adversarial attacks. The adversarial robustness of different neural flow GNNs is empirically compared on several benchmark datasets under a variety of adversarial attacks. Extensive numerical experiments demonstrate that GNNs leveraging conservative Hamiltonian flows with Lyapunov stability substantially improve robustness against adversarial perturbations. The implementation code of experiments is available at \url{https://github.com/zknus/NeurIPS-2023-HANG-Robustness}.

Matching landmark patches from a real-time image captured by an on-vehicle camera with landmark patches in an image database plays an important role in various computer perception tasks for autonomous driving. Current methods focus on local matching for regions of interest and do not take into account spatial neighborhood relationships among the image patches, which typically correspond to objects in the environment. In this paper, we construct a spatial graph with the graph vertices corresponding to patches and edges capturing the spatial neighborhood information. We propose a joint feature and metric learning model with graph-based learning. We provide a theoretical basis for the graph-based loss by showing that the information distance between the distributions conditioned on matched and unmatched pairs is maximized under our framework. We evaluate our model using several street-scene datasets and demonstrate that our approach achieves state-of-the-art matching results.

For autonomous vehicles (AVs), visual perception techniques based on sensors like cameras play crucial roles in information acquisition and processing. In various computer perception tasks for AVs, it may be helpful to match landmark patches taken by an onboard camera with other landmark patches captured at a different time or saved in a street scene image database. To perform matching under challenging driving environments caused by changing seasons, weather, and illumination, we utilize the spatial neighborhood information of each patch. We propose an approach, named RobustMat, which derives its robustness to perturbations from neural differential equations. A convolutional neural ODE diffusion module is used to learn the feature representation for the landmark patches. A graph neural PDE diffusion module then aggregates information from neighboring landmark patches in the street scene. Finally, feature similarity learning outputs the final matching score. Our approach is evaluated on several street scene datasets and demonstrated to achieve state-of-the-art matching results under environmental perturbations.

Recent advances in Graph Neural Networks (GNNs) have achieved superior results in many challenging tasks, such as few-shot learning. Despite its capacity to learn and generalize a model from only a few annotated samples, GNN is limited in scalability, as deep GNN models usually suffer from severe over-fitting and over-smoothing. In this work, we propose a novel GNN framework with atriple-attention mechanism,i.e.node self-attention, neighbor attention, and layer memory attention, to tackle these challenges. We provide both theoretical analysis and illustrations to explain why the proposed attentive modules can improve GNN scalability for few-shot learning tasks. Our experiments show that the proposed Attentive GNN model outperforms the state-of-the-art few-shot learning methods using both GNN and non-GNN approaches. The improvement is consistent over the mini-ImageNet, tiered-ImageNet, CUB-200-2011, and Flowers-102 benchmarks, using both ConvNet-4 and ResNet-12 backbones, and under both the inductive and transductive settings. Furthermore, we demonstrate the superiority of our method for few-shot fine-grained and semi-supervised classification tasks with extensive experiments. The code for this work is publicly available athttps://github.com/chenghao-ch94/AGNN.

A reconfigurable intelligent surface (RIS) allows a reflection transmission path between a base station (BS) and user equipment (UE). In wireless localization, this reflection path aids in positioning accuracy, especially when the line-of-sight (LOS) path is subject to severe blockage and fading. In this paper, we develop a RIS-aided positioning framework to locate a UE in environments where the LOS path may or may not be available. We first estimate the RIS-aided channel parameters from the received signals at the UE. To infer the UE position and clock bias from the estimated channel parameters, we propose a fusion method consisting of weighted least squares over the estimates of the LOS and reflection paths. We show that this approximates the maximum likelihood estimator under the large-sample regime and when the estimates from different paths are independent. We then optimize the RIS phase shifts to improve the positioning accuracy and extend the proposed approach to the case with multiple BSs and UEs. We derive Cramér-Rao bound (CRB) and demonstrate numerically that our proposed positioning method approaches the CRB.

We consider a probabilistic model for graph signal processing (GSP) in a generalized framework where each vertex of a graph is associated with an element from a Hilbert space. We introduce the notion of joint wide-sense stationarity in this generalized GSP (GGSP) framework, which allows us to characterize a random graph process as a combination of uncorrelated oscillation modes across both the vertex and Hilbert space domains. We also propose a method for joint power spectral density estimation in case of missing features. Experiment results corroborate the effectiveness of our estimation approach.

Real-time data transmissions from a vehicle enhance road safety and traffic efficiency by aggregating data in a central server for data analytics. When drivers share their instantaneous vehicular information for a service provider to perform a legitimate task, a curious service provider may also infer private information it has not been authorized for. In this paper, we propose a privacy preservation framework based on the Hilbert Schmidt Independence Criterion (HSIC) to sanitize driving data to protect the vehicle’s trajectory from adversarial inference while ensuring the data is still useful for driver behavior detection. We develop a deep learning model to learn the HSIC sanitizer and demonstrate through two datasets that our approach achieves better utility-privacy trade-offs when compared to three other benchmarks.

Heterogeneous graphs have multiple node and edge types and are semantically richer than homogeneous graphs. To learn such complex semantics, many graph neural network approaches for heterogeneous graphs use metapaths to capture multi-hop interactions between nodes. Typically, features from non-target nodes are not incorporated into the learning procedure. However, there can be nonlinear, high-order interactions involving multiple nodes or edges. In this paper, we present Simplicial Graph Attention Network (SGAT), a simplicial complex approach to represent such high-order interactions by placing features from non-target nodes on the simplices. We then use attention mechanisms and upper adjacencies to generate representations. We empirically demonstrate the efficacy of our approach with node classification tasks on heterogeneous graph datasets and further show SGAT's ability in extracting structural information by employing random node features. Numerical experiments indicate that SGAT performs better than other current state-of-the-art heterogeneous graph learning methods.

Neural diffusion on graphs is a novel class of graph neural networks that has attracted increasing attention recently. The capability of graph neural partial differential equations (PDEs) in addressing common hurdles of graph neural networks (GNNs), such as the problems of over-smoothing and bottlenecks, has been investigated but not their robustness to adversarial attacks. In this work, we explore the robustness properties of graph neural PDEs. We empirically demonstrate that graph neural PDEs are intrinsically more robust against topology perturbation as compared to other GNNs. We provide insights into this phenomenon by exploiting the stability of the heat semigroup under graph topology perturbations. We discuss various graph diffusion operators and relate them to existing graph neural PDEs. Furthermore, we propose a general graph neural PDE framework based on which a new class of robust GNNs can be defined. We verify that the new model achieves comparable state-of-the-art performance on several benchmark datasets.

Intelligent transportation system significantly relies on accurate positioning information of land vehicles for both safety and non-safety related applications, such as hard-braking ahead warning and red-light violation warning. However, existing Global Navigation Satellite System (GNSS) based solutions suffer from positioning performance degradation in challenging environments, such as urban canyons and tunnels. In this paper, we focus on the positioning performance enhancement of land vehicles via cooperative positioning under a partial GNSS environment in a Vehicular Ad-hoc NETwork (VANET). The availability of Time-of-Flight (ToF) based inter-vehicle or vehicle-to-infrastructure ranges is verified via 5.9 GHz Dedicated Short-Range Communication (DSRC) vehicle-to-everything communication with RTS/CTS unicast mechanism. An inertial navigation sensor aided, tightly coupled integration approach for land vehicle cooperative positioning using DSRC ToF ranges and carrier frequency offset range-rates is proposed, where a digital map is used to constrain the position estimates. If available, the GNSS pseudorange and Doppler shift under partial GNSS environment can also be incorporated. A Rao–Blackwellized particle filter is utilized to estimate the unknown variables allowing for reduced computational complexity in comparison with the conventional particle filter. The posterior Cramer–Rao lower bound is also derived to give a theoretical performance guideline. Both simulation and experimental results show the validity of our proposed approach.

