Rule-based Knowledge Research Articles

Collaborative Learning With Heterogeneous Local Models: A Rule-Based Knowledge Fusion Approach

Collaborative Learning With Heterogeneous Local Models: A Rule-Based Knowledge Fusion Approach

Modeling and simulation of Fuzzy-rule based WBAN using multi-level fuzzy colored petri-nets and reinforcement learning

By developing IoTs (Internet of Things), a new concept is arising in WBANs (Wireless Body Area Networks), called IoBs (Internet of Bodies). WBANs are the safety-critical devices that use rule-based knowledge with fuzzy variables for monitoring patients. There are two issues with the knowledgebase: (1) appropriate knowledge representation and inference and (2) handling the situations for which there is no rule in the knowledgebase. In this paper, for the first issue we present MFCPN (Multi-level Fuzzy Colored Petri-nets) and for the second one we present a RL (Reinforcement Learning) mechanism for decision making based on the past device’s feedback to the environment. Through a few scenarios and two datasets, the proposed approach has been applied to pacemakers to show the approach's effectiveness. The comparison between decision making by our approach and two real datasets indicates the proposed approach has 79.3 % accuracy.

Debt Collection Model for Mass Receivables Based on Decision Rules—A Path to Efficiency and Sustainability

Debt collection companies buy overdue debts on the market in order to collect them and recover the highest possible amount of a debt. The pursuit of debt recovery by employees of collection agencies is a very demanding task. The aim of the article is to propose a rule-based model for managing the process of mass debt collection in a debt collection company, which will make the debt collection process more efficient. To achieve this, we have chosen a decision tree as a machine learning technique best suited for creating rules based on extensive data from the debt collection company. The classification accuracy of the decision tree, regardless of the possibility of acquiring rule-based knowledge, proved to be the highest among the tested machine learning methods, with an accuracy rate of 85.5%. Through experiments, we generated 16 stable rules to assist in the debt collection process. The proposed approach allows for the elimination of debts that are difficult to recover at the initial stage of the recovery process and to decide whether to pursue amicable debt collection or to escalate the debt recovery process to legal action. Our approach also enables the determination of specific actions during each stage of the proceedings. Abandoning certain actions or reducing their frequency will alleviate the burden on collection agency employees and help to avoid the typical burnout associated with this line of work. This is the path to making the organizational culture of a collection agency more sustainable. Our model also confirms the possibility of using data from debt collection companies to automatically generate procedural rules and automate the process of purchasing and collecting debts. However, this would require a larger set of attributes than what we currently possess.

Expert System for Diagnosing Gastric Disease Using the Forward Chaining Method

Health is the most crucial aspect of life, as it signifies a state of well-being and vitality. Gastric diseases pose a significant health concern, often requiring prompt and accurate diagnosis for effective treatment. In this study, an expert system utilizing the forward chaining method is developed to aid in diagnosing gastric diseases. The system aims to streamline the diagnostic process, providing users with timely and reliable results to facilitate early intervention and management of stomach disorders. Through the integration of technology and medical expertise, this system offers a valuable tool in enhancing healthcare delivery and improving patient outcomes in the realm of gastric health. The research findings highlight the effectiveness of an expert system employing the forward chaining method in diagnosing gastric diseases. The system demonstrates the capability to efficiently identify and assess symptoms related to stomach disorders, offering users a reliable and accessible platform for obtaining accurate diagnostic results. By leveraging current knowledge base rules and user-provided symptom data, the system enhances the diagnostic process, potentially reducing both time and costs associated with diagnosing gastric diseases. This study underscores the significance of technological advancements in healthcare, particularly in the realm of gastric health, by providing a valuable tool for early detection and management of stomach disorders.

