Proceedings of the Seventh International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’23): Volume 1 (Lecture Notes in Networks and Systems)

This book contains the works connected with the key advances in Industrial Artificial Intelligence presented at IITI 2023, the Seventh International Scientific Conference on Intelligent Information Technologies for Industry held on September 25-30, 2023 in St. Petersburg, Russia. The works were written by the experts in the field of applied artificial intelligence including topics such as Machine Learning, Explainable AI, Decision-Making, Fuzzy Logic, Multi-Agent and Bioinspired Systems. The following industrial application domains were touched: railway automation, cyber security, intelligent medical systems, navigation and energetic systems. The editors believe that this book will be helpful for all scientists and engineers interested in the modern state of applied artificial intelligence. Preface Organization Contents Invited Papers Intelligent Interfaces and Systems for Human-Computer Interaction 1 Introduction 2 Intelligent Synthesis 2.1 Video Modality 2.2 Audio Modality 2.3 Text Modality 2.4 Multimodality 3 Intelligent Analysis 3.1 Video Modality 3.2 Audio Modality 3.3 Text Modality 3.4 Multimodality 4 Conclusions References Three Knowledge Sources and Three Constituents of Artificial Intelligence Foundation 1 Introduction 2 Data vs. Knowledge 3 Data Science 4 From Knowledge Engineering to Knowledge Science: A Perspective 5 Digital Twin As a Source of Various Knowledge 6 Discussion: Integration and Interaction of Knowledge from Different Types of Sources 7 Conclusion References Machine Learning and Its Applications Comparative Analysis of Data Synthesis Methods for Prognostic Models Development in Cardiology 1 Introduction 2 Related works 3 Methods and materials 4 Results 5 Discussion 6 Conclusion References Artificial Intelligence Approach to Palladium Nanocatalysts Diagnostics Automation 1 Introduction 2 Model-Based Training DRL Agents 3 Optimal Control Algorithms and Spectral Profile Analysis 4 Conclusion References Methodology for Detecting and Feature Selection of an Information Attack in the Process of Mediatization 1 Introduction 2 Related Work 3 Methodology for Detection and Feature Selection 3.1 Information Attack Model 3.2 Approach to Detection and Feature Selection 4 Case Study 5 Conclusion References Planning Maneuvers for Autonomous Driving Based on Offline Reinforcement Learning: Comparative Study 1 Introduction 2 Related Work 3 Background 3.1 Offline RL 3.2 Conservative Q-learning 4 Approach 4.1 CommonRoad Scenarios 4.2 Trajectories Generation 5 Experiments 6 Conclusion References Big Five: What User Posts Say? 1 Introduction 1.1 Prerequisites for Research 1.2 Big Five 1.3 Related Work 2 Problem Statement 3 Methods 3.1 Description Dataset 3.2 Data Preprocessing 3.3 Used Models 3.4 Methodology for Evaluating the Results of Experiments 3.5 Results 4 Discussion 5 Conclusion References Gated Recurrent Unit Autoencoder for Fault Detection in Penicillin Fermentation Process 1 Introduction 2 Literature Review 3 Domain Adaptation 4 Deep Learning Models 4.1 Gated Recurrent Unit 4.2 GRU Based Autoencoder 4.3 GRU-AE Based Fault Detection 5 Experiments and Results 5.1 Dataset 5.2 Data Preprocessing 5.3 Model Training 5.4 Results 6 Conclusion References Resume Recommendation using RNN Classification and Cosine Similarity 1 Introduction 2 Related Works 2.1 Resume Classifier Related Works 2.2 Resume Recommender Related Works 3 The Proposed Two-Fold Algorithm Approach 3.1 Resume Classification 3.2 Experimental Results and Evaluation 4 Resume Recommendation 4.1 Similarity Function 4.2 Computing Similarity Function 5 Discussion and Conclusion References Machine Learning for Adaptive Analysis and Evaluation of Soil Slopes 1 Introduction 2 Background 3 Method 3.1 The Model of Cluster Analysis 3.2 BCubed Clustering Quality Metric 3.3 Agglomerative Clustering Algorithm 4 Results 4.1 Pre-processing of the Data Frame 4.2 Field Segmentation 4.3 Formation of ``Winners'' and ``Losers'' 4.4 Clustering 4.5 Angle Calculation 4.6 Quality Assessment 5 Proposed Decision 6 Conclusion and Future Work References Research on Video Pedestrian Tracking Based on the Combination of Optical Flow Method and Target Tracking Network 1 Introduction 1.1 Tracking Model Based on Traditional Methods 1.2 Tracking Model Based on Deep