The problem of estimating event truths from conflicting agent opinions in a social network is investigated. An autoencoder learns the complex relationships between event truths, agent reliabilities and agent observations. A Bayesian network model is proposed to guide the learning process by modeling the relationship of the autoencoder's outputs with different variables. At the same time, it also models the social relationships between agents in the network. The proposed approach is unsupervised and is applicable when ground truth labels of events are unavailable. A variational inference method is used to jointly estimate the hidden variables in the Bayesian network and the parameters in the autoencoder. Experiments on three real datasets demonstrate that our proposed approach is competitive with, and in most cases better than, several state-of-the-art benchmark methods.

Workers participating in a crowdsourcing platform can have a wide range of abilities and interests. An important problem in crowdsourcing is the task recommendation problem, in which tasks that best match a particular worker's preferences and reliabilities are recommended to that worker. A task recommendation scheme that assigns tasks more likely to be accepted by a worker who is more likely to complete it reliably results in better performance for the task requester. Without prior information about a worker, his preferences and reliabilities need to be learned over time. In this article, we propose a multi-armed bandit (MAB) framework to learn a worker's preferences and his reliabilities for different categories of tasks. However, unlike the classical MAB problem, the reward from the worker's completion of a task is unobservable. We therefore include the use of gold tasks (i.e., tasks whose solutions are knowna prioriand which do not produce any rewards) in our task recommendation procedure. Our model could be viewed as a new variant of MAB, in which the random rewards can only be observed at those time steps where gold tasks are used, and the accuracy of estimating the expected reward of recommending a task to a worker depends on the number of gold tasks used. We show that the optimal regret is$O(\sqrt{n})$, where$n$is the number of tasks recommended to the worker. We develop three task recommendation strategies to determine the number of gold tasks for different task categories, and show that they are order optimal. Simulations verify the efficiency of our approaches.

Though deep learning has been applied successfully in many scenarios, malicious inputs with human-imperceptible perturbations can make it vulnerable in real applications. This paper proposes an error-correcting neural network (ECNN) that combines a set of binary classifiers to combat adversarial examples in the multi-class classification problem. To build an ECNN, we propose to design a code matrix so that the minimum Hamming distance between any two rows (i.e., two codewords) and the minimum shared information distance between any two columns (i.e., two partitions of class labels) are simultaneously maximized. Maximizing row distances can increase the system fault tolerance while maximizing column distances helps increase the diversity between binary classifiers. We propose an end-to-end training method for our ECNN, which allows further improvement of the diversity between binary classifiers. The end-to-end training renders our proposed ECNN different from the traditional error-correcting output code (ECOC) based methods that train binary classifiers independently. ECNN is complementary to other existing defense approaches such as adversarial training and can be applied in conjunction with them. We empirically demonstrate that our proposed ECNN is effective against the state-of-the-art white-box and black-box attacks on several datasets while maintaining good classification accuracy on normal examples.

A heterogeneous graph consists of different vertices and edges types. Learning on heterogeneous graphs typically employs meta-paths to deal with the heterogeneity by reducing the graph to a homogeneous network, guide random walks or capture semantics. These methods are however sensitive to the choice of meta-paths, with suboptimal paths leading to poor performance. In this paper, we propose an approach for learning on heterogeneous graphs without using meta-paths. Specifically, we decompose a heterogeneous graph into different homogeneous relation-type graphs, which are then combined to create higher-order relation-type representations. These representations preserve the heterogeneity of edges and retain their edge directions while capturing the interaction of different vertex types multiple hops apart. This is then complemented with attention mechanisms to distinguish the importance of the relation-type based neighbors and the relation-types themselves. Experiments demonstrate that our model generally outperforms other state-of-the-art baselines in the vertex classification task on three commonly studied heterogeneous graph datasets.

Deep neural networks (DNNs) are well-known to be vulnerable to adversarial attacks, where malicious human-imperceptible perturbations are included in the input to the deep network to fool it into making a wrong classification. Recent studies have demonstrated that neural Ordinary Differential Equations (ODEs) are intrinsically more robust against adversarial attacks compared to vanilla DNNs. In this work, we propose a neural ODE with Lyapunov-stable equilibrium points for defending against adversarial attacks (SODEF). By ensuring that the equilibrium points of the ODE solution used as part of SODEF are Lyapunov-stable, the ODE solution for an input with a small perturbation converges to the same solution as the unperturbed input. We provide theoretical results that give insights into the stability of SODEF as well as the choice of regularizers to ensure its stability. Our analysis suggests that our proposed regularizers force the extracted feature points to be within a neighborhood of the Lyapunov-stable equilibrium points of the SODEF ODE. SODEF is compatible with many defense methods and can be applied to any neural network's final regressor layer to enhance its stability against adversarial attacks.

To facilitate 5G-based positioning applications, Release 16 of the 3GPP 5G standard has defined the Positioning Reference Signal (PRS), which can be used to measure Time of Arrival (TOA) for downlink positioning. However, Orthogonal Frequency Division Multiplexing (OFDM) signals are sensitive and vulnerable to synchronization errors. Moreover, the highly configurable 5G PRS in Release 16 calls for a unique allocation pattern on the subcarriers. Existing timing recovery methods that have been employed for reference signals, which are evenly inserted in the subcarrier symbols, may not perform well. To solve the timing recovery issue of the OFDM signal through 5G standard-compliant PRS, we propose a three-stage timing recovery scheme. We use the 5G PRS as pilot symbols to estimate the path time delay and complete receiver sampling clock synchronization. We propose a generalized path time delay estimation method that can correct timing errors larger than one sample. In addition, we incorporate a delay-locked loop (DLL) that can track the PRS code-phase when the phase errors are within one sample, which showcases the precise positioning possible with a standard-compliant 5G New Radio (NR) signal.

In this paper, we propose a multi-level localization algorithm that breaks a centralized localization problem into a cluster-level distributed localization problem, where each cluster is a centralized unit. In contrast to fully distributed localization, the cluster-level distributed scheme results in reduction in contention, communication overheads, convergence time and energy consumption because cluster heads are responsible for the intra cluster positioning on behalf of the whole cluster. To generate a global map, the cluster heads communicate with their direct neighbors to carry out inter-cluster ranging and positioning. The proposed method is suitable for large ad-hoc networks where most agents are low-cost, low-power RF transceivers used for ranging only while some agents are integrated with microcomputers such as Raspberry Pis capable of running intra and inter-cluster localization algorithms. The proposed system can work without anchor nodes and thus it can be deployed in the environments such as urban canyon, inside multi-story buildings, airports, and underground shopping malls where access to anchors or Global Navigation Satellite System (GNSS) is limited or prohibitive. We exploit a hybrid of two well-known methods: multidimensional scaling (MDS) and extended Kalman filtering (EKF) to effectively construct local and global position maps, even in the absence of GNSS information, anchors, or a complete ranging matrix.

We propose a privacy preserving framework to sanitize data so as to eliminate private information while maximally retaining non-sensitive information. We regularize the domain of the sanitized data to make it compatible with a service provider's learning systems already in place for the raw data. Thus, our privacy preserving framework incurs no additional cost for the service provider. We present a probabilistic sanitizer to privatize the raw data and a variational method to approximate the mutual information between the sanitized data and raw data. We include maximum mean discrepancy and domain adaption as the domain regularization techniques, and average information leakage as the privacy metric. We present a deep learning model as an example of the proposed framework where the input data is an image. Numerical experiments verify the feasibility of our approach.

In view of the huge success of convolution neural networks (CNN) for image classification and object recognition, there have been attempts to generalize the method to general graph-structured data. One major direction is based on spectral graph theory. In this paper, we study the problem from a different perspective, by introducing parallel flow decomposition of graphs. The essential idea is to decompose a graph into families of non-intersecting one dimensional (1D) paths, after which, we may apply a 1D CNN along each family of paths. We demonstrate that the our method, which we call GFCN (graph flow convolutional network), is able to transfer CNN architectures to general graphs. We demonstrate effectiveness of the method with synthetic and real applications.

