Learning Structures of Conceptual Models from Observed Dynamics Using Evolutionary Echo State Networks

Abdelbari, Hassan; Shafi, Kamran

doi:10.1515/jaiscr-2018-0010

Cited by 7 publications

(5 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Two other approaches to modeling causal worlds are system-dynamics (SD) models [12], [13] and Bayesian belief networks (BBNs) [14].…”

Section: Comparison With Other Methodsmentioning

confidence: 99%

“…A FCM's overall cyclic signed digraph structure resembles a feedback neural or semantic network. FCMs also resemble causal loop diagrams and systems-dynamics models [12], [13] because nonlinear difference or differential equations describe FCM concept nodes (and sometimes FCM causal edges). A FCM's directed graph structure permits inference through forward chaining.…”

Section: Fuzzy Cognitive Maps As Models Of Causal Feedbackmentioning

confidence: 99%

See 1 more Smart Citation

Causal Modeling with Feedback Fuzzy Cognitive Maps

Osoba

Kosko

2019

Social‐Behavioral Modeling for Complex Systems

View full text Add to dashboard Cite

Fuzzy cognitive maps (FCMs) model feedback causal relations in interwoven webs of causality and policy variables. FCMs are fuzzy signed directed graphs that allow degrees of causal influence and event occurrence. Such causal models can simulate a wide range of policy scenarios and decision processes. Their directed loops or cycles directly model causal feedback. Their nonlinear dynamics permit forward-chaining inference from input causes and policy options to output effects. Users can add detailed dynamics and feedback links directly to the causal model or infer them with statistical learning laws. Users can fuse or combine FCMs from multiple experts by weighting and adding the underlying fuzzy edge matrices and do so recursively if needed. The combined FCM tends to better represent domain knowledge as the expert sample size increases if the expert sample approximates a random sample. Many causal models use more restrictive directed acyclic graphs (DAGs) and Bayesian probabilities. DAGs do not model causal feedback because they do not contain closed loops. Combining DAGs also tends to produce cycles and thus tends not to produce a new DAG. Combining DAGs tends to produce a FCM. FCM causal influence is also transitive whereas probabilistic causal influence is not transitive in general. Overall: FCMs trade the numerical precision of probabilistic DAGs for pattern prediction, faster and scalable computation, ease of combination, and richer feedback representation. We show how FCMs can apply to problems of public support for insurgency and terrorism and to US-China conflict relations in Graham Allison's Thucydides-trap framework. The Thucydides-trap FCM predicts war-like conflict for nearly 80% of input scenarios where we treat all such scenarios

show abstract

“…Two other approaches to modeling causal worlds are system-dynamics (SD) models [12], [13] and Bayesian belief networks (BBNs) [14].…”

Section: Comparison With Other Methodsmentioning

confidence: 99%

Section: Fuzzy Cognitive Maps As Models Of Causal Feedbackmentioning

confidence: 99%

Causal Modeling with Feedback Fuzzy Cognitive Maps

Osoba

Kosko

2019

Social‐Behavioral Modeling for Complex Systems

View full text Add to dashboard Cite

show abstract

“…The dynamical reservoir weights are designed to maintain an echo state property by generating initial random sparse weights within the [−1, 1] range. Finally, the reservoir weights matrix is multiplied by a chosen value called spectral radius [52]. The ReLU activation function is used for all these five models' hidden layers and the softmax activation function is used for the output layer.…”

Section: B Experimental Setupsmentioning

confidence: 99%

Replay Attack Detection in Smart Cities Using Deep Learning

et al. 2020

View full text Add to dashboard Cite

Intrusion detection is an important and challenging problem that has a major impact on quality and reliability of smart city services. To this extent, replay attacks have been one of the most common threats on smart city infrastructure, which compromises authentication in a smart city network. For example, a replay attack may physically damage smart city infrastructure resulting in loss of sensitive data, incurring considerable financial damages. Therefore, towards securing smart cities from reply attacks, intrusion detection systems and frameworks based on deep learning have been proposed in the recent literature. However, the absence of the time dimension of these proposals is a major limitation. Therefore, we have developed a deep learning-based model for replay attack detection in smart cities. The novelty of the proposed methodology resides in the adoption of deep learning based models as an application for detecting replay attacks to improve detection accuracy. The performance of this model is evaluated by applying it to a real life smart city dataset, where replay attacks were simulated. Our results show that the proposed model is capable of distinguishing between normal and attack behaviours with relatively high accuracy. In addition, according to the results, our proposed model outperforms traditional classification and deep learning models. Last but not least, as an additional contribution, this paper presents a real life smart city data set with simulated replay attacks for future research. INDEX TERMS Smart cities, intrusion detection, replay attack, deep learning, convolutional neural networks.

show abstract

“…The GP algorithm is based on a typical schema of a population-based algorithm (see e.g. [2,42,43,49]). It uses two populations, i.e.…”

Section: Description Of the Proposed Gp Algorithmmentioning

confidence: 99%

Genetic Programming Algorithm for Designing of Control Systems

Cpałka¹,

Łapa²,

Przybył³

2018

ITC

View full text Add to dashboard Cite

,andrzej.przybyl}@iisi.pcz.pl Genetic programming algorithms and other population-based methods are a convenient tool for solving complex interdisciplinary problems. Their characteristic feature is that they flexibly adapt to the problem and expectations of a designer. In this paper, they are used for designing complex control systems. In particular, a new approach for automatic designing of PID-based controllers which are resistant to noises in the measuring path is proposed. It is based on the knowledge about an object model and capabilities of the genetic algorithm and genetic programming. Not only do these make it possible to tune parameters and select the structure of PID-based controllers, but they also tune parameters of some additional components of a complex controller structure, like parameters of finite impulse response (FIR) filters. The idea of the proposed approach relies on proper encoding of the controller and a dedicated way of evolutionary processing of encoded solutions. The approach proposed in this paper has been tested using a typical control problem, i.e. a DC motor control problem.

show abstract

Learning Structures of Conceptual Models from Observed Dynamics Using Evolutionary Echo State Networks

Cited by 7 publications

References 37 publications

Causal Modeling with Feedback Fuzzy Cognitive Maps

Causal Modeling with Feedback Fuzzy Cognitive Maps

Replay Attack Detection in Smart Cities Using Deep Learning

Genetic Programming Algorithm for Designing of Control Systems

Contact Info

Product

Resources

About