Learning Methods 2 Related Work 2.1 Siamese-Based Tracking Model 2.2 Self-attention 2.3 Variational Optical Flow Method 3 Transformer-Based Target Tracking Model 3.1 Overall Framework of SiamSA Tracking Model 3.2 Variational Optical Flow Model Based on Deep Learning Priors 3.3 Optical Flow Vector Correction Search Area 4 Experiments 4.1 Effect of Self-attention 4.2 Improved Optical Flow Model 4.3 Effect of Search Area Update Model 5 Conclusion References Development and Testing Intelligent Video Surveillance Systems Based on the CNN Algorithm 1 Introduction 2 Analog Overview 2.1 Solutions Based on OpenCV Methods 2.2 Convolutional Neural Network Models 2.3 Determining the Accuracy of Forecast Data 3 Suggested Solutions 3.1 Object Detection Based on OpenCV and Convolution Neural Network 3.2 Tracking and Dynamic Identity Stack Creation Algorithm 4 Computer Simulation Scenarios as Datasets 5 Experimental Results 5.1 Testing the Object Detection Algorithm 5.2 Testing the Tracking Algorithm 6 Conclusion References Impact of Loss Functions on the Training of LiDAR-based Place Recognition Models 1 Introduction 2 Related Work 2.1 LiDAR-based Place Recognition Methods 2.2 Loss Functions in Place Recognition 3 Methodology 3.1 Loss Function 3.2 Methods 3.3 Training Details 4 Experiments 4.1 Datasets 4.2 Evaluation Metrics 4.3 Results and Discussion 5 Conclusion References Neural Attention Forests: Transformer-Based Forest Improvement 1 Introduction 2 Preliminaries 2.1 Nadaraya-Watson Regression and Attention 2.2 Attention-based Random Forest 3 The Neural Attention Forest Architecture 4 The Neural Attention Forest as a Transformer 5 Numerical Experiments 6 Concluding Remarks References Audio-Visual Multi-modal Meeting Recording System 1 Introduction 1.1 Background 1.2 Main Contribution 1.3 Paper Structure 2 Training Strategy 2.1 Contrast Learning 2.2 Domain Generalization 3 Dataset 3.1 Dataset of AVSR 3.2 Dataset of SPR 4 Models 4.1 AVSR Model 4.2 SPR Model 5 Experiment 5.1 AVSR Model Training 5.2 SPR Model Training 5.3 Test 6 System Deployment 6.1 User Interface 6.2 Function Realization 7 Conclusion References Evolutional Modeling Studying the Efficiency of Parameter Scaling in Optimal Control Problems with Parallel Memetic Algorithm 1 Introduction 2 Problem Formulation 3 Gasoline’s Catalytic Reforming Reaction 4 Memetic Parallel Mind Evolutionary Computation Algorithm 5 Computational Experiments 5.1 Study of the Parameter Scaling Efficiency 5.2 Analysis from the Chemical Perspective 5.3 Analysis from the Optimization Perspective 6 Conclusions References Canonical Representation of Transport Networks and Their Identification Based on Evolutionary Modeling 1 Introduction 2 Operations on Basic Structures of Transport Networks 2.1 Serial Connection 2.2 Adding a Feedback Loop 2.3 Parallel Connection 3 Model Parameter Identification Method 4 Learning Theory Methods in Transport Network Identification 5 Conclusion References Modified Adaptive Particle Swarm Algorithm 1 Introduction 2 Approach to the Representation of Solutions in an Algorithm based on Swarm Intelligence 3 Modified Algorithm for the Adaptive Behavior of a Bee Colony 4 Hybridization of the Structure of Swarm Intelligence 5 Experimental Studies 6 Conclusion References Development and Research of Algorithms for the Synthesis of Combinational Logic Circuits Based on the Evolutionary Approach 1 Introduction 2 Problem Statement 3 Evolutionary Algorithm for Synthesizing Combinational Circuits 4 Results of Computational Experiments 5 Conclusions References Fuzzy Models Co-active of Fuzzy Temporal Ontological Models and Fuzzy Temporal Cognitive Models for the Analysis and Forecasting of Complicated Systems 1 Introduction 2 Fuzzy Temporal Ontological Model 3 Fuzzy Relational Temporal Cognitive Model 4 Forecast Assessment of the Condition and Risks of the IEMS Malfunction 5 Conclusion References Synthesis of Intelligent Tracking Filter with Fuzzy for Parameter Setting in Problems of Air Traffic Management Automation 1 Introduction 2 Formulation of the Problem 3 Construction of the Equation of the Target Motion Based on the Condition for the Maximum of the Generalized Power Function for a Discrete Time Setting 4 The Synthesis of the Intelligent Tracking Filter 5 Mathematical Simulation 6 Conclusion References Intelligent Decision-Making Making Diagnostic Decisions Based on the Assessment of Mixed Production Rules 1 