Integrating regular expressions into neural networks for relation extraction

Relation extraction aims to identify semantic relationships between entities from the given sentences. The development of deep learning prompts a variety of neural-based relation extraction models. These models can learn features from sentences automatically, but controlling their learned patterns is equally difficult. Therefore, it is hard to take steps to improve their performance further. A possible way to modify the models is by utilizing human-defined rules to introduce rule-based knowledge and guiding the models to capture the desired features based on human intent. As an efficient human knowledge, regular expression (RE) is helpful in extracting particular targets from text, which is also logical and interpretable. A novel embedding method is proposed to transform REs into learnable representations available to neural networks (NNs). Then, the rule-based representations are investigated to be integrated into different levels of NNs. Three representative relation extraction models based on recurrent neural network (RNN), convolutional neural network (CNN), and Transformer encoder, respectively, are selected as baselines to incorporate our rule-based representations. The experimental results show that the models with rule-based information outperform their original ones on the F1-score by 3.75%, 4.06%, and 2.57%, and the identification accuracy of complex samples is also significantly improved, verifying the effectiveness of our approaches. Our work is easy to reproduce and can be generally applied to various neural networks, proving that data-driven relation extraction methods can take advantage of the rule-based knowledge provided by REs.

ПИТАННЯ ФОРМУВАННЯ МУЛЬТИ-НЕЧІТКИХ МОДЕЛЕЙ І ПРОЦЕДУР СИСТЕМ МОНІТОРИНГУ ПРОГРАМНОГО ЗАБЕЗПЕЧЕННЯ

The article details issues related to the features, tasks and the development of fuzzy models and intelligent procedures designed for software monitoring systems. At the same time, improved forms of fuzzy control models, called multi-fuzzy, are proposed. They generalize the methods of fuzzy modeling in the structure of rules and procedures by representing the dependence of the rules in the knowledge base, establishing various models of input data and applying various fuzzy inference methods for individual knowledge base rules.

Autonomous Behavior Selection For Self-driving Cars Using Probabilistic Logic Factored Markov Decision Processes

ABSTRACT We propose probabilistic logic factored Markov decision processes (PL-fMDPs) as a behavior selection scheme for self-driving cars. Probabilistic logic combines logic programming with probability theory to achieve clear, rule-based knowledge descriptions of multivariate probability distributions, and a flexible mixture of deductive and probabilistic inferences. Factored Markov decision processes (fMDPs) are widely used to generate reward-optimal action policies for stochastic sequential decision problems. For evaluation, we developed a simulated self-driving car with reliable modules for behavior selection, perception, and control. The behavior selection module is composed of a two-level structure of four action policies obtained from PL-fMDPs. Three main tests were conducted focused on the selection of the appropriate actions in specific driving scenarios, and the overtaking of static obstacle vehicles and dynamic obstacle vehicles. We performed 520 repetitions of these tests. The self-driving car completed its task without collisions in 99.2% of the repetitions. Results show the suitability of the overall self-driving strategy and PL-fMDPs to construct safe action policies for self-driving cars.

Computer programmers show distinct, expertise-dependent brain responses to violations in form and meaning when reading code

As computer programming becomes more central to the workforce, the need for better models of how it is effectively learned has become more apparent. The current study addressed this gap by recording electrophysiological brain responses as 62 Python programmers with varying skill levels read lines of code with manipulations of form (syntax) and meaning (semantics). At the group level, results showed that manipulations of form resulted in P600 effects, with syntactically invalid code generating more positive deflections in the 500–800 ms range than syntactically valid code. Meaning manipulations resulted in N400 effects, with semantically implausible code generating more negative deflections in the 300–500 ms range than semantically plausible code. Greater Python expertise within the group was associated with greater sensitivity to violations in form. These results support the notion that skilled programming, like skilled natural language learning, is associated with the incorporation of rule-based knowledge into online comprehension processes. Conversely, programmers at all skill levels showed neural sensitivity to meaning manipulations, suggesting that reliance on pre-existing semantic relationships facilitates code comprehension across skill levels.

