Trust2Vec: Large-Scale IoT Trust Management System Based on Signed Network Embeddings

Abstract

A trust management system (TMS) is an integral component of any IoT network. A reliable trust management system must guarantee the network security, data integrity, and act as a referee that promotes legitimate devices, and punishes any malicious activities. Trust scores assigned by TMSs reflect devices' reputations, which can help predict the future behaviours of network entities and subsequently judge the reliability of different entities in IoT networks. Many TMSs have been proposed in the literature, these systems are designed for smallscale trust attacks, and can deal with attacks where a malicious device tries to undermine TMS by spreading fake trust reports. However, these systems are prone to large-scale trust attacks. To address this problem, in this paper, we propose a TMS for large-scale IoT systems called Trust2Vec, which can manage trust relationships in large-scale IoT systems and can mitigate largescale trust attacks that are performed by hundreds of malicious devices. Trust2Vec leverages a random-walk network exploration algorithm that navigates the trust relationship among devices and computes trust network embeddings, which enables it to analyze the latent network structure of trust relationships, even if there is no direct trust rating between two malicious devices. To detect large-scale attacks, such as self-promoting and bad-mouthing, we propose a network embeddings community detection algorithm that detects and blocks communities of malicious nodes. The effectiveness of Trust2Vec is validated through large-scale IoT network simulation. The results show that Trust2Vec can achieve up to 94% mitigation rate in various network settings.

Hybrid Fuzzy Rule Algorithm and Trust Planning Mechanism for Robust Trust Management in IoT-Embedded Systems Integration

“…Trust Prediction Accuracy (%) Trust2Vec [24] 56 Trust-FTSR [25] 81 RFA [26] 62 EMBTR [27] 71 Trust-FTSR [25], RFA [26] and EMBTR [27].…”

VeSoNet: Traffic-Aware Content Caching for Vehicular Social Networks Using Deep Reinforcement Learning

“…The content network is represented as a graph data structure, where the nodes represent the data content, and the edges represent the content similarity between these nodes. Vehicle2vec learns the content node low dimensional vector that preserves the neighbourhood of nodes in the original graph [18]. To build such content node embeddings, vehi-cle2vec uses stochastic gradient descent (SGD) to optimize the objective function, hence learning the low dimensional representation.…”

Bipolar Fuzzy PROMETHEE‐based Decision‐Making Trust Model for guaranteeing data dissemination reliability in Internet of Things

“…This LSTM‐SMART approach also confirmed 99.76% of F‐measure and 99.87% of accuracy with 100 number of iterations, compared to the baseline approaches. Dhelim et al 18 proposed trust management scheme termed Trust2Vec for managing the relationships of trust in IoT environment. This Trust2Vec was proposed for handling the impacts of large‐scale trust attacks that are generally launched by huge number of malicious nodes in the network.…”

“…In the proposed BFPSMTM methodology, the node alternative with lesser incoming and greater outgoing flow is said to be cooperative and is preferred during routing. The nodes from the set of alternatives can be chosen depending on both positive and negative outranking flows found from expressions as shown in Equations ( 17) and (18).…”