This paper considers the problem of the quickest detection of a change in distribution while taking privacy considerations into account. Our goal is to sanitize the signal to satisfy information privacy requirements while being able to detect a change quickly. We formulate the privacy-aware quickest change detection (QCD) problem by including a privacy constraint to Lorden's minimax formulation. We show that the Generalized Likelihood Ratio (GLR) CuSum achieves asymptotic optimality with a properly designed sanitization channel and formulate the design of this sanitization channel as an optimization problem. For computational tractability, a continuous relaxation for the discrete counting constraint is proposed and the augmented Lagrangian method is applied to obtain locally optimal solutions.

We consider the quasi-synchronization problem of a continuous time generalized Markovian switching heterogeneous network with time-varying connectivity, using pinned nodes that are event-triggered to reduce the frequency of controller updates and internode communications. We propose a pinning strategy algorithm to determine how many and which nodes should be pinned in the network. Based on the assumption that a network has limited control efficiency, we derive a criterion for stability, which relates the pinning feedback gains, the coupling strength, and the inner coupling matrix. By utilizing the stochastic Lyapunov stability analysis, we obtain sufficient conditions for exponential quasi-synchronization under our stochastic event-triggering mechanism, and a bound for the quasi-synchronization error. Numerical simulations are conducted to verify the effectiveness of the proposed control strategy.

We consider a decentralized detection network whose aim is to infer a public hypothesis of interest. However, the raw sensor observations also allow the fusion center to infer private hypotheses that we wish to protect. We consider the case where there are an uncountable number of private hypotheses belonging to an uncertainty set, and develop local privacy mappings at every sensor so that the sanitized sensor information minimizes the Bayes error of detecting the public hypothesis at the fusion center while achieving information privacy for all private hypotheses. We introduce the concept of a most favorable hypothesis (MFH) and show how to find an MFH in the set of private hypotheses. By protecting the information privacy of the MFH, information privacy for every other private hypothesis is also achieved. We provide an iterative algorithm to find the optimal local privacy mappings, and derive some theoretical properties of these privacy mappings. The simulation results demonstrate that our proposed approach allows the fusion center to infer the public hypothesis with low error while protecting information privacy of all the private hypotheses.

In a decentralized Internet of Things (IoT) network, a fusion center receives information from multiple sensors to infer a public hypothesis of interest. To prevent the fusion center from abusing the sensor information, each sensor sanitizes its local observation using a local privacy mapping, which is designed to achieve both inference privacy of a private hypothesis and data privacy of the sensor raw observations. Various inference and data privacy metrics have been proposed in the literature. We introduce the concept of privacy implication (with vanishing budget) to study the relationships between these privacy metrics. We propose an optimization framework in which both local differential privacy (data privacy) and information privacy (inference privacy) metrics are incorporated. In the parametric case where sensor observations' distributions are known a priori, we propose a two-stage local privacy mapping at each sensor, and show that such an architecture is able to achieve information privacy and local differential privacy to within the predefined budgets. For the nonparametric case where sensor distributions are unknown, we adopt an empirical optimization approach. Simulation and experiment results demonstrate that our proposed approaches allow the fusion center to accurately infer the public hypothesis while protecting both inference and data privacy.

We consider a linear dynamical system in which the state vector consists of both public and private states. One or more sensors make measurements of the state vector and sends information to a fusion center, which performs the final state estimation. To achieve an optimal tradeoff between the utility of estimating the public states and protection of the private states, the measurements at each time step are linearly compressed into a lower dimensional space. Under the centralized setting where all measurements are collected by a single sensor, we propose an optimization problem and an algorithm to find the best compression matrix. Under the decentralized setting where measurements are made separately at multiple sensors, each sensor optimizes its own local compression matrix. We propose methods to separate the overall optimization problem into multiple sub-problems that can be solved locally at each sensor. We consider the cases where there is no message exchange between the sensors; and where each sensor takes turns to transmit messages to the other sensors. Simulations and empirical experiments demonstrate the efficiency of our proposed approach in allowing the fusion center to estimate the public states with good accuracy while preventing it from estimating the private states accurately.

This paper considers the problem of quickest detection of a change in distribution where the signal may undergo both nuisance and critical changes. Our goal is to detect the critical change without raising a false alarm over the nuisance change. We formulate the quickest change detection (QCD) problem in the presence of a nuisance change following Lorden's formulation. We propose a window-limited sequential change detection procedure based on the generalized likelihood ratio test statistic for the problem of QCD in which both nuisance and critical changes may occur. We derive a recursive update scheme for our proposed test statistic and show that our test is asymptotically optimal under mild technical conditions. We compare our proposed stopping rules with a naive 2-stage stopping time, which attempts to detect the changes using separate CuSum stopping procedures for the nuisance and critical changes. Simulations suggest that our proposed stopping time outperforms the naive 2-stage procedures.

We develop a privacy-preserving distributed strategy over multitask diffusion networks, where each agent is interested in not only improving its local inference performance via in-network cooperation, but also protecting its own individual task against privacy leakage. In the proposed strategy, at each time instant, each agent sends a noisy estimate, which is its local intermediate estimate corrupted by a zero-mean additive noise, to its neighboring agents. We derive a sufficient condition to determine the amount of noise to add to each agent's intermediate estimate to achieve an optimal trade-off between the steady-state network mean-square-deviation and an inference privacy constraint. We show that the proposed noise powers are bounded and convergent, which leads to mean-square convergence of the proposed privacy-preserving multitask diffusion scheme. Simulation results demonstrate that the proposed strategy is able to balance the trade-off between estimation accuracy and privacy preservation.

We equip an ultra-wideband (UWB) node and a 2D LiDAR sensor a.k.a. 2D laser rangefinder on a mobile robot, and place UWB beacon nodes at unknown locations in an unknown environment. All UWB nodes can do ranging with each other thus forming a cooperative sensor network. We propose to fuse the peer-to-peer ranges measured between UWB nodes and laser scanning information, i.e., range measured between robot and nearby objects/obstacles, for simultaneous localization of the robot, all UWB beacons and LiDAR mapping. The fusion is inspired by two facts: 1) LiDAR may improve UWB-only localization accuracy as it gives a more precise and comprehensive picture of the surrounding environment; 2) on the other hand, UWB ranging measurements may remove the error accumulated in the LiDAR-based SLAM algorithm. Our experiments demonstrate that UWB/LiDAR fusion enables drift-free SLAM in real-time based on ranging measurements only.

In Internet of Things, with large amounts of sensor data gathered in fusion center, it is important to detect a public hypothesis, but at the same time it is crucial to prevent a private hypothesis being detected. In order to achieve this goal, a multilayer nonlinear processing procedure is proposed to distort the sensor's data before it is sent to the fusion center. In particular, each sensor applies linear and nonlinear distortions to balance the public hypothesis test and the privacy distortion. Mirror descent methodology is reformulated to optimize the distortion matrices so as to ensure that the regularized empirical risk of detecting the private hypothesis is above a given privacy threshold, while minimizing the regularized empirical risk of detecting the public hypothesis. Experiments on empirical datasets demonstrate that the proposed approach achieves a good tradeoff between the error rates of the public and private hypotheses.

Social sensing relies on a large number of observations reported by different, possibly unreliable, agents to determine if an event has occurred or not. In this paper, we consider the truth discovery problem in social sensing, in which an agent may receive another agent’s observation (known as an information flow), and may change its observation to match the observation it receives. If an agent’s observation is influenced by another agent, we say that the former is a dependent agent. We propose an Iterative Expectation Maximization algorithm for Truth Discovery (IEMTD) in social sensing with dependent agents. Compared with other popular truth discovery approaches, which assume either the agents’ observations are independent, or their dependency is known a priori, IEMTD allows to infer each agent’s reliability, the observations’ dependency and the events’ truth jointly. Simulation results on synthetic data and three real world data sets demonstrate that in almost all our experiments, IEMTD achieves a higher truth discovery accuracy than the existing algorithms when dependencies exist between agents’ observations.