Construction of knowledge constraints: a case study of 3D structural modeling

The uncertainty of structural interpretation complicates the practical production and application of data-driven complex geological structure modeling technology. Intelligent structural modeling excavates and extracts structural knowledge from structural interpretation through human–machine collaboration and combines structural interpretation to form a new model of complex structural modeling guided by knowledge. Specifically, we focus on utilizing knowledge rule reasoning technology to extract topological semantic knowledge from interpretive data and employ knowledge inference to derive structural constraint information from complex geological structure models, thus effectively constraining the 3D geological structure modeling process. To achieve this, we develop a rule-based knowledge inference system that derives theoretical models consistent with expert cognition from interpretive data and prior knowledge. Additionally, we represent the extracted knowledge as a topological semantic knowledge graph, which facilitates computer recognition and allows estimation of intersection lines during 3D geological modeling, resulting in the creation of accurate models. The applicability of our proposed method to various complex geological structures is validated through application tests using real-world data. Furthermore, our method effectively supports the realization of intelligent structure modeling in real working area.

The role of metalinguistic knowledge in third-language development among heritage and late bilinguals

Aims and objectives: This study examined whether bilinguals’ metalinguistic knowledge (MK) across both of their languages as well as heritage/late bilingual experience contributed to the initial development of third-language (L3) morphosyntax using Japlish, a semi-artificial language. Methodology: 48 heritage and 63 late English-Spanish bilinguals were exposed aurally to Japlish sentences containing word order patterns and case markers. Participants completed an acceptability judgment task (AJT) and a picture-matching task (PMT) immediately after exposure and again 2 weeks later as well as English and Spanish MK tests. Data and analysis: Logistic mixed effects analyses were run on AJT and PMT accuracy scores, grammatical/ungrammatical items, linguistic structure, and time of testing. MK scores and heritage/late bilingual status were included as predictors in the statistical model. Findings: MK significantly contributed to overall AJT scores, including grammatical and ungrammatical items and time of testing. MK also contributed to ungrammatical word order and case markers. For the PMT, MK significantly contributed to scores, especially during delayed testing. A weak correlation emerged between participants’ MK across both of their languages and their rule-based knowledge of Japlish linguistic structures. Heritage and late bilingual experiences did not significantly contribute to these results. Originality: This study included two different bilingual groups to investigate how variation in bilingual experience shapes additional language learning. It also advances our knowledge on the role of MK in the learning of linguistic structures not involved in transfer. Furthermore, a more nuanced examination of MK is considered as it relates to type of assessment, linguistic structure, and time of testing. Significance: The findings highlight the critical role of MK for both bilingual populations and its facilitative role in the development of initial L3 morphosyntax. The results are interpreted under the lens of structural sensitivity theory.

Interval Type-2 Mutual Subsethood Cauchy Fuzzy Neural Inference System (IT2MSCFuNIS)

An interval type-2 (IT2) mutual subsethood Cauchy fuzzy neural inference system has been proposed in this paper. The network architecture consists of 3-layers with all connection weights being IT2 Cauchy fuzzy membership functions (CMFs). The crisp inputs to the system are fuzzified into IT2CMFs with fixed centers and uncertain spreads. The hidden layer represents the rule-based knowledge. The firing degree of the antecedent part of each rule at the hidden layer is computed by aggregating the product of the mutual subsethood similarity measures between the inputs and the connection weights. A volume defuzzification is used to compute the numeric output. A gradient descent back-propagation algorithm is used to train the model. The novelty of the proposed model is threefold. First, is enriching the theory of the mutual subsethood fuzzy neural models by adopting the Cauchy membership function (CMF) as another powerful fuzzy basis function (FBF) rather than the classical choice of Gaussian fuzzy membership functions (GMFs). Second, is the success of computing the mutual subsethood similarity measure between the IT2CMFs and all the model parameters’ updating equations in analytic closed-form formulas, not numerically or approximately. Third, is the ability to extract the type-1 (T1) mutual subsethood Cauchy fuzzy neural inference system (T1MSCFuNIS) with all its analytic closed-form formulas directly as a special case from the general formulas of IT2MSCFuNIS model. Such a novelty makes the proposed model a concrete and effective development of the theory of mutual subsethood fuzzy neural models. Both IT2MSCFuNIS and T1MSCFuNIS models have been tested using different examples from the domains of function approximation, classification, and prediction. The results ensure the efficacy of both models compared with other models reported in the literature.

