Multiple Edges Research Articles

Towards Collaborative Edge Intelligence: Blockchain-Based Data Valuation and Scheduling for Improved Quality of Service

Collaborative edge intelligence, a distributed computing paradigm, refers to a system where multiple edge devices work together to process data and perform distributed machine learning (DML) tasks locally. Decentralized Internet of Things (IoT) devices share knowledge and resources to improve the quality of service (QoS) of the system with reduced reliance on centralized cloud infrastructure. However, the paradigm is vulnerable to free-riding attacks, where some devices benefit from the collective intelligence without contributing their fair share, potentially disincentivizing collaboration and undermining the system’s effectiveness. Moreover, data collected from heterogeneous IoT devices may contain biased information that decreases the prediction accuracy of DML models. To address these challenges, we propose a novel incentive mechanism that relies on time-dependent blockchain records and multi-access edge computing (MEC). We formulate the QoS problem as an unbounded multiple knapsack problem at the network edge. Furthermore, a decentralized valuation protocol is introduced atop blockchain to incentivize contributors and disincentivize free-riders. To improve model prediction accuracy within latency requirements, a data scheduling algorithm is given based on a curriculum learning framework. Based on our computer simulations using heterogeneous datasets, we identify two critical factors for enhancing the QoS in collaborative edge intelligence systems: (1) mitigating the impact of information loss and free-riders via decentralized data valuation and (2) optimizing the marginal utility of individual data samples by adaptive data scheduling.

Energy-efficiency optimization for heterogeneous computing-assisted NOMA-MEC edge AI tasks

Edge artificial intelligence (AI) is an emerging paradigm that leverages edge computing to pave the last-mile delivery of AI. To satisfy the increasing demand for high-performance computing and low latency of edge service, heterogeneous computing accelerators, especially Neural Processor Units (NPUs), are widely deployed on edge nodes. However, the edge AI heterogeneous computing system encounters the challenge of energy constraints. In this paper, we investigate the energy efficiency (EE) optimization problem in heterogeneous computing-assisted non-orthogonal multiple access mobile edge computing (NOMA-MEC) network model. Edge devices are configured with CPU-NPU heterogeneous computing processors to improve the AI task processing, and NOMA is applied to edge task offloading to enable massive connectivity. The system-centric energy efficiency maximization problem is studied. We formulate a system-centric energy efficiency maximization problem. To solve this problem, we decouple the original problem into two subproblems, i.e., the optimization problems of heterogeneous computing workload splitting and task offloading. A heuristic algorithm and binary search algorithm are proposed to optimize NOMA user pairing and task offloading delay, respectively. Furthermore, we propose a multi-objective alternative iterative algorithm to optimize system energy efficiency. Numerical results show that the proposed scheme can significantly improve the energy efficiency of heterogeneous computing-assisted NOMA-MEC networks compared to the existing baselines.

Deep reinforcement learning based task offloading and resource allocation strategy across multiple edge servers

Deep reinforcement learning based task offloading and resource allocation strategy across multiple edge servers

Research on airport baggage anomaly retention detection technology based on machine vision, edge computing, and blockchain

AbstractAirport checked luggage entails specific requirements for speed, stability, and reliability. The issue of abnormal retention of checked luggage presents a significant challenge to aviation safety and transportation efficiency. Traditional luggage monitoring systems exhibit limitations in terms of accuracy and timeliness. To address this challenge, this paper proposes a real‐time detection and alerting of luggage anomaly retention based on the YOLOv5 object detection model, leveraging visual algorithms. By eliminating cloud servers and deploying multiple edge servers to establish a private chain, images of anomalously retained luggage are encrypted and stored on the chain. Data users can verify the authenticity of accessed images through anti‐tampering algorithms, ensuring the security of data transmission and storage. The deployment of edge computing servers can significantly reduce algorithm latency and enhance real‐time performance. This solution employs computer vision technology and an edge computing framework to address the speed and stability of checked luggage transportation. Furthermore, blockchain technology greatly enhances system security during operation. A model trained on a sample set of 4600 images achieved a luggage recognition rate of 96.9% and an anomaly detection rate of 95.8% in simulated test videos.

