Inference System Research Articles

A fundamental analysis on PAHs in food products to detect the toxicity index using fuzzy logic system

Polycyclic Aromatic Hydrocarbons (PAHs) are complex chemical compounds that occur naturally in unprocessed food when it is exposed to contaminated air during transportation, natural emission such as volcano, forest fire and through pesticides spray. It is reported by different agencies that there are 16 types of PAHs in which BaP (Benzo[a] pyrene), BaA (Benz[a]anthracene), BbF(Benzo [b] fluoranthene), Chr (Chrysene) are considered to be carcinogenic and it can occur due to different processes. In processed food it occurs due to various processing methods like overheating, incomplete burning, drying etc. The presence of PAH in food is conventionally found through analytical, traditional, and semi-automatic methods. These methods are found to be valuable but expensive and time-consuming. Further, these methods are used only for the detection of PAHs and the toxicity level is measured or identified based on expert knowledge of researchers and the Standards. Therefore, in this research, a simple harmfulness index system has been developed using Fuzzy Logic System(FLS). The proposed system has been designed based on the PAH values of different food and food products. Hence to initiate the study and to determine the significance of the results, PAH data have been collected from different articles that have investigated food products experimentally. These PAH data were analyzed using statistical measures such as Min, Mean, Max, Standard Deviation, Variance and Kurtosis method. Based on the observations from the results, the fuzzy sets were designed with four membership functions for each PAH and the rules were framed. The strength output from the inference engine has been associated with harmfulness index such as normal, low risk, medium risk, and high risk. From the evaluation, it can be observed that 89.72% of the food samples were recognized along with their degree of harmfulness. Also it can be inferred that 11% of the misclassified samples showed clear metrics of their harmfulness with PAH variations.

GraphPPL.jl: A Probabilistic Programming Language for Graphical Models

This paper presents GraphPPL.jl, a novel probabilistic programming language designed for graphical models. GraphPPL.jl uniquely represents probabilistic models as factor graphs. A notable feature of GraphPPL.jl is its model nesting capability, which facilitates the creation of modular graphical models and significantly simplifies the development of large (hierarchical) graphical models. Furthermore, GraphPPL.jl offers a plugin system to incorporate inference-specific information into the graph, allowing integration with various well-known inference engines. To demonstrate this, GraphPPL.jl includes a flexible plugin to define a Constrained Bethe Free Energy minimization process, also known as variational inference. In particular, the Constrained Bethe Free Energy defined by GraphPPL.jl serves as a potential inference framework for numerous well-known inference backends, making it a versatile tool for diverse applications. This paper details the design and implementation of GraphPPL.jl, highlighting its power, expressiveness, and user-friendliness. It also emphasizes the clear separation between model definition and inference while providing developers with extensibility and customization options. This establishes GraphPPL.jl as a high-level user interface language that allows users to create complex graphical models without being burdened with the complexity of inference while allowing backend developers to easily adopt GraphPPL.jl as their frontend language.

RDSSMM: A FUZZY EXPERT SYSTEM FOR ENHANCED MALARIA DIAGNOSIS AND MANAGEMENT IN ENDEMIC REGIONS

Malaria continues to be a pressing public health issue in tropical regions, with its deleterious effects on human health well documented. While a minority of cases may be life threatening, delayed diagnosis can significantly worsen the severity of malaria. This paper presents the Recommender and Decision Support System for Malaria Management (RDSSMM), designed to assist researchers, physicians and healthcare professionals in malaria endemic areas. RDSSMM is made up of four main components: a knowledge base, fuzzification module, inference engine, and defuzzification module. Mayo Hospital Lahore, Pakistan was used to develop the fuzzy expert system based on experts’ opinions in the field of medicine. For ease of access, a mobile application has been included which allows patients to get primary diagnosis by keying in their physical signs that are then transformed into fuzzy membership functions. The data was collected from the outpatient department of Mayo Hospital Lahore as one way of evaluating RDSSMM’s effectiveness. Comparing its predictions to expert medical diagnoses and reports pointed out this system’s performance. Malaria risk was correctly predicted 80% of the time by RDSSMM. These findings show that RDSSMM can greatly improve malaria management in rural areas through provision of quick and accurate results, which would facilitate timely decisions regarding malaria testing and monitoring.