Fuzzy Logic Speed Regulator for D.C. Motor Tuning

A D.C. motor's rotational speed is regulated in this study using a PID controller and a fuzzy logic controller. In contrast to the fuzzy logic controller, which uses rules based on knowledge and experience, the proportional-integral-derivative (PID) controller requires a mathematical system model. This study investigates the regulation of a DC motor's velocity using PID and fuzzy logic controllers. The PID controller utilizes a mathematical model and parameter tuning by trial and error. Still, the fuzzy logic controller (FLC) operates on rule-based knowledge, enabling it to handle the nonlinear features of the DC motor effectively. The FLC design entails intricate determinations, including the establishment of a rule base and the process of fuzzification. A total of 49 fuzzy rules have been devised to achieve precise control. Based on MATLAB/SIMULINK simulations, the study concludes that the Fuzzy Logic Controller (FLC) beats the Proportional-Integral-Derivative (PID) controller. The FLC exhibits superior transient and steady-state responses, shorter response times, reduced steady-state errors, and higher precision. This study emphasizes the efficacy of the FLC (Fuzzy Logic Controller) in dealing with the difficulties associated with DC motor control. It presents a strong argument for the suitability and efficiency of FLCs in industrial environments compared to conventional PID (Proportional-Integral-Derivative) controllers. There are a wide variety of ways to construct a fuzzy logic controller. The speed error and the rate of change in the speed error are two inputs to the FLC. Defuzzification is done by focusing on the core of the problem. The results show that FLC is superior to PID controllers in efficiency and effectiveness due to its reduced transient and steady-state factors.

Intelligent Technology Assessment of High-Speed Railway Based on Knowledge Graphs

This paper introduces an intelligent technology assessment framework for high-speed railways based on a knowledge graph approach. We employ rule-based knowledge extraction algorithms and a Bert-BiLSTM-CRF model for entity extraction from technical texts. Subsequently, we establish relationships among various entities, constructing a knowledge graph specific to high-peed railways. The knowledge graph is stored in a Neo4j graph database in triple format. Furthermore, we establish a comprehensive evaluation metric system, integrating knowledge graph insights to assess the utility of enabling technologies. We employ the CRITIC weighting method to calculate the value assessment results for target technologies. Simulation results indicate that the training results for word segmentation and sentence splitting are favorable, and the Bert-BiLSTM-CRF model achieves convergence in accuracy, recall, and F1 score after 30 iterations. The technical assessment results in this paper align closely with the actual technological value assessment.