Joint Optimization Algorithm for Latency-Energy and Privacy Combining Discrete Differential Evolution and Convex Optimization

A privacy protection task offloading algorithm combining discrete differential evolution and convex optimization is proposed to resolve the new position privacy leakage problem for mobile terminal due to transferring tasks to multiple edge servers for computation via wireless networks. First, the privacy cost of offloading decision is defined for the characteristics of the position privacy problem combined with the standard deviation theory. Second, the objective problem is modeled as a joint optimization of position privacy protection, offloading decision and transmission power, and the total cost of task offloading is an optimization objective. Finally, to handle the objective problem efficiently, offloading decision variable is decoupled from the transmission power variable to decrease the computational complexity of the problem; the discrete differential evolution algorithm is designed to accumulate the approximate optimal offloading decision, and the optimal transmission power is efficiently solved by variable substitution method and subgradient method. Simulations reveal that the algorithm has excellent convergence; it improves the position privacy protection level of mobile terminal and reduces the total cost of task offloading while maintaining lower energy consumption and latency.

Graph Theoretical Models for Enhancing Highway Connectivity and Safety in Vehicular Networks

Vehicular networks play a crucial role in modern transportation systems, significantly impacting connectivity and safety on highways. This paper explores the application of graph theoretical models to enhance both connectivity and safety in vehicular networks. Graph theory, a branch of discrete mathematics, provides a robust framework for modeling and analyzing complex networks, including those formed by vehicles on highways. Our study begins by defining the vehicular network as a graph where nodes represent vehicles, and edges denote communication links between them. We employ various graph theoretical concepts such as connectivity, centrality, and network flow to evaluate and improve the network's performance. Key metrics, including the degree of nodes, clustering coefficients, and shortest path lengths, are utilized to quantify network connectivity and identify critical nodes and edges that influence overall network efficiency. One of the primary objectives is to ensure uninterrupted connectivity in the presence of dynamic and often unpredictable vehicular movement. To this end, we analyze the network's resilience to node failures and propose strategies to enhance robustness using redundancy and alternative routing paths. By incorporating concepts like k-connectivity and network diameter, we develop models that maintain high levels of connectivity despite the removal or failure of multiple nodes or edges. Safety is addressed through the lens of network stability and reliability. We investigate the impact of vehicular density, speed, and communication range on the network's ability to sustain reliable communication channels. Techniques such as dynamic topology management and adaptive power control are proposed to mitigate the risks associated with network fragmentation and communication delays. Furthermore, we introduce optimization algorithms that leverage graph partitioning and community detection to improve the management of vehicular clusters, facilitating efficient data dissemination and reducing the likelihood of congestion-related incidents. The proposed models are validated through simulations that mimic real-world highway conditions, demonstrating significant improvements in both connectivity and safety metrics. In conclusion, the application of graph theoretical models offers a promising approach to enhancing highway connectivity and safety in vehicular networks

Observation of topological frequency combs.

On-chip generation of optical frequency combs using nonlinear ring resonators has enabled numerous applications of combs that were otherwise limited to mode-locked lasers. Nevertheless, on-chip frequency combs have relied predominantly on single-ring resonators. In this study, we experimentally demonstrate the generation of a novel class of frequency combs, the topological frequency combs, in a two-dimensional lattice of hundreds of ring resonators that hosts fabrication-robust topological edge states with linear dispersion. By pumping these edge states, we demonstrate the generation of a nested frequency comb that shows oscillation of multiple edge state resonances across ≈40 longitudinal modes and is spatially confined at the lattice edge. Our results provide an opportunity to explore the interplay between topological physics and nonlinear frequency comb generation in a commercially available nanophotonic platform.

Quantitative study on the acupoint effects based on complex network analysis: with the examples of the textbook networks formed by Chinese Acupuncture and Moxibustion Therapy and Practical Acupuncture and Moxibustion

By extracting the acupoint names and their main indications from cases in Chinese Acupuncture and Moxibustion Therapy and Practical Acupuncture and Moxibustion, the acupoints and their main indications are represented in a reduced dimension, establishing an "acupoint-indication" linkage. Using complex network detection results (node degree values), the specificity of acupoints was assessed. The small-world characteristics of the "acupoint-indication" network are utilized to analyze the consistency of acupoint selection in acupuncture prescriptions and strategies to avoid redundant acupoints. The results show that the "acupoint-indication" network formed by both texts exhibited an approximate "long-tail" distribution, with a large number of node degree values concentrated between 0 and 4 000, while a few nodes have degree values exceeding 10 000. There are significant differences in the number and distribution of nodes with degree values> 10 000 between the two texts. Chinese Acupuncture and Moxibustion Therapy includes 11 acupoints with multiple edges across the body, whereas Practical Acupuncture and Moxibustion contains only 2 such acupoints, located in the lower limbs. Clinically, some acupoints have a broad therapeutic effect and appear in numerous prescriptions. The division of acupoints based on node degree values can coarsely evaluate the body region specificity of acupoints' regulatory effects. The "acupoint-indication" network of Chinese Acupuncture and Moxibustion Therapy has a higher number of edges than that of Practical Acupuncture and Moxibustion, which might be related to the different historical contexts of the two texts. In the future, diagnostic and therapeutic patterns with historical continuity can be utilized to optimize acupuncture prescriptions.

