International Journal of Reasoning-based Intelligent Systems (10 papers in press)
A Metaphorical Situation Annotation Framework
by Mehdi Fesharaki, Abdolhamid Fetanat, Mehdi Amjadi
Abstract: Recent advances in cognitive sciences show that metaphorical thinking plays an essential role in human cognition. Metaphors are basic building blocks that construct a human minds understanding of the world. However, formal models of cognition, intelligence and even natural language have mostly ignored metaphorical thinking. Consequently, formal processing machines and intelligent agents developed based on them cannot process metaphors and thus, cannot use human minds metaphorical capabilities. There are formal models of metaphorical reasoning mainly developed in psychology that can be used as the basis for metaphorical intelligent systems. These models describe metaphorical reasoning as a process comprised of two logically distinct phases: metaphorical structure mapping and metaphorical knowledge transfer. They even propose simple algorithms for implementing metaphorical reasoning. In this paper, we propose a formal framework that provides tools and methods for a metaphorical communication between human and machine agents in a situation awareness environment, based on the formal models of metaphorical reasoning. The framework leaves the first phase of metaphorical reasoning to human minds and the second phase to machine agents. Then, it utilizes the metaphorical semantic tags as the mediators between the two phases of metaphor processing. Therefore, the metaphorical tags enable the sharing of metaphorical wisdom of human agents with machine agents. The framework implements metaphorical tags and their processing schemes based on common semantic technologies and thus, is an easily implementable extension to existing semantic applications.
Keywords: Metaphorical Reasoning; semantic web; situation awareness; tag-based annotation; human-computer interaction.
Analysis of IDS Alerts by Generalizing Features and Discovering Emerging Patterns
by Mahdi Maleki, Seyed Mansour Shahidi
Abstract: One of the significant problems in using intrusion detection systems is the high volume of low-level Alerts. In this paper, an appropriate analysis of cyber alerts has been used to reduce low-level alerts utilizing a range of available features of attacks. It has also benefited from the discovery of emerging patterns to improve situational awareness in cyber attacks. Moving to different levels of generalization and extraction of rules; based on Attribute-Oriented Induction and emerging patterns is a remarkable achievement of this. To evaluate the proposed method, a new CICIDS2017 database is used to eliminate the defects of the previous datasets. The results show a decrease in alerts at the rate of 99% at the lowest generalization level and an average of 25% at other generalization levels. In addition to normal traffic, 14 different types of attacks have been identified. The Dos Hulk attack has the highest frequency with 8.16%, and the Heartbleed attack having the lowest frequency with 0.0004% frequency. On average, 18 overlap (TO-EP) pattern, 63 relatively subsumption-overlap pattern (SO-EP), and 92 similar (SIM-EP) pattern have been extracted at four generalization levels.
Keywords: Intrusion Detection System; Feature Generalization; Multidimensional Data Mining; OLAP; Multistage Attacks.
A Systematic Mapping Study about Applications of Knowledge Graphs in Agribusiness
by Mario José Diván, Valerio Frittelli
Abstract: Knowledge management refers to data jointly with their meaning and relationships to foster a consistent, accessible, and integrated view of knowledge. Complemented by the recommendation strategies, it is useful for supporting different decision-making processes in intelligent systems. The knowledge graphs (KGs) allows describing entities, objects, events, or concepts, placing data in every context, fostering an articulation between data and its associated meaning. This article aims to address a broad analysis of the modelling and application of the KG in the agribusiness sector, following the systematic mapping study methodology. The Scopus, IEEE, ScienceDirect, and ACM databases are considered for performing the different queries according to the research questions aligned with the objective. The main scope of this study is constituted of articles written in the English language, published within the last ten years. Initially, a total of 156 documents are obtained. Finally, 12 are retained and considered according to the defined filters and after a careful and full reading. The KG applied in the agribusiness sector keeps being a challenge, existing different opportunities to be addressed.
Keywords: systematic mapping study; SMS; agribusiness applications; knowledge graphs; frameworks; platforms.
Image-based Anomaly Detection Using CNN Cues Generalization in Face Recognition System
by Margarita Favorskaya, Andrey I. Pakhirka
Abstract: Anomaly detection caused by face presentation attacks fundamentally reduces the vulnerability of the face recognition system. Face presentation attacks can be regarded as presentations obtained from secondary sources that are different in nature. This fact means a wide range of presentation attacks, both now and in the future, as well as different approaches to preventing them. We adhere to a one-class classifier approach, but present it via an ensemble classifier based on the analysis of visual maps. During training, we first create a set of maps looking for chromatic colours, specular reflection, blur, and colour diversity features. Second, each of the maps is generalised using CNN stream with a simple architecture followed by an ensemble classifier. Such hybrid detection of face presentation attacks provides fast implementation and protects against unknown attacks in the future. Experiments with public face datasets and own face dataset confirmed our approach.
Keywords: anomaly detection; face recognition; presentation attacks; cue generalisation; ensemble classifier.