Vehicle self-positioning is of significant importance for intelligent transportation applications. However, accurate positioning (e.g., with lane-level accuracy) is very difficult to obtain due to the lack of measurements with high confidence, especially in an environment without full access to a global navigation satellite system (GNSS). In this paper, a novel information fusion algorithm based on a particle filter is proposed to achieve lane-level tracking accuracy under a GNSS-denied environment. We consider the use of both coarse-scale and fine-scale signal measurements for positioning. Time-of-arrival measurements using the radio frequency signals from known transmitters or roadside units, and acceleration or gyroscope measurements from an inertial measurement unit (IMU) allow us to form a coarse estimate of the vehicle position using an extended Kalman filter. Subsequently, fine-scale measurements, including lane-change detection, radar ranging from the known obstacles (e.g., guardrails), and information from a high-resolution digital map, are incorporated to refine the position estimates. A probabilistic model is introduced to characterize the lane changing behaviors, and a multi-hypothesis model is formulated for the radar range measurements to robustly weigh the particles and refine the tracking results. Moreover, a decision fusion mechanism is proposed to achieve a higher reliability in the lane-change detection as compared to each individual detector using IMU and visual (if available) information. The posterior Cramér-Rao lower bound is also derived to provide a theoretical performance guideline. The performance of the proposed tracking framework is verified by simulations and real measured IMU data in a four-lane highway.

We consider a crowdsourcing platform where workers' responses to questions posed by a crowdsourcer are used to determine the hidden state of a multi-class labeling problem. As workers may be unreliable, we propose to perform sequential questioning in which the questions posed to the workers are designed based on previous questions and answers. We propose a Partially-Observable Markov Decision Process (POMDP) framework to determine the best questioning strategy, subject to the crowdsourcer's budget constraint. As this POMDP formulation is in general intractable, we develop a suboptimal approach based on a q-ary Ulam-Renyi game. We also propose a sampling heuristic, which can be used in tandem with standard POMDP solvers, using our Ulam-Renyi strategy. We demonstrate through simulations that our approaches outperform a non-sequential strategy based on error correction coding and which does not utilize workers' previous responses.

Insider attacks can result in significant costs to an organization. There is an urgent need for an automatic insider threat detector with good accuracy and low false alarms. In this work, we propose a graph based insider threat detector to identify potential insider attackers based on identifying not only self-anomalous behaviors of an employee but also anomalies relative to other employees with similar job roles. A machine learning approach is developed to first infer the correlation graph among the organization's employees. Then, a graph signal processing method is designed to identify the potential insiders with detection and false positive rates better than performing detection independently on each employee. Our approach demonstrates that the correlated behaviors of an organization's employees should be exploited for a better detection of suspicious behaviors.

We consider the problem of quickest change detection (QCD) for a signal which may undergo both a nuisance and a critical change. Our goal is to detect the critical change without raising a false alarm over the nuisance change. An optimal sequential change detection procedure is proposed for the Bayesian formulation of our QCD problem. A sequential change detection procedure based on the generalized likelihood ratio test (GLRT) statistic is also proposed for the non-Bayesian formulation. We show that our proposed test statistics can be computed efficiently via respective recursive update schemes. We compare our proposed stopping rules with the naive 2-stage procedures, which attempt to detect the changes using separate optimal stopping procedures (i.e., the Shiryaev procedure in the Bayesian formulation, and the CuSum procedure in the non-Bayesian formulation) for the nuisance and critical changes. Simulations demonstrate that our proposed rules outperform the 2-stage procedures.

We are concerned with a privacy-preserving problem in Kalman filter: a sensor releases a set of measurements to fusion center, who has perfect knowledge of the dynamical model, to allow it to estimate the public state, while prevent it from estimating the private state. We propose to linearly transform the original observation into a lower dimensional space before sending them to fusion center. Two privacy-utility tradeoffs are formulated: one concerns only at the current time step and the other concerns over two time steps. The transformation that leads to the optimal tradeoff can be found in closed-form. The privacy (estimation of private state) and utility (estimation of public state) are measured based on recursive Bayesian Cramér-Rao bound.

We consider the problem of blind calibration of a sensor network, where the sensor gains and offsets are estimated from noisy observations of unknown signals. This is in general a nonidentifiable problem, unless restrictive assumptions on the signal subspace or sensor observations are imposed. We show that if each signal observed by the sensors follows a known dynamic model with additive noise, then the sensor gains and offsets are identifiable. We propose a dynamic Bayesian nonparametric model to infer the sensors' gains and offsets. Our model allows different sensor clusters to observe different unknown signals, without knowing the sensor clusters a priori. We develop an offline algorithm using block Gibbs sampling and a linearized forward filtering backward sampling method that estimates the sensor clusters, gains, and offsets jointly. Furthermore, for practical implementation, we also propose an online inference algorithm based on particle filtering and local Markov chain Monte Carlo. Simulations using a synthetic dataset, and experiments on two real datasets suggest that our proposed methods perform better than several other blind calibration methods, including a sparse Bayesian learning approach, and methods that first cluster the sensor observations and then estimate the gains and offsets.

We study the problem of identifying infection sources in a network based on the network topology, and a subset of infection timestamps. In the case of a single infection source in a tree network, we derive the maximum likelihood estimator of the source and the unknown diffusion parameters. We then introduce a new heuristic involving an optimization over a parametrized family of Gromov matrices to develop a single source estimation algorithm for general graphs. Compared with the breadth-first search tree heuristic commonly adopted in the literature, simulations demonstrate that our approach achieves better estimation accuracy than several other benchmark algorithms, even though these require more information like the diffusion parameters. We next develop a multiple sources estimation algorithm for general graphs, which first partitions the graph into source candidate clusters, and then applies our single source estimation algorithm to each cluster. We show that if the graph is a tree, then each source candidate cluster contains at least one source. Simulations using synthetic and real networks, and experiments using real-world data suggest that our proposed algorithms are able to estimate the true infection source(s) to within a small number of hops with a small portion of the infection timestamps being observed.

We propose a non-Bayesian social learning update rule for agents in a network, which minimizes the sum of the Kullback-Leibler divergence between the true distribution generating the agents’ local observations and the agents’ beliefs (parameterized by a hypothesis set), and a weighted varentropy-related term. The varentropy-related term allows us to control the rate of convergence of our update rule, which also reuses some of the most recent observations of each agent to speed up convergence. Under mild technical conditions, we show that the belief of each agent concentrates on the optimal hypothesis set, and we derive a bound for the convergence rate. Furthermore, to overcome the performance degradation due to misinforming agents, who use a corrupted likelihood functions in their belief updates, we propose to use multiple social networks that update their beliefs independently and a convex combination mechanism among the beliefs of all the networks. Simulations with applications to location identification and group recommendation demonstrate that our proposed methods offer improvements over two other current state-of-the art non-Bayesian social learning algorithms.

A sensor network wishes to transmit information to a fusion center to allow it to detect a public hypothesis, but at the same time prevent it from inferring a private hypothesis. We propose a multilayer sensor network structure, where each sensor first applies a nonlinear fusion function on the information it receives from sensors in a previous layer, and then a linear weighting matrix to distort the information it sends to sensors in the next layer. We adopt a nonparametric approach and develop an algorithm to optimize the weighting matrices so as to ensure that the regularized empirical risk of detecting the private hypothesis is above a given privacy threshold, while minimizing the regularized empirical risk of detecting the public hypothesis. Simulations on a synthetic dataset and an empirical experiment demonstrate that our approach is able to achieve a better trade-off between the error rates of the public and private hypothesis than using only linear precoding to achieve information privacy.