Almost complex structure on finite points from bidirected graphs

We show that there is an almost complex structure on a differential calculus on finite points coming from a bidirected finite graph without multiple edges or loops. We concentrate on a polygon as a concrete case. In particular, a ‘holomorphic structure on the exterior bundle’ built from the polygon is studied. Also, a positive Hochschild 2-cocycle on the vertex set of the polygon, albeit a trivial one, is shown to arise naturally from the almost complex structure.

On the Perron root and eigenvectors of a non-negative integer matrix

In this paper, we obtain a combinatorial expression for the Perron root and eigenvectors of a non-negative integer matrix using techniques from symbolic dynamics. We associate such a matrix with a multigraph and consider the edge shift corresponding to it. This gives rise to a collection of forbidden words F which correspond to the non-existence of an edge between two vertices, and a collection of repeated words R with multiplicities which correspond to multiple edges between two vertices. In general, for given collections F of forbidden words and R of repeated words with pre-assigned multiplicities, we construct a generalized language as a multiset. A combinatorial expression that enumerates the number of words of fixed length in this generalized language gives the Perron root and eigenvectors of the adjacency matrix. We also obtain conditions under which such a generalized language is a language of an edge shift.

Non-Abelian Braiding of Topological Edge Bands.

Braiding is a geometric concept that manifests itself in a variety of scientific contexts from biology to physics, and has been employed to classify bulk band topology in topological materials. Topological edge states can also form braiding structures, as demonstrated recently in a type of topological insulators known as Möbius insulators, whose topological edge states form two braided bands exhibiting a Möbius twist. While the formation of Möbius twist is inspiring, it belongs to the simple Abelian braid group B_{2}. The most fascinating features about topological braids rely on the non-Abelianness in the higher-order braid group B_{N} (N≥3), which necessitates multiple edge bands, but so far it has not been discussed. Here, based on the gauge enriched symmetry, we develop a scheme to realize non-Abelian braiding of multiple topological edge bands. We propose tight-binding models of topological insulators that are able to generate topological edge states forming non-Abelian braiding structures. Experimental demonstrations are conducted in two acoustic crystals, which carry three and four braided acoustic edge bands, respectively. The observed braiding structure can correspond to the topological winding in the complex eigenvalue space of projective translation operator, akin to the previously established point-gap winding topology in the bulk of the Hatano-Nelson model. Our Letter also constitutes the realization of non-Abelian braiding topology on an actual crystal platform, but not based on the "virtual" synthetic dimensions.

Actions on the quiver: discrete quotients on the Coulomb branch

This paper introduces two operations in quiver gauge theories. The first operation, collapse, takes a quiver with a permutation symmetry Sn and gives a quiver with adjoint loops. The corresponding 3d N\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$ \\mathcal{N} $$\\end{document} = 4 Coulomb branches are related by an orbifold of Sn. The second operation, multi-lacing, takes a quiver with n nodes connected by edges of multiplicity k and replaces them by n nodes of multiplicity qk. The corresponding Coulomb branch moduli spaces are related by an orbifold of type ℤqn−1\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$ {\\mathbb{Z}}_q^{n-1} $$\\end{document}. Collapse generalises known cases that appeared in the literature [1–3]. These two operations can be combined to generate new relations between moduli spaces that are constructed using the magnetic construction.

Dynamic modeling of a sliding ring on an elastic rod with incremental potential formulation

Abstract Mechanical interactions between rigid rings and flexible cables find broad application in both daily life (hanging clothes) and engineering system (closing a tether-net). A reduced-order method for the dynamic analysis of sliding rings on a deformable one-dimensional (1D) rod-like object is proposed. In contrast to the conventional approach of discretising joint rings into multiple nodes and edges for contact detection and numerical simulation, a single point is used to reduce the order of the model. To ensure that the sliding ring and flexible rod do not deviate from their desired positions, a new barrier function is formulated using the incremental potential theory. Subsequently, the interaction between tangent frictional forces is obtained through a delayed dissipative approach. The proposed barrier functional and the associated frictional functional are C2 continuous, hence the nonlinear elastodynamic system can be solved variationally by an implicit time-stepping scheme. The numerical framework is initially applied to simple examples where the analytical solutions are available for validation. Then, multiple complex practical engineering examples are considered to showcase the effectiveness of the proposed method. The simplified ring-to-rod interaction model has the capacity to enhance the realism of visual effects in image animations, while simultaneously facilitating the optimisation of designs for space debris removal systems.