Development of intelligent suite for malaria pathogen detection in microscopy images

The identification of malaria infection using microscope images of blood smears is considered as a ‘gold standard’. The diagnosis of malaria needs expert microscopists which are scarce in remote areas where malaria is endemic. Therefore, it is desirable to automate the repetitive task of pathogen detection in the blood samples received as microscope images. This study provides an easy to use and deploy method for implementing a malaria pathogen detection software- the Intelligent Suite. The Intelligent Suite features a graphical user interface (GUI) implemented using ‘cvui’ library to interact with the OpenVINO’s inference engine for model optimisation and deployment across several inference devices. The intelligent Suite uses a custom YOLO-mp-3l model trained on Darknet framework for detection of malaria pathogen in thick smear microscope images. Moreover, the Intelligent Suite provides user interface for inference device/mode selection, alter model parameters, and generate detection reports along with the model performance metrics. The Intelligent Suite was executed on a CPU computer with model inference running on a plug-and-play Neural Compute Stick (NCS2) and performance reported.

Prediction of confined compressive strength of concrete column strengthened with FRCM composites

AbstractNowadays, retrofitting and rehabilitation of deteriorated reinforced concrete structures are becoming a growing need of the construction industry instead of demolishing aged structures. The application of fabric‐reinforced cementitious matrix (FRCM) on the existing concrete structures is one of the sustainable solutions to retrofit the concrete structures. This study used machine learning (ML) models such as linear regression (LR), support vector machines (SVM), and adaptive neuro‐fuzzy inference systems (ANFIS) to estimate the compressive strength (CS) of columns wrapped with FRCM. The experimental dataset of 301 column specimens was collected including input parameters such as cross‐sectional properties, mechanical properties of concrete and steel, and characteristics of FRCM material. Apart from ML models, seven analytical models were also used to compare the accuracy and precision of ML models. The results illustrate that the ANFIS model outperformed other ML models and established itself as a dependable and precise model. The R‐value of the ANFIS model was 0.9816, whereas R‐values of 0.9269 and 0.9572 were achieved by LR and SVM models, respectively. In addition, the MAPE value acquired by the ANFIS model was 1.52% which was lower than those of the LR model by 73.24%, and the SVM model by 60.60%, respectively. As the precision of the ANFIS model was higher as compared with SVM and LR models, so, the developed ANFIS‐based mathematical model can be easily used to predict the CS of FRCM‐strengthened concrete columns. The developed model is accurate, economical, and fast; and can be utilized by FRCM applicators and structural designers.

Development and evaluation of a panel of newly screened Y chromosome InDels for inferring paternal ancestry information in Southwest China.

Y-InDels (insertions/deletions) are genetic markers which are extremely understudied. It is unknown whether this type of markers can be utilized for genetic ancestry inference. We have developed an innovative Y chromosome ancestry inference system tailored for forensic applications. This panel amplifies 21 Y chromosome loci, encompassing Y-InDels and Y-SNPs (Single Nucleotide Polymorphism), utilizing the capillary electrophoresis (CE) platform. The system performed well at DNA concentrations greater than 0.125 ng/ul and produced accurate results at a 1:100 mixing ratio of male and female DNA. The Cumulative probability of matching (CPM) was between 0.95 and 0.97 in the experimental population. The system's efficacy in inferring ancestral origins was demonstrated through intercontinental population discrimination, revealing high discrimination power between African and East Asian populations. Population genetic analyses conducted on Han, Qiang and Hui populations in Southwest China, where the smallest FST value was 0.0002 between Han Chinese in Beijing (from 1000 Genomes Project) and Qiang Chinese from Sichuan (CQSC). Phylogenetic tree construction further illuminated distinct haplotypes among populations, with ethnically unique haplotypes observed in 34.6% of Hui and 7.1% of Qiang populations. K-fold cross-validation show the system's inference abilities at the intercontinental level. In addition, our investigations identified potential associations between the Y-InDel locus Y: 15,385,547 (GRCh37) and haplogroup R1a1a1b2a2- Z2124, as well as locus Y: 13,990,180 (GRCh37) and haplogroup F-M89. In conclusion, we have established a Y-chromosome inference system tailored for grassroots-level application, underscoring the value of incorporating Y-InDel markers in forensic analyses.