This paper considers the problem of quickest detection of a change in distribution under the assumption that the pre-change distribution π is known, and the post-change distribution μ is unknown and belongs to a general class of distributions. Using the knowledge of the pre-change distribution π, the sample space is partitioned into equiprobable intervals and the number of samples falling into each of these intervals is monitored to detect the change. A test statistic that approximates the generalized likelihood ratio test is proposed. A recursive update scheme to compute the statistic efficiently and an approximation of the average run-length to false alarm are also derived. Simulations show that our approach is comparable in performance to two other non-parametric quickest change detection methods if the change is either a shift in distribution mean or variance, respectively. But our method significantly outperforms them if these distribution change assumptions are violated.

In this paper we study the problem of identifying an infection source in a network based only on the network topology and a stream of infection timestamps. We propose a sequential source estimation algorithm (SSE) using a particle filter that is based on an approximate hidden Markov chain model, which can be interpreted as a “reverse” propagation process. Simulations using synthetic networks and experiments using real-world social network data suggest that SSE is able to estimate the true infection source to within a small number of hops with less than 20% of the infection timestamps being observed.

In the sensor network blind calibration problem, the gains and offsets of sensors are estimated from noisy observations of unknown underlying signals. This is in general a non-identifiable problem, unless restrictive assumptions on the signal subspace or sensor observations are imposed. To overcome these assumptions, we propose a dynamic Bayesian nonparametric model. We show that if the unknown underlying signals follow the first-order auto-regressive process, then the sensor gains and offsets are identifiable. Furthermore, our model allows sensors to form clusters, where each cluster observes the same underlying signal. The clusters are however not known a priori, and are learned through the sensor data. We present a block Gibbs sampling inference method based on the forward filtering backward sampling algorithm. Simulation results suggest that our approach can estimate the sensor gains and offsets with good accuracy, and performs better than methods that first perform clustering and then blind calibration.

A recently proposed non-Bayesian social learning rule by [1] finds an update rule that achieves the least Kullback-Leibler (LKL) divergence between a proxy of the true distribution generating agent's local observations and the agent's beliefs. In this paper, we investigate improvements to the LKL social learning algorithm by reusing the M most recent observations of each agent to perform updates of each agent's local belief. We call this the observation reusing LKL (OR-LKL) social learning algorithm. Under some technical conditions, we show that under OR-LKL updates, every agent's belief concentrates on an optimal hypothesis set. Furthermore, in order to achieve robust performance when there are misinforming agents in the network, we propose to use two social networks that update their beliefs independently using OR-LKL, and a soft switching mechanism between the beliefs of the two networks. Simulation results demonstrate that the proposed OR-LKL and robust OR-LKL algorithms offer significant improvements over two other current state-of-the art non-Bayesian social learning algorithms.

This paper considers the problem of range-based decentralized localization in wireless sensor networks when the impulsive measurement noise is present. We develop a robust localization estimator requiring no a priori knowledge of the noise distribution. The approach to robust localization presented here follows the concept of M-estimation and is implemented in a decentralized manner thus suiting the dynamic nature of wireless networks. The performance of the proposed scheme is verified by simulations.

This paper considers the problem of decentralized, cooperative, and dynamic self-localization in wireless sensor networks. In particular, we are interested in a restrictive but very realistic scenario where few anchors are deployed and each anchor whose location is priori known may only communicate with very few agents (e.g. just one agent) whose location is unknown and to-be-estimated. The lack of agent-to-anchor communication links renders slow estimation convergence thereby demanding more message exchanges among the nodes i.e. agent-to-agent and agent-to-anchor. This urges us to propose an efficient localization method that needs less iterations (i.e. less message exchanges) to achieve a certain accuracy.

We study the problem of inferring the graph structure of a network using knowledge of information cascades in the network. Unlike previous studies, which assume knowledge of the distributions of information diffusion across edges in the network, we only require that diffusion along different edges in the network be independent together with limited information on their distributions (e.g., just the means). We introduce the concept of a separating vertex set for a graph, which is a set of vertices in which for any two given distinct vertices of the graph, one can find a vertex whose distance to them are different. We show that a necessary condition for reconstructing a tree perfectly using distance information between pairs of vertices is given by the size of an observed separating vertex set. We then propose an algorithm to recover the tree structure using infection times, whose differences have means corresponding to the distance between two vertices. To improve the accuracy, we propose the concept of redundant vertices, which allows us to perform averaging to better estimate the distance between two vertices. Though the theory is developed mainly for trees, we demonstrate how the algorithm can be extended heuristically to general graphs. Simulation results suggest that our proposed algorithm performs better than some current state-of-the-art network reconstruction methods.

We consider a crowdsourcing platform where workers are posed questions by a crowdsourcer, who then uses their responses to determine the hidden state of a multi-class labeling problem. Workers may be unreliable, therefore by designing the questions using error correction coding approaches, the crowdsourcer can achieve a more reliable overall result. We propose to perform sequential questioning in which workers are asked q-ary questions sequentially, and questions are determined based on the workers' previous responses. We propose an optimization framework to determine the best q and questioning strategy to use, subject to a crowdsourcer budget constraint. For a fixed q, this problem is equivalent to finding an optimal questioning strategy to a q-ary Ulam-Rényi game, which is in general intractable. We propose a heuristic to find a suboptimal strategy, and demonstrate through simulations that our solution outperforms another error correction coding strategy that does not utilize previous workers' responses. Simulations also suggest that q can in general be chosen to be much smaller than the number of classes in the multi-class labeling problem.

Finding the infection sources in a network when we only know the network topology and infected nodes, but not the rates of infection, is a challenging combinatorial problem, and it is even more difficult in practice where the underlying infection spreading model is usually unknown a priori. In this paper, we are interested in finding a source estimator that is applicable to various spreading models, including the susceptible-infected (SI), susceptible-infected-recovered (SIR), susceptible-infected-recovered-infected (SIRI), and susceptible- infected-susceptible (SIS) models. We show that under the SI, SIR, and SIRI spreading models and with mild technical assumptions, the Jordan center is the infection source associated with the most likely infection path in a tree network with a single infection source. This conclusion applies for a wide range of spreading parameters, while it holds for regular trees under the SIS model with homogeneous infection and recovery rates. Since the Jordan center does not depend on the infection, recovery, and reinfection rates, it can be regarded as a universal source estimator. We also consider the case where there are k 1 infection sources, generalize the Jordan center definition to a k-Jordan center set, and show that this is an optimal infection source set estimator in a tree network for the SI model. Simulation results on various general synthetic networks and real-world networks suggest that Jordan center-based estimators consistently outperform the betweenness, closeness, distance, degree, eigenvector, and pagerank centrality-based heuristics, even if the network is not a tree.

Cloud radio access network (C-RAN) is emerging as a potential alternative for the next generation RAN by merging RAN and cloud computing together. In this paper, we consider the baseband unit (BBU) pool of C-RAN as a collection of virtual machines (VMs). We allow each user equipment (UE) to associate with multiple VMs in the BBU pool, and each remote radio head (RRH) can only serve a limited number of UEs. Under this model, we jointly optimize the VM activation in the BBU pool and sparse beamforming in the coordinated RRH cluster, which is constrained by limited fronthaul capacity, to minimize the system cost of C-RAN. We formulate this problem as a mixed-integer nonlinear programming problem, and then propose efficient methods to optimize the number of active VMs, as well as the sparse beamforming vectors. Moreover, we derive a closed-form solution for the beamforming vectors. Simulation results suggest that our proposed algorithms have better performance than the benchmark algorithms in terms of both system cost and robustness.