CAVE: Concurrency-Aware Graph Processing on SSDs

Large-scale graph analytics has become increasingly common in areas like social networks, physical sciences, transportation networks, and recommendation systems. Since many such practical graphs do not fit in main memory, graph analytics performance depends on efficiently utilizing underlying storage devices. These out-of-core graph processing systems employ sharding and sub-graph partitioning to optimize for storage while relying on efficient sequential access of traditional hard disks. However, today's storage is increasingly based on solid-state drives (SSDs) that exhibit high internal parallelism and efficient random accesses. Yet, state-of-the-art graph processing systems do not explicitly exploit those properties, resulting in subpar performance. In this paper, we develop CAVE, the first graph processing engine that optimally exploits underlying SSD-based storage by harnessing the available storage device parallelism via carefully selecting which I/Os to graph data can be issued concurrently. Thus, CAVE traverses multiple paths and processes multiple nodes and edges concurrently, achieving parallelization at a granular level. We identify two key ways to parallelize graph traversal algorithms based on the graph structure and algorithm: intra-subgraph and inter-subgraph parallelization. The first identifies subgraphs that contain vertices that can be accessed in parallel, while the latter identifies subgraphs that can be processed in their entirety in parallel. To showcase the benefit of our approach, we build within CAVE parallelized versions of five popular graph algorithms (Breadth-First Search, Depth-First Search, Weakly Connected Components, PageRank, Random Walk) that exploit the full bandwidth of the underlying device. CAVE uses a blocked file format based on adjacency lists and employs a concurrent cache pool that is essential to the parallelization of graph algorithms. By experimenting with different types of graphs on three SSD devices, we demonstrate that CAVE utilizes the available parallelism, and scales to diverse real-world graph datasets. CAVE achieves up to one order of magnitude speedup compared to the popular out-of-core systems Mosaic and GridGraph, and up to three orders of magnitude speedup in runtime compared to GraphChi.

Interface states and optical coupling functionalities in the super-SSH lattices

We theoretically address the coupling between trimer lattices and reveal the existence of stable multiple edge and interface states. It is shown the superlattice can provides a tunable number of topologically protected edge and interface states depending on the coupling strength and topological phase of the connecting lattices. Dynamics and transport properties of interface states are also investigated, Due to the interference of linear modes with different propagation constants, stable oscillations resulted from the coupling of interface states in finite trimerized waveguide arrays are observed and can give rise to optical coupling functionalities, including directional coupling, beam splitting and beam oscillator.

A high-performance dataflow-centric optimization framework for deep learning inference on the edge

Edge computing has been emerging as a popular scenario for model inference. However, the inference performance on edge devices (e.g., Multi-Core DSP, FGPA, etc.) suffers from inefficiency due to the lack of highly optimized inference frameworks. Previous model inference frameworks are mainly developed in an operator-centric way, which provides insufficient acceleration to edge-based inference. Besides, the operator-centric framework incurs significant costs for continuous development and maintenance.Targeting the existing drawbacks of operator-centric frameworks, we design Xenos, which can automatically conduct dataflow-centric optimization of the computation graph and accelerate inference in two dimensions. Vertically, Xenos develops operator linking technique to improve data locality by restructuring the inter-operator dataflow. Horizontally, Xenos develops DSP-aware operator split technique to enable higher parallelism across multiple DSP units. Our evaluation demonstrates the effectiveness of vertical and horizontal dataflow optimization, which reduce the inference time by 15.0%–84.9% and 17.9%–89.9% , respectively. Besides, Xenos also outperforms the widely-used TVM by 1.1×–1.9×. Moreover, we extend Xenos to a distributed solution, which we call d-Xenos. d-Xenos employs multiple edge devices to jointly conduct the inference task and achieves a speedup of 3.68×–3.78× compared with the single device.