AUTOMAP: Inferring Rank-Polymorphic Function Applications with Integer Linear Programming

Dynamically typed array languages such as Python, APL, and Matlab lift scalar operations to arrays and replicate scalars to fit applications. We present a mechanism for automatically inferring map and replicate operations in a statically-typed language in a way that resembles the programming experience of a dynamically-typed language while preserving the static typing guarantees. Our type system---which supports parametric polymorphism, higher-order functions, and top-level let-generalization---makes use of integer linear programming in order to find the minimum number of operations needed to elaborate to a well-typed program. We argue that the inference system provides useful and unsurprising guarantees to the programmer. We demonstrate important theoretical properties of the mechanism and report on the implementation of the mechanism in the statically-typed array programming language Futhark.

Programmable MCMC with Soundly Composed Guide Programs

Probabilistic programming languages (PPLs) provide language support for expressing flexible probabilistic models and solving Bayesian inference problems. PPLs with programmable inference make it possible for users to obtain improved results by customizing inference engines using guide programs that are tailored to a corresponding model program. However, errors in guide programs can compromise the statistical soundness of the inference. This article introduces a novel coroutine-based framework for verifying the correctness of user-written guide programs for a broad class of Markov chain Monte Carlo (MCMC) inference algorithms. Our approach rests on a novel type system for describing communication protocols between a model program and a sequence of guides that each update only a subset of random variables. We prove that, by translating guide types to context-free processes with finite norms, it is possible to check structural type equality between models and guides in polynomial time. This connection gives rise to an efficient type-inference algorithm for probabilistic programs with flexible constructs such as general recursion and branching. We also contribute a coverage-checking algorithm that verifies the support of sequentially composed guide programs agrees with that of the model program, which is a key soundness condition for MCMC inference with multiple guides. Evaluations on diverse benchmarks show that our type-inference and coverage-checking algorithms efficiently infer types and detect sound and unsound guides for programs that existing static analyses cannot handle.

Declarative probabilistic logic programming in discrete-continuous domains

Over the past three decades, the logic programming paradigm has been successfully expanded to support probabilistic modeling, inference and learning. The resulting paradigm of probabilistic logic programming (PLP) and its programming languages owes much of its success to a declarative semantics, the so-called distribution semantics. However, the distribution semantics is limited to discrete random variables only. While PLP has been extended in various ways for supporting hybrid, that is, mixed discrete and continuous random variables, we are still lacking a declarative semantics for hybrid PLP that not only generalizes the distribution semantics and the modeling language but also the standard inference algorithm that is based on knowledge compilation. We contribute the measure semantics together with the hybrid PLP language DC-ProbLog (where DC stands for distributional clauses) and its inference engine infinitesimal algebraic likelihood weighting (IALW). These have the original distribution semantics, standard PLP languages such as ProbLog, and standard inference engines for PLP based on knowledge compilation as special cases. Thus, we generalize the state of the art of PLP towards hybrid PLP in three different aspects: semantics, language and inference. Furthermore, IALW is the first inference algorithm for hybrid probabilistic programming based on knowledge compilation.

An expert system of recommendations for combating cyber threats using CVSS metrics and game theory

This study is focused on the creation of an expert system for generating recommendations on cyber security. The developed expert system uses a game-theoretic model as a inference engine to transform expert knowledge into recommendations for end-users, who may be chief IT security officers (CISOs), system administrators, or cyber security engineers. Expert knowledge is presented in the form of an estimate of the base group of CVSS metrics - Common Vulnerability Score System, for each type of attack and adjusted values of CVSS in the case that the counterattack strategy is applied. Given a set of attacks and a base of expert attack knowledge, the system generates a game matrix of zero-sum game with a cybercriminal and a cyberdefense expert as players. The inference engine of the expert system is a game-theoretic model responsible for solving the game using the Brown-Robinson iterative method and generating cyber protection recommendations. An experiment was conducted on the convergence of the BrownRobinson algorithm on the 2022 vulnerability dataset from the Cybersecurity and Infrastructure Security Agency database, as a result of which it was determined that the convergence of the algorithm for solving the matrix game is achieved at a number of iterations of 1000. As a result of the work, expert system was designed and implemented along with the Web interface, which provides input by experts of CVSS level assessments of collected threats, threats countermeasures and output of recommendations for combating cyber threats