We consider the problem of decentralized detection of a hypothesis H using multiple sensors. The sensors also want to keep the fusion center from inferring about another hypothesis G. Each sensor makes an observation and summarizes the observation using a local decision rule. The sensor summaries are communicated to the fusion center to perform an overall decision making. As the underlying joint distribution of the hypotheses and sensor observations is unknown, we aim at finding sensor decision rules that minimize the regularized empirical risk of deciding H at the fusion center, while ensuring that the regularized risk of the fusion center deciding G correctly is more than a given threshold. We propose an optimization approach based on the Gauss-Seidel method, and show that it converges to a critical point.

We formulate and study a multi-user multi-armed bandit (MAB) problem that exploits the temporal-spatial reuse of primary user (PU) channels so that secondary users (SUs) who do not interfere with each other can make use of the same PU channel. We first propose a centralized channel allocation policy that has logarithmic regret, but requires a central processor to solve a NP-complete optimization problem at exponentially increasing time intervals. To avoid the high computation complexity at the central processor and the need for SU synchronization, we propose a heuristic distributed policy that incorporates channel access rank learning in a local procedure at each SU at the cost of a higher regret. We compare the performance of our proposed policies with other distributed policies recently proposed for opportunistic spectrum access. Simulations suggest that our proposed policies significantly outperform the benchmark algorithms when spectrum temporal-spatial reuse is allowed.

We propose a method for tracking a moving receiver using RF messages broadcasted from opportunistic beacons. The beacons are not required to be synchronized with each other or with the receiver. By measuring the time-of- arrival (TOA) of the RF messages transmitted from the beacons at different locations along the receiver's trajectory, the receiver is able to track its own location, velocity and local oscillator (LO) parameters with respect to each beacon. A major challenge lies in separating the time offset due to the movement of the receiver from the time offset due to local LO drifts. We propose a Kalman filtering framework that is able to track the LO drifts, the receiver location and the receiver velocity simultaneously, using the measured TOAs and the beacons' locations as inputs. Simulation results are presented to show the feasibility of the proposed method. Finally, the proposed method is implemented on a software-defined radio (SDR) testbed. Experiment results demonstrate that our approach can successfully track the receiver trajectories with good accuracy.

We consider the problem of angle-of-arrival (AoA) estimation of a RF transmitter using a mobile receiver. We develop a method very similar to synthetic aperture radar which is compatible with cellular technology. By considering the successive packets received along the receiver trajectory, we implicitly create a virtual MIMO array, which allows us to utilize conventional MIMO theory for AoA estimation. For this method to work, the first major challenge is the need to separate the phase offset due to receiver movement from the phase offset due to local oscillator (LO) offset. Two approaches are proposed to do this: i) a stop-and-start approach, where the receiver first stands still long enough to estimate the LO offset and then estimates the AoA while moving, and ii) a joint nonlinear estimator where the AoA and LO offset are estimated simultaneously. The second major difficulty is the need to estimate the receiver's relative position with sub-wavelength accuracy. We solve this by using a three-dimensional inertial measurement unit, which provides reasonably good relative position estimates if the measurement period is sufficiently short. Simulations and experimental results based on a software-defined radio platform in an anechoic chamber show the feasibility of the proposed method.

We formulate and study a multi-user multi-armed bandit problem that exploits the temporal–spatial opportunistic spectrum access (OSA) of primary user (PU) channels, so that secondary users (SUs) who do not interfere with each other can make use of the same PU channel. We first propose a centralized channel allocation policy that has logarithmic regret, but requires a central processor to solve an NP-complete optimization problem at exponentially increasing time intervals. To overcome the high computation complexity at the central processor, we also propose heuristic distributed policies that, however, have linear regrets. Our first distributed policy utilizes a distributed graph coloring and consensus algorithm to determine SUs’ channel access ranks, while our second distributed policy incorporates channel access rank learning in a local procedure at each SU at the cost of a higher regret. We compare the performance of our proposed policies with other distributed policies recently proposed for temporal (but not spatial) OSA. We show that all these policies have linear regrets in our temporal–spatial OSA framework. Simulations suggest that our proposed policies have significantly smaller regrets than the other policies when spectrum temporal–spatial reuse is allowed.

We consider the problem of tracking a receiver using signals-of-opportunity (SOOPs) from beacons and a reference anchor with known positions and velocities, and where all devices have asynchronous local clocks or oscillators. We model the clock drift at individual devices by a two-state model with unknown clock offset and clock skew and analyze the biases introduced by clock asynchronism in the received signals. Based on an extended Kalman filter, we propose a sequential estimator to jointly track the receiver location, velocity, and its clock parameters using altitude information together with time-difference-of-arrival and frequency-difference-of-arrival measurements obtained from the SOOP samples collected by the receiver and a reference anchor. The receiver was implemented on a software-defined radio testbed, and field experiments are carried out using Iridium satellites as the SOOP beacons. The experiment and simulation results demonstrate that our measurement model has a good fit, and our proposed estimator can successfully track both the receiver location, velocity, and the relative clock offset and skew with respect to the reference anchor with good accuracy.

The goal of a rumor source node in a social network is to spread its rumor to as many nodes as possible, while remaining hidden from the network administrator. On the other hand, the network administrator aims to identify the source node based on knowledge of which nodes have accepted the rumor (which are called infected nodes). We model the rumor spreading and source identification problem as a strategic game, where the rumor source and the network administrator are the two players. As the Jordan center estimator is a minimax source estimator that has been shown to be robust in recent works, we assume that the network administrator utilizes a source estimation strategy that probes every node within a given radius of the Jordan center. Given any estimation strategy, we design a best-response infection strategy for the rumor source. Given any infection strategy, we design a best-response estimation strategy for the network administrator. We derive conditions under which the Nash equilibria of the strategic game exist. Simulations in both synthetic and real-world networks demonstrate that our proposed infection strategy infects more nodes while maintaining the same safety margin between the true source node and the Jordan center source estimator.

We propose an alternating optimization algorithm for localizing a mobile non-cooperative target using a wireless sensor network. We consider the scenario where sensors receive single-bounce non-line-of-sight signals from the moving target. Each sensor is able to measure the target signal's angle-of-arrival and received signal strength. The transmit powers of the non-cooperative target at different locations are unknown, and estimated jointly with its locations and the orientations of the scatterers off which the target signals are reflected before reaching the sensors. We formulate the problem as a non-convex least squares problem, and then transform and approximate it into a form that is solvable by an alternating algorithm. We show that our algorithm converges, and simulation results demonstrate that our algorithm is able to localize the target with good accuracy.

We study the problem of identifying a single infection source in a network under the susceptible-infected-recovered-infected (SIRI) model. We describe the infection model via a state-space model, and utilizing a state propagation approach, we derive an algorithm known as the heterogeneous infection spreading source (HISS) estimator, to infer the infection source. The HISS estimator uses the observations of node states at a particular time, where the elapsed time from the start of the infection is unknown. It is able to incorporate side information (if any) of the observed states of a subset of nodes at different times, and of the prior probability of each infected or recovered node to be the infection source. Simulation results suggest that the HISS estimator outperforms the dynamic message passing and Jordan center estimators over a wide range of infection and reinfection rates.

In this paper we propose a method to localize a periodic RF transmitter using a single mobile receiver. The receiver measures the time-of-arrival (TOA) of the periodic messages at different locations along its trajectory. By comparing the TOA of successive messages at different points along its trajectory, the receiver can eventually estimate the transmitter location. The challenge lies in separating the time offset due to receiver movement from that caused by local oscillator (LO) drift. We propose an extended Kalman filter framework that estimates the LO drift and the transmitter location simultaneously, using the TOA measurements and the receiver location as inputs. The proposed algorithm is implemented and tested on a software-defined radio testbed, and experimental results demonstrate that the proposed method is able to simultaneously locate the transmitter and estimate the LO drift with good accuracy.