ФОРМИРОВАНИЕ И РЕДУКЦИЯ БАЗЫ ЗНАНИЙ ДЛЯ ОЦЕНКИ ЦЕЛЕСООБРАЗНОСТИ БУРЕНИЯ СКВАЖИН НА НЕФТЯНЫХ МЕСТОРОЖДЕНИЯХ

The article discusses the issues of forming and reducing a fuzzy knowledge base designed to assess the feasibility of drilling wells in oil fields as part of a decision support system. The initial knowledge base was formed using a specially developed software package that implements neuro-fuzzy modeling technology based on a group of fuzzy neural networks. To train them and form a knowledge base, initial data were required that described the dependence of the influence of the characteristics of producing wells on their efficiency in terms of the production rate of oil produced. The initial data for the analysis of 3187 records were presented for 46 oil field objects with carbonate reservoirs. Each record was characterized by the values of the input parameters “absolute rock permeability”, “rock porosity”, “average flow rate of neighboring wells” and “minimum water cut of neighboring wells”. The output parameter determined the efficiency of the well as the average oil flow rate for the first 3 years of operation. When training a team of 5 fuzzy neural networks, 5 gradations of input parameters were selected. The total number of fuzzy rules in the original knowledge base was 319. To assess its classification ability, the values of the Accuracy, Precision, Recall and F1-Score metrics were calculated. To reduce the knowledge base, a specially developed software package was used that implements 2 stages of genetic optimization: structural-parametric transformations of the fuzzy rules of the original knowledge base to obtain an intermediate knowledge base and minimizing the composition of the fuzzy rules of the intermediate knowledge base to obtain the desired (reduced) knowledge base. As a result of the reduction, the quantitative composition of knowledge bases, as well as the values of classification quality metrics, changed. The number of rules in the intermediate knowledge base was 131, and in the reduced one - 33. The overall reduction in the volume of the knowledge base was 89.66%. After the 1st stage of reduction, the values of the Accuracy, Precision, Recall and F1-Score metrics on the test data set increased, respectively, by 7.88%, 6.8%, 0.34% and 0.53%. After the 2nd stage of reduction, the overall increase in the values of these metrics was, respectively, 9.02%, 7.55%, 1.58% and 1.68%. The results obtained were tested at Tatneft and used to form and optimize the knowledge base when developing a decision support system for assessing the feasibility of drilling wells in oil fields.

Reasons for using parallel activation of logical rules in solving management tasks

The complexity of the subject areas in which intelligent information systems operate is steadily increasing. The tasks set for such systems are increasingly aimed at automating and robotizing spheres of human activity. The solution of such problems requires adaptive and flexible methods capable of taking into account dynamic changes in the environment in real time. The mivar approach to creating intelligent decision-making systems allows working with adaptive discrete structures and provides methods for making management decisions based on adaptive active logical inference based on the knowledge base of mivar rules. The mivar machine of logical inference is the core of expert systems based on the mivar approach. As a result of the historical development of the mivar approach when working on different subject areas, various versions of mivar machines of logical inference with their algorithms for bypassing rules in the knowledge base were obtained. This paper discusses the reasons for the emergence of such options and demonstrates the need to use a parallel algorithm for activating rules in an adaptive network of logical rules when solving problems of action planning in technical systems.

Text processing using LLM for automatic creation of agricultural crops knowledge bases

The complexity of subject areas in which intelligent information systems operate is steadily increasing. Tasks assigned to smart agriculture systems are increasingly focused on automating and robotizing areas of human activity. Solving such tasks requires adaptive and flexible methods capable of accommodating dynamic changes in the environment in real-time. The mivar approach to creating intelligent decision-making systems enables working with adaptive discrete structures and provides methods for managerial decision-making based on adaptive active logical inference from the mivar rule knowledge base. The mivar logical inference machine forms the core of expert systems based on the mivar approach. As a result of the development of the mivar approach across various subject areas, different versions of mivar logical inference machines with their algorithms for rule traversal in the knowledge base have been created. Recent advancements in artificial intelligence and machine learning have opened new opportunities for enhancing the mivar approach. The integration of large language models for automating text processing in mivar systems significantly enhances the accuracy and efficiency of decision-making processes based on expert systems for sustainable agriculture. This paper demonstrates the feasibility of using automated text processing, intended for human training, through large language models, and its subsequent application in action planning tasks within technical systems. The proposed methodology is aimed at creating extensive knowledge bases based on textual information for real-time monitoring and decision-making in smart agriculture systems.

An advanced control system for nitrogen removal and energy consumption optimization in full-scale wastewater treatment plants

The development and implementation of advanced control systems enable the successful and efficient conversion of wastewater treatment plants into water resource recovery facilities. This work presents an advanced control system based on low-cost sensors (pH, ORP and DO) which enhances biological nitrogen removal and improves energy efficiency. It consists of three independent fuzzy-logic-based control loops (for nitrification, simultaneous nitrification-denitrification, and denitrification, respectively) combined with a set of knowledge-based rules. The control system was implemented in three facilities and test-run for over two years. During this time, proper process performances were achieved both in terms of water quality and of operating cost savings. Energy consumption, calculated as kWh per kg COD equivalent removed (including nitrogen removal), was reduced between 14 % and 33 %.