ShaderNN: A lightweight and efficient inference engine for real-time applications on mobile GPUs

Inference using deep neural networks on mobile devices has been an active area of research in recent years. The design of a deep learning inference framework targeted for mobile devices needs to consider various factors, such as the limited computational capacity of the devices, low power budget, varied memory access methods, and I/O bus bandwidth governed by the underlying processor’s architecture. Furthermore, integrating an inference framework with time-sensitive applications — such as games and video-based software to perform tasks like ray tracing denoising and video processing — introduces the need to minimize data movement between processors and increase data locality in the target processor. In this paper, we propose Shader Neural Network (ShaderNN), an OpenGL-based, fast, and power-efficient inference framework designed for mobile devices to address these challenges. Our contributions include the following: (1) the texture-based input/output provides an efficient, zero-copy integration with real-time graphics pipelines or image processing applications, thereby saving expensive data transfers between CPU and GPU, which are unavoidable in most existing inference engines; (2) we are the first to leverage fragment shaders based on the OpenGL backend in neural network inference operators, which has an advantage in deploying parametrically small neural network models; (3) a hybrid implementation of the compute shader and fragment shader is proposed that enables layer-level shader selection to boost performance; and (4) we utilize OpenGL features — such as normalization, interpolation and texture padding — to improve performance. Experiments illustrate the favorable performance of ShaderNN over other popular on-device deep learning frameworks such as TensorFlow-Lite on the latest mobile devices powered by Qualcomm and MediaTek chips. A case study further demonstrates the usability and integration of the ShaderNN framework with a media processing Android application seamlessly. ShaderNN is available open source at Github (https://github.com/inferenceengine/shadernn).

A decision support framework for best-fitting blockchain platform selection in sustainable supply chains under uncertainty

Despite blockchain’s potential to enhance visibility and traceability in sustainable supply chains (SCs), its adoption is complex due to the various criteria (e.g., interoperability and cost) required for the best-fitting platform selection. This study aims to investigate conflicting criteria in the blockchain technology (BT) platform selection process for decision-making under uncertainty. We propose a three-phase decision support framework to study BT adoption considering technological, organizational, and environmental contexts. In the first phase, after exploring the evaluation criteria from multiple contexts, the developed framework incorporates uncertainty and reliability to deal with the BT platform evaluation problem. Then, fuzzy cognitive map modeling, advanced by a Z-number-based inference system, is introduced to model the causal relationships between criteria. This is followed by implementing a hybrid learning algorithm to assess the impact of each criterion on adoption decisions. Finally, the fuzzy combined compromise solution embedded in the framework prioritizes BT platforms to identify the most suitable ones for sustainable SC. The findings imply that performance efficiency, implementation costs, maintainability and operability can significantly affect the BT platform selection decisions. The outcomes offer more stable, reliable, and distinguishable solutions for the proposed problem compared to the traditional approaches. The results introduce Hyperledger and R3 Corda as the best-fitting platforms for adoption based on the identified criteria.

Intelligent Fuzzy Traffic Signal Control System for Complex Intersections Using Fuzzy Rule Base Reduction