An improvement of hidden Markov model for stock market predictions
by Seyed Kazem Chavoshi, Azadeh Mansouri, Sorour Sheidani
Abstract: This paper predicts Tehran Exchange Dividend and Price Indexrn(TEDPIX) by finding a pattern in TEDPIX through settled transactions andrnopen orders volume effects. To do so, we improve an Autoregressive HiddenrnMarkov Model (AR-HMM) by adding a more hidden layer. Then, we utilized arngenetic algorithm for long term daily trend predictions. By exploiting thernobtained information of predicted 5 days using the genetic algorithm, wernupdate the parameters of improved AR-HMM. This stepwise prediction updating process continues until all desired number of future days stockrnexchange indices get predicted. Comparing our new scheme with other studiedrnMarkov family models shows that the added features lead to achieve morernaccuracy and less prediction errors. Experimental results show that MeanrnAbsolute Percentage Error of all predictions by our improved AR-HMMrnapproach are less than 5 percent which indicates far better performance of ourrnmethod against Markov and Hidden Markov Models.
Keywords: Autoregressive; Hidden Markov Models; TEDPIX; Settled transactions; Open orders.
A Reasoning about Number Theory and Brain Signals
by Stevo Bozinovski, Adrijan Bozinovski
Abstract: This paper presents a process of reasoning on a relation between Number Theory and brain rhythms. So far these scientific disciplines have been considered separate. The relation is discovered observing the Pascal's Triangle and central frequencies of brain rhythms. That way the paper defines a theoretical process underlying the ranges of brain rhythms. The paper also comments on an oscillatory process from Number Theory, as well as on the applications of brain rhythms in Brain-Computer Interface
Keywords: Number Theory; brain rhythms; Pascal Triangle; Brain-Computer Interface.
Sustainable Intuitionistic Fuzzy Inventory Models with Preservation Technology Investment and Shortages
by Swagatika Sahoo, Milu Acharya, Srikanta Patnaik
Abstract: In this paper, complete backordering and partial backordering EOQ models are developed in crisp and fuzzy sense. Spending on preservation technology, the rate of deterioration of the product is kept under control. We provide a cost for carbon emissions in the EOQ model because carbon emissions are a major contributor to climate change and global warming. This paper also describes a fuzzy model by taking a generalised pentagonal intuitionistic fuzzy number for per unit holding cost and for de-fuzzification purposes, ranking generalised pentagonal intuitionistic fuzzy number is used. The main purpose of the models is to minimise the total costs. Illustration of the suggested model is conducted through numerical examples. In this problem, a convexity check of the cost function is performed and a sensitivity test is also tried to study the effect of change in the values of the parameters.
Keywords: deteriorating items; preservation technology; complete and partial backordering; disposal cost; salvage value; carbon emission; intuitionistic fuzzy; ranking formula.
Defect Detection through Customised Reduction and Hybrid Convolution Classification over Super-pixel Clusters
by Matthew Immanuel Samson, Hossam A. Gaber
Abstract: Defect detection is the process of locating defects or anomalies within an object that include changes in textures, features, patterns, missing part along with other object modifications. The paper discusses some of the main challenges of defect detection including details on sample selection, object orientation, semantic segmentation and image defect classification. This paper focuses on applying modified machine and deep learning models to analyse defects with wide object invariance. We demonstrate algorithms that perform multi-class classification with improvements in the image segmentation process that directly connect to the deep model architecture. Before applying learning algorithms, the paper also demonstrates the value of sample selection together with a more simplified normalised dimension reduction based on image downscaling even before using the convolution operation of the convolutional neural networks (CNN).
Keywords: image processing; computer vision; machine learning; deep learning; defect detection; classification; localisation; segmentation.
Exploiting Discourse Relations to Produce Arabic Extracts
by Samira Lagrini, Nabiha Azizi, Mohammed Redjimi
Abstract: Text summarisation is one of the interesting tools for a quick and optimal exploitation of the huge amount of online textual documents. Several approaches have been proposed to date to produce extractive summaries in Arabic. However, in most cases, the linguistic qualities of the generated summary are not satisfactory. In this paper, we attempt to overcome this limitation by proposing a new approach for single-document summarisation that combines a discourse analysis following the rhetorical structure theory (RST) framework and a score-based method. Unlike traditional RST-based approaches, the proposed approach relies on exploiting intra-sentence discourse relations instead of text discourse structure to produce a primary summary. Then, each sentence within the primary summary is evaluated based on a combination of statistical and linguistic features to produce the final summary considering user compression rate. The proposed approach was evaluated under Essex Arabic Summaries Corpus (EASC) using ROUGE-1 and ROUGE-2 measures, and compared against other existing methods. A human evaluation was also conducted in order to assess the linguistic qualities of generated summaries. Experimental results are very encouraging and prove that, exploiting discourse relations is very useful to produce Arabic extractive summaries with good linguistic qualities.
Keywords: extractive single document summarisation; Arabic discourse analysis; Arabic discourse relations; score based; statistical features.
A method of fine-grained image fuzzy main color segmentation based on visual perception
by Juanjuan Liu, Feng Gao
Abstract: In order to overcome the low segmentation accuracy of traditional main colour segmentation methods, this paper proposes a fine-grained image fuzzy main colour segmentation method based on visual perception. The colour features of fine-grained graph are optimised by using visual perception technology, and the fine-grained region features are obtained through the maximum colour link region and the surrounding colour roughness of fine-grained graph. A bilateral filter is used to enhance the details of fine-grained image, and fuzzy clustering is applied to the time-domain difference image. The edge contour of the target image is cut through the edge detection step to obtain the foreground area and complete the colour segmentation. The experimental results show that the effect of image brightness preservation is good, the segmentation accuracy is close to 100%, JS value is close to 1, and the segmentation effect is good.
Keywords: visual perception; fine-grained image; colour roughness; image foreground; main colour segmentation.