Magnetically controllable multimode interference in topological photonic crystals

Topological photonic insulators show promise for applications in compact integrated photonic circuits due to their ability to transport light robustly through sharp bendings. The number of topological edge states relies on the difference between the bulk Chern numbers across the boundary, as dictated by the bulk edge correspondence. The interference among multiple topological edge modes in topological photonics systems may allow for controllable functionalities that are particularly desirable for constructing reconfigurable photonic devices. In this work, we demonstrate magnetically controllable multimode interference based on gyromagnetic topological photonic insulators that support two unidirectional edge modes with different dispersions. We successfully achieve controllable power splitting in experiments by engineering multimode interference with the magnetic field intensity or the frequency of wave. Our work demonstrates that manipulating the interference among multiple chiral edge modes can facilitate the advancement of highly efficient and adaptable microwave devices.

Introduction of Local Resonators to a Nonlinear Metamaterial With Topological Features

Abstract Recent work in nonlinear topological metamaterials has revealed many useful properties such as amplitude dependent localized vibration modes and nonreciprocal wave propagation. However, thus far, there have not been any studies to include the use of local resonators in these systems. This work seeks to fill that gap through investigating a nonlinear quasi-periodic metamaterial with periodic local resonator attachments. We model a one-dimensional metamaterial lattice as a spring-mass chain with coupled local resonators. Quasi-periodic modulation in the nonlinear connecting springs is utilized to achieve topological features. For comparison, a similar system without local resonators is also modeled. Both analytical and numerical methods are used to study this system. The dispersion relation of the infinite chain of the proposed system is determined analytically through the perturbation method of multiple scales. This analytical solution is compared to the finite chain response, estimated using the method of harmonic balance and solved numerically. The resulting band structures and mode shapes are used to study the effects of quasi-periodic parameters and excitation amplitude on the system behavior both with and without the presence of local resonators. Specifically, the impact of local resonators on topological features such as edge modes is established, demonstrating the appearance of a trivial bandgap and multiple localized edge states for both main cells and local resonators. These results highlight the interplay between local resonance and nonlinearity in a topological metamaterial demonstrating for the first time the presence of an amplitude invariant bandgap alongside amplitude dependent topological bandgaps.

An offloading and pricing mechanism based on virtualization in edge–cloud computing

Integrate users, edge servers and cloud into a system can improve resource utilization and quality of service (QoS). The users can send the same request to multiple Edge Servers. However, edge servers have limited scope, which may result in repeated offloading tasks for the same user. In this paper, we consider the cloud resource migration in the form of virtual machines (VMs) to achieve high computing efficiency and low latency requirements. Obviously, the migration of VMs causes changes in the number and price of edge resources, which further affects the task offloading strategy. Therefore, it is necessary to optimize the VMs migration strategy and resource pricing strategy successively to achieve the optimal offloading strategy suitable for users. First, based on the competition between edge servers, a static VMs migration algorithm (SVMMA) is designed to formulate the strategy and price of VMs migration. Then, according to the users’ resource needs and the resource possession of different edge servers, a resource partitioning approach (RPA) have been proposed. Finally, two different QoS functions are evaluated, representing the QoS matching degree between tasks and multiple resource combination blocks (MRCBs). Wherein, a dynamic matching and pricing of tasks algorithm (DMPT) is conducted to obtain the optimal offloading strategy and price of the task. Experiments show that our algorithms can improve the offloading rate of tasks and achieve the relative balance among user, cloud server, and edge server benefits.

An efficient mechanism for function scheduling and placement in function as a service edge environment

With the advancement of Internet of Things (IoT) applications and the substantial surge in associated traffic, coupled with the rise in demand for low-latency mobile applications, the utilization of remote cloud computing infrastructure has encountered notable challenges. As a solution, edge computing has emerged, decentralizing computational resources closer to data sources. This results in quicker processing and diminished reliance on distant servers. To address resource limitations and prevent congestion in edge servers, applications are now being developed as smaller, modular components, referred to as the Function as a Service (FaaS) pattern. FaaS brings advantages such as distributing the workload across multiple edge servers and offering improved flexibility in resource management. However, the FaaS model introduces new challenges when compared to the traditional edge computing environment. These include issues like function placement and minimizing cold start delays. Assignment mechanisms play a crucial role in characterizing the interactions between edge servers and user devices. They excel at devising optimal allocation strategies and ensuring mutual satisfaction between edge servers and user devices, especially in scenarios with limited resources. This article introduces the Edge FaaS-Top Trading Cycles (EF-TTC), an efficient mechanism for implementing FaaS-based applications within the edge platform. The proposed approach models the problem of function placement and scheduling in a manner akin to the well-known School Choice problem. It subsequently employs the Top Trading Cycle mechanism to address this challenge. Experimental results demonstrate that the proposed approach significantly improves the average completion time for users’ requests, outperforming other baseline algorithm in this domain.