We propose a novel technique to localize a periodic transmitter using a single mobile receiver. By having the receiver measure the time-of-arrival (TOA) of the periodic messages sent by the transmitter, the receiver is able to measure the virtual time-difference-of-arrival (V-TDOA) between different locations along its trajectory. The V-TDOA can then be used in place of TDOA measurements in various TDOA-based localization methods for the transmitter. However, a major challenge in obtaining an accurate V-TDOA measurement lies in separating it from the time offset between the transmitter and the receiver caused by the local oscillator (LO) drift. We propose a two-stage technique: in a first stage, the receiver stands still and uses the TOA measurements to train a Kalman filter to estimate and predict the LO skew between itself and the transmitter. In a second stage, the receiver starts moving and uses its prediction of the LO skew to estimate the V-TDOA. The proposed algorithm is implemented on a software-defined radio testbed, and experimental results show that our setup is able to successfully measure the V-TDOA.

We propose a novel distributed expectation maximization (EM) method for non-cooperative RF target localization using a wireless sensor network. We consider the scenario where few or no sensors receive line-of-sight signals from the target. In the case of non-line-of-sight signals, the signal path consists of a single reflection between the transmitter and receiver. Each sensor is able to measure the time difference of arrival of the target's signal with respect to a reference sensor, as well as the angle of arrival of the target's signal. We derive a distributed EM algorithm where each node makes use of its local information to compute summary statistics, and then shares these statistics with its neighbors to improve its estimate of the target localization. We show that our distributed algorithm converges, and simulation results suggest that our method achieves an accuracy close to the centralized EM algorithm. We apply the distributed EM algorithm to a set of experimental measurements with a network of four nodes, which confirm that the algorithm is able to localize a RF target in a realistic non-line-of-sight scenario.

Cloud radio access network (C-RAN) aims to improve spectrum and energy efficiency of wireless networks by migrating conventional distributed base station functionalities into a centralized cloud baseband unit (BBU) pool. We propose and investigate a cross-layer resource allocation model for C-RAN to minimize the overall system power consumption in the BBU pool, fiber links and the remote radio heads (RRHs). We characterize the cross-layer resource allocation problem as a mixed-integer nonlinear programming (MINLP), which jointly considers elastic service scaling, RRH selection, and joint beamforming. The MINLP is however a combinatorial optimization problem and NP-hard. We relax the original MINLP problem into an extended sum-utility maximization (ESUM) problem, and propose two different solution approaches. We also propose a low-complexity Shaping-and-Pruning (SP) algorithm to obtain a sparse solution for the active RRH set. Simulation results suggest that the average sparsity of the solution given by our SP algorithm is close to that obtained by a recently proposed greedy selection algorithm, which has higher computational complexity. Furthermore, our proposed cross-layer resource allocation is more energy efficient than the greedy selection and successive selection algorithms.

We consider a binary hypothesis testing problem in a tandem network where the distribution of the agent observations under each hypothesis comes from an uncertainty class. When agents know their positions in the tandem, and the contamination of the uncertainty classes are non-zero, we show that asymptotic learning of the true hypothesis under social learning is not possible even when the log likelihood ratio of the nominal distributions of the uncertainty classes is unbounded. Furthermore, asymptotic learning in social learning is achievable if and only if the uncertainty classes contamination converge to zero. When agents do not know their positions, the minimax error probability is bounded from zero, and we provide tight bounds for it.

In a cognitive radio network, a primary user (PU) shares its spectrum with secondary users (SUs) temporally and spatially, while allowing for some interference. We consider the problem of estimating the no-talk region of the PU, i.e., the region outside which SUs may utilize the PU's spectrum regardless of whether the PU is transmitting or not. We propose a distributed boundary estimation algorithm that allows SUs to estimate the boundary of the no-talk region collaboratively through message passing between SUs, and analyze the trade-offs between estimation error, communication cost, setup complexity, throughput and robustness. Simulations suggest that our proposed scheme has better estimation performance and communication cost trade-off compared to several other alternative benchmark methods, and is more robust to SU sensing errors, except when compared to the least squares support vector machine approach, which however incurs a much higher communication cost.

We consider the problem of optimally redirecting user requests in a cloud-centric media network (CCMN) to multiple destination Virtual Machines (VMs), which elastically scale their service capacities in order to minimize a cost function that includes service response times, computing costs, and routing costs. We also allow the request arrival process to switch between normal and flash crowd modes to model user requests to a CCMN. We quantify the trade-offs in flash crowd detection delay and false alarm frequency, request allocation rates, and service capacities at the VMs. We show that under each request arrival mode (normal or flash crowd), the optimal redirection policy can be found in terms of a price for each VM, which is a function of the VM's service cost, with requests redirected to VMs in order of nondecreasing prices, and no redirection to VMs with prices above a threshold price. Applying our proposed strategy to a YouTube request trace data set shows that our strategy outperforms various benchmark strategies. We also present simulation results when various arrival traffic characteristics are varied, which again suggest that our proposed strategy performs well under these conditions.

We consider the problem of localizing two devices using signals of opportunity from beacons with known positions. Beacons and devices have asynchronous local clocks or oscillators with unknown clock skews and offsets. We model clock skews as random, and analyze the biases introduced by clock asynchronism in the received signals. By deriving the equivalent Fisher information matrix for the modified Bayesian Cramér-Rao lower bound (CRLB) of device position and velocity estimation, we quantify the errors caused by clock asynchronism. We propose an algorithm based on differential time-difference-of-arrival (DTDOA) and frequency-difference-of-arrival (FDOA) that mitigates the effects of clock asynchronism to estimate the device positions and velocities. Simulation results suggest that our proposed algorithm is robust and approaches the CRLB when clock skews have small standard deviations.

We consider the problem of identifying an infection source based only on an observed set of infected nodes in a network, assuming that the infection process follows a Susceptible-Infected-Susceptible (SIS) model. We derive an estimator based on estimating the most likely infection source associated with the most likely infection path. Simulation results on regular trees suggest that our estimator performs consistently better than the minimum distance centrality based heuristic.

We consider the problem of localizing two sensors using signals-of-opportunity from beacons with known positions. Beacons and sensors have asynchronous local clocks or oscillators with unknown clock skews and offsets. We analyze the biases introduced by clock asynchronism in the received signals, and derive the Cramér-Rao lower bound (CRLB) for sensor position and velocity estimation errors. We quantify the error caused by beacon clock asynchronism. We also propose an algorithm that mitigates the effects of clock asynchronism to estimate the sensor positions and velocities. Simulation results suggest that our proposed algorithm is robust and approaches the CRLB when sensor clock skews have small standard deviations.

In a cognitive radio network, a primary user (PU) shares its spectrum with secondary users (SUs) temporally and spatially, while allowing for some interference. We consider the problem of estimating the interference coverage region of the PU, i.e., the region outside of which SUs may utilize the PU's spectrum regardless of whether the PU is transmitting or not. We propose a distributed boundary estimation algorithm that allows SUs to estimate the boundary of the coverage region collaboratively through message passing between SUs. We also propose a spatial spectrum sensing scheme based on the estimated boundary. Simulation results suggest that our proposed scheme has better estimation performance and communication cost trade-offs compared to centralized boundary estimation methods, and has better weighted throughputs than traditional fusion center based collaborative spectrum sensing methods.

We consider randomized broadcast or information dissemination in wireless networks with switching network topologies. We show that an upper bound for the$\epsilon$-dissemination time consists of the conductance bound for a network without switching, and an adjustment that accounts for the number of informed nodes in each period between topology changes. Through numerical simulations, we show that our bound is asymptotically tight. We apply our results to the case of mobile wireless networks with unreliable communication links and establish an upper bound for the dissemination time when the network undergoes topology changes and periods of communication link erasures.