In this study, the concept of symmetry is employed to implement an intelligent fuzzy traffic signal control system for complex intersections. This approach suggests that the implementation of reduced fuzzy rules through the reduction method, without compromising the performance of the original fuzzy rule base, constitutes a symmetrical approach. In recent decades, urban and city traffic congestion has become a significant issue because of the time lost as a result of heavy traffic, which negatively affects economic productivity and efficiency and leads to energy loss, and also because of the heavy environmental pollution effect. In addition, traffic congestion prevents an immediate response by the ambulance, police, and fire brigades to urgent events. To mitigate these problems, a three-stage intelligent and flexible fuzzy traffic control system for complex intersections, using a novel hybrid reduction approach was proposed. The three-stage fuzzy traffic control system performs four primary functions. The first stage prioritizes emergency car(s) and identifies the degree of urgency of the traffic conditions in the red-light phase. The second stage guarantees a fair distribution of green-light durations even for periods of extremely unbalanced traffic with long vehicle queues in certain directions and, especially, when heavy traffic is loaded for an extended period in one direction and the short vehicle queues in the conflicting directions require passing in a reasonable time. The third stage adjusts the green-light time to the traffic conditions, to the appearance of one or more emergency car(s), and to the overall waiting times of the other vehicles by using a fuzzy inference engine. The original complete fuzzy rule base set up by listing all possible input combinations was reduced using a novel hybrid reduction algorithm for fuzzy rule bases, which resulted in a significant reduction of the original base, namely, by 72.1%. The proposed novel approach, including the model and the hybrid reduction algorithm, were implemented and simulated using Python 3.9 and SUMO (version 1.14.1). Subsequently, the obtained fuzzy rule system was compared in terms of running time and efficiency with a traffic control system using the original fuzzy rules. The results showed that the reduced fuzzy rule base had better results in terms of the average waiting time, calculated fuel consumption, and CO2 emission. Furthermore, the fuzzy traffic control system with reduced fuzzy rules performed better as it required less execution time and thus lower computational costs. Summarizing the above results, it may be stated that this new approach to intersection traffic light control is a practical solution for managing complex traffic conditions at lower computational costs.

Career Guide Expert System for Nigerian Secondary Schools Based on Howard Gardner Categorisation of Intelligence

Career Guide Expert System for Nigerian secondary schools is an applied research targeted at developing a web-based career guidance system for students transitioning from the Universal Basic Education (UBE) Scheme to Senior secondary school (SSS). The transitioning process requires students to choose among groups of subjects (Science, Art, or Commercial) to study in SSS which ideally requires the services of counsellors. However, there is a shortage of counsellors in schools. This research will significantly address the problem of shortage of professional counsellors and inefficient associated manual career guidance, which can negatively affect the career trajectory of students. Three parameters (Howard Gardner Intelligence Categorization, Academic performance and student aspirations) were used in determining which of the three alternatives (Science, Art, and Commercial) suits a student the best. The inference engine of the Expert System was implemented using a weighted sum multiple decision criteria model to determine which alternative suits a student the most based on the data generated from the quiz attempted by a student. The system was validated by a professional counsellor who subjected the system to a test with twenty samples of student data and certified that the system was accurate. Thus, if this system is incorporated into the Nigerian educational system, it will revolutionize the career guidance process, leading to better performance of students and future career satisfaction.

Internet of Conscious Things: Ontology-Based Social Capabilities for Smart Objects

Emerging distributed intelligence paradigms for the Internet of Things (IoT) call for flexible and dynamic reconfiguration of elementary services, resources and devices. In order to achieve such capability, this paper faces complex interoperability and autonomous decision problems by proposing a thorough framework based on the integration of the Semantic Web of Things (SWoT) and Social Internet of Things (SIoT) paradigms. SWoT enables low-power knowledge representation and autonomous reasoning at the edge of the network through carefully optimized inference services and engines. This layer provides service/resource management and discovery primitives for a decentralized collaborative social protocol in the IoT, based on the Linked Data Notifications(LDN) over Linked Data Platform on Constrained Application Protocol (LDP-CoAP). The creation and evolution of friend and follower relationships between pairs of devices is regulated by means of novel dynamic models assessing trust as a usefulness reputation score. The close SWoT-SIoT integration overcomes the functional limitations of existing proposals, which focus on either social device or semantic resource management only. A smart mobility case study on Plug-in Electric Vehicles (PEVs) illustrates the benefits of the proposal in pervasive collaborative scenarios, while experiments show the computational sustainability of the dynamic relationship management approach.