A Hybrid Model for 30-Day Syncope Prognosis Prediction in the Emergency Department.

Syncope is a challenging problem in the emergency department (ED) as the available risk prediction tools have suboptimal predictive performances. Predictive models based on machine learning (ML) are promising tools whose application in the context of syncope remains underexplored. The aim of the present study was to develop and compare the performance of ML-based models in predicting the risk of clinically significant outcomes in patients presenting to the ED for syncope. We enrolled 266 consecutive patients (age 73, IQR 58-83; 52% males) admitted for syncope at three tertiary centers. We collected demographic and clinical information as well as the occurrence of clinically significant outcomes at a 30-day telephone follow-up. We implemented an XGBoost model based on the best-performing candidate predictors. Subsequently, we integrated the XGboost predictors with knowledge-based rules. The obtained hybrid model outperformed the XGboost model (AUC = 0.81 vs. 0.73, p < 0.001) with acceptable calibration. In conclusion, we developed an ML-based model characterized by a commendable capability to predict adverse events within 30 days post-syncope evaluation in the ED. This model relies solely on clinical data routinely collected during a patient's initial syncope evaluation, thus obviating the need for laboratory tests or syncope experienced clinical judgment.

Logical Rule-Based Knowledge Graph Reasoning: A Comprehensive Survey

With its powerful expressive capability and intuitive presentation, the knowledge graph has emerged as one of the primary forms of knowledge representation and management. However, the presence of biases in our cognitive and construction processes often leads to varying degrees of incompleteness and errors within knowledge graphs. To address this, reasoning becomes essential for supplementing and rectifying these shortcomings. Logical rule-based knowledge graph reasoning methods excel at performing inference by uncovering underlying logical rules, showcasing remarkable generalization ability and interpretability. Moreover, the flexibility of logical rules allows for seamless integration with diverse neural network models, thereby offering promising prospects for research and application. Despite the growing number of logical rule-based knowledge graph reasoning methods, a systematic classification and analysis of these approaches is lacking. In this review, we delve into the relevant research on logical rule-based knowledge graph reasoning, classifying them into four categories: methods based on inductive logic programming (ILP), methods that unify probabilistic graphical models and logical rules, methods that unify embedding techniques and logical rules, and methods that jointly use neural networks (NNs) and logical rules. We introduce and analyze the core concepts and key techniques, as well as the advantages and disadvantages associated with these methods, while also providing a comparative evaluation of their performance. Furthermore, we summarize the main problems and challenges, and offer insights into potential directions for future research.

A novel continual reinforcement learning-based expert system for self-optimization of soft real-time systems

Virtual globes are soft real-time systems, which stream multi-resolution data sets and render world-scale landscapes in real-time. Such systems require an adaptation mechanism to fulfill user and quality of experience (QoE) requirements. Configuration and tuning of parameters to optimize system performance is challenging due to run-time uncertainties. In this work, we propose a continual reinforcement learning-based expert system for self-optimization of a soft real-time system to meet user and quality objectives. The proposed system is capable of learning continuously by utilizing the previously learned knowledge to fulfill adaptation requirements. The proposed managing system interacts with a managed system and evolves a knowledge base of adaptation rules for run-time optimization of the managed system to meet user and quality objectives. We devise a learning classifier system for the evolution, storage, and transfer of adaptation rules using a code fragments-based rich encoding scheme. We transform user goals, system states, and execution environment conditions into observable states. We validated the proposed system for the same and cross-domain knowledge transfer and reuse scenarios for different user goals and execution environments. The results of the proposed system are compared with baseline and state-of-the-art methods where the proposed system outperformed the methods in terms of adaptation accuracy.