The need for bandwidth and the incitation to reduce power consumption lead to the reduction of cell size in wireless networks. This allows reducing the distance between a user and the base station, thus increasing the capacity. A relatively inexpensive way of deploying small-cell networks is to use femtocells. However, the reduction in cell size causes problems for coordination and network deployment, especially due to the intra- and cross-tier interference. In this paper, we consider a two-tier multiple-input multiple-output (MIMO) network in the downlink, where a single macrocell base station with multiple transmit antennas coexists with multiple closed-access MIMO femtocells. With multiple receive antennas at both the macrocell and femtocell users, we propose an opportunistic interference alignment scheme to design the transmit and receive beamformers in order to mitigate intra- (or inter-) and cross-tier interference. Moreover, to reduce the number of macrocell and femtocell users coexisting in the same spectrum, we apply a random spectrum allocation on top of the opportunistic interference alignment. Using stochastic geometry, we analyze the proposed scheme in terms of the distribution of a received signal-to-interference-plus-noise ratio, spatial average capacity, network throughput, and energy efficiency. In the presence of imperfect channel state information, we further quantify the performance loss in spatial average capacity. Numerical results show the effectiveness of our proposed scheme in improving the performance of random MIMO femtocell networks.

We consider the problem of sensor localization in wireless networks in a multipath environment. We propose a distributed and cooperative algorithm based on belief propagation, which allows sensors to cooperatively self-localize with respect to a single anchor node in the network, using range and direction of arrival measurements. In the algorithm, neighboring sensors exchange limited information to update their local mean location estimates and covariance matrices. We show that the covariance matrix for each sensor converges for connected networks, and its mean location estimate converges if all scatters are either parallel or orthogonal to each other. Furthermore, these estimates are asymptotically unbiased. Simulations show that cooperation amongst neighboring nodes significantly improves the localization accuracy.

We consider the problem of intrusion detection in a computer network, and investigate the use of extreme learning machines (ELMs) to classify and detect the intrusions. With increasing connectivity between networks, the risk of information systems to external attacks or intrusions has increased tremendously. Machine learning methods like support vector machines (SVMs) and neural networks have been widely used for intrusion detection. These methods generally suffer from long training times, require parameter tuning, or do not perform well in multi-class classification. We propose a basic ELM method based on random features, and a kernel based ELM method for classification. We compare our methods with commonly used SVM techniques in both binary and multi-class classifications. Simulation results show that the proposed basic ELM approach outperforms SVM in training and testing speed, while the proposed kernel based ELM achieves higher detection accuracy than SVM in multi-class classification case.

We consider the problem of localizing two sensors using their TDOA measurements of signals from beacons with known positions. Beacons and sensors have asynchronous local clocks or oscillators with unknown clock skews and offsets. We analyze the effect of the clock skews and offsets on the TDOA estimations, and introduce a novel approach using the difference in TDOA (DTDOA) measurements to mitigate the effects of clock skews and offsets on the location estimates. We propose an algorithm for location and velocity estimation based on DTDOA, and perform simulations to verify the robustness of our algorithm to asynchronous local clocks in the beacons and sensors.

Estimating which nodes in a network are the infection sources, including the individuals who started a rumor in a social network, the computers that introduce a virus into a computer network, or the index cases of a contagious disease, plays a critical role in identifying the influential nodes in a network, and in some applications, limiting the damage caused by the infection through timely quarantine of the sources. We consider the problem of estimating the infection sources, based only on knowledge of the underlying network connections. We derive estimators based on approximations of the infection sequences counts. We show that if there are at most two infection sources in a geometric tree, our estimator identifies these sources with probability going to one as the number of infected nodes increases. When there are more than two infection sources, we present heuristics that have quadratic complexity. We show through simulations that our proposed estimators can correctly identify the infection sources to within a few hops with high probability.

We consider the decentralized binary hypothesis testing problem in networks with feedback, where some or all of the sensors have access to compressed summaries of other sensors' observations. We study certain two-message feedback architectures, in which every sensor sends two messages to a fusion center, with the second message based on full or partial knowledge of the first messages of the other sensors. We also study one-message feedback architectures, in which each sensor sends one message to a fusion center, with a group of sensors having full or partial knowledge of the messages from the sensors not in that group. Under either a Neyman-Pearson or a Bayesian formulation, we show that the asymptotically optimal (in the limit of a large number of sensors) detection performance (as quantified by error exponents) does not benefit from the feedback messages, if the fusion center remembers all sensor messages. However, feedback can improve the Bayesian detection performance in the one-message feedback architecture if the fusion center has limited memory; for that case, we determine the corresponding optimal error exponents.

We consider the decentralized Bayesian binary hypothesis testing problem in feedback architectures, in which the fusion center broadcasts information based on the messages of some sensors to some or all sensors in the network. We show that the asymptotically optimal detection performance (as quantified by error exponents) does not benefit from the feedback messages. In addition, we determine the corresponding optimal error exponents.

We consider the problem of decentralized detection in a network consisting of a large number of nodes arranged as a tree of bounded height, under the assumption of conditionally independent and identically distributed (i.i.d.) observations. We characterize the optimal error exponent under a Neyman-Pearson formulation. We show that the Type II error probability decays exponentially fast with the number of nodes, and the optimal error exponent is often the same as that corresponding to a parallel configuration. We provide sufficient, as well as necessary, conditions for this to happen. For those networks satisfying the sufficient conditions, we propose a simple strategy that nearly achieves the optimal error exponent, and in which all non-leaf nodes need only send 1-bit messages.

We consider the problem of Bayesian decentralized binary hypothesis testing in a network of sensors arranged in a tandem. We show that the rate of error probability decay is always subexponential, establishing the validity of a long-standing conjecture. Under the additional assumption of bounded Kullback-Leibler (KL) divergences, we show that for alld> 1/2, the error probability is Omega(e-cnd), wherecis a positive constant. Furthermore, the bound Omega(e-c(logn)d) , for alld> 1, holds under an additional mild condition on the distributions. This latter bound is shown to be tight.

We study the asymptotic detection performance of large sensor networks, configured as trees with bounded height, in which information is progressively compressed as it moves towards the root of the tree. We show that the error probability decays exponentially fast, and we provide bounds for the error exponent. We analyze further the case where the tree has certain symmetry properties, and derive simple, easily implementable, suboptimal strategies

We consider the problem of decentralized binary hypothesis testing in a network of sensors arranged in a tandem. We show that the rate of error probability decay is always sub-exponential, establishing the validity of a long-standing conjecture. Under the additional assumption of bounded Kullback-Leibler divergences, we show that for all d > 1/2, the error probability is Omega(e-cnd), where c is a positive constant. Furthermore, the bound Omega(e-c(logn)d), for all d > 1, holds under an additional mild condition on the distributions. This latter bound is shown to be tight.

We study decentralized detection in tree networks with bounded height, and in which there are either sensor failures or unreliable communications between sensors. We characterize the asymptotically optimal performance of such tree networks, when certain parameters are allowed to become large, to model the case of dense sensor networks. We also develop simple strategies that nearly achieve the optimal performance.

We consider the problem of decentralized binary detection in a sensor network where the sensors have access to side information that affects the statistics of their measurements, or reflects the quality of the available channel to a fusion center. Sensors can decide whether or not to make a measurement and transmit a message to the fusion center ("censoring"), and also have a choice of the mapping from measurements to messages. We consider the case of a large number of sensors, and an asymptotic criterion involving error exponents. We study both a Neyman-Pearson and a , Bayesian formulation, characterize the optimal error exponent, and derive asymptotically optimal strategies for the case where sensor decisions are only allowed to depend on locally available information. Furthermore, we show that for the Neyman-Pearson case, global sharing of side information ("sensor cooperation") does not improve asymptotic performance, when the Type I error is constrained to be small.

We consider the problem of Bayesian decentralized binary detection in a sensor network in which the sensors have access to some side information that affects the statistics of the measurements they make. Sensors can decide whether or not to make a measurement and transmit a message to the fusion center ("censoring"), and also have a choice of the transmission function from measurements to messages. We consider the case of a large number of sensors, characterize the optimal error exponent, and derive asymptotically optimal strategies. We show that the optimal strategy consists of dividing the sensors into two groups, with sensors in each group using the same policy