A smart system for student performance assessment (SPA)

The proposed smart system for Student Performance Assessment (SPA) is a system that evaluates students' knowledge and skill attainment in a specific course by measuring their achievements of the Course Learning Outcomes (CLOs). The instructor defines the aspects, weights, and rating scale used by SPA to analyze each course. The system calculates the average of students' marks in each learning outcome and compares them with the CLO targets and scores to determine the effectiveness of the teaching and learning methods used. The system uses facts and rules extracted from the course syllabus and Bloom's Taxonomy to build its knowledge base. This paper presents the development of the SPA inference engine, which is used to find CLO targets based on the course level. The inference engine uses efficient procedures and a prediction process to determine the correct target and score, providing a reliable and understandable methodology for reasoning about the information in the knowledge base and formulating conclusions. SPA is a highly responsive and intelligent system that can be a valuable tool for measuring students' achievements. Its characteristics include high performance, reliability, and intelligibility, and its combination of cognitive systems and cognitive theory has led to remarkable progress in measuring student performance. Limitations include dependency on accurate course content and initial setup time, potential bias in CLO weight assignments, challenges in integrating SPA with existing institutional databases, the need for continuous updates to the knowledge base to reflect curriculum changes, and potential resistance from educators to adopt new technologies. Future improvements could involve adaptive learning integrations, enhanced user interfaces, and broader applicability across diverse educational settings.

Pemodelan Sistem Pakar Identifikasi Tingkat Stres Karyawan dengan Metode Certainty Factor

Expert systems work by collecting knowledge from experts in the form of rules or knowledge bases, then using an inference engine to analyze situations and produce decisions or suggestions. Apart from that, it is often used in the fields of medicine, engineering, finance, etc., related to this, especially the process of identifying the level of stress experienced by employees in the company so that they can find out more, therefore an expert system application design is needed for identification using the method. certainty factor, based on the certainty factor which has a concept based on symptoms and diagnoses from the weight of an expert's value and the user's value, calculated using the formula in CF, the final result is the creation of an expert system application for identifying employee stress levels based on a website.

Implementation of A Fuzzy Logic for Early Detecting of A Pregnant Women Risk of Hypertension

Hypertension is one of the diseases that can cause fatal effects to health, especially for pregnant women. Hypertension during pregnancy can cause serious effects such as pre-eclampsia and eclampsia which can threaten the lives of pregnant women and fetuses. The purpose of this study is to determine the input and output variables related to hypertension during pregnancy and used for analysis using Tsukamoto FIS to determine risk factors for hypertension in pregnant women. This study uses input variables in the form of maternal age, systolic blood pressure (SBP), diastolic blood pressure (DBP), history of hypertension and genetic. Furthermore, it is analyzed using fuzzy logic through the process of fuzzyfication, inference engine and defuzzyfication. The examination of data samples of pregnant women with systolic and diastolic blood pressure categories included in the high category or potentially have hypertension. The Z value (defuzzyfication) shows that the output is in the category of severe hypertension so that the patient needs immediate treatment by medical personnel. The results showed that the risk of a mother having hypertension can be obtained through Tsukamoto FIS in the form of concrete values that can describe risk factors in the form of normal, hypertension and severe hypertension as well as recommended actions related to these risk factors.

Development of fuzzy inference engine for prediction of mechanical properties in TIG-welded Al-65,032 alloy through reduced experimentation using Taguchi method

Development of fuzzy inference engine for prediction of mechanical properties in TIG-welded Al-65,032 alloy through reduced experimentation using Taguchi method

Reducing the Effect of Uncertainty in Mamdani–Type Fuzzy Inference System Using Qubits

The success of fuzzy logic in effectively capturing subjective knowledge which are uncertain or ambiguous in nature lies on the fact that fuzzy sets assigns a degree of membership to each linguistic variable of a crisp value. As a condition the degree of membership values to all linguistic variables should add up to one. This abstract condition is neither probabilistic nor realistic in quantum mechanics. Probabilistic is in a sense that the normalization condition of quantum mechanics isn’t satisfied. Realistic is in a sense that the condition isn’t physically proven experimentally. In this paper, a novel quantum Mamdani-type fuzzy inference system is proposed. The proposed system consists of quantum fuzzification, quantum fuzzy inference and quantum defuzzification steps. Instead of assigning degree of membership values for linguistic variables, the proposed inference engine assigns a quantum state for a crisp value. Hence, a crisp input is fuzzified into superposition of linguistic variables represented by qubits. The proposed quantum Mamdani-type fuzzy inference system is tested to captures expert knowledge more realistically than the Mamdani-type fuzzy inference system, is thus less vulnerable to effects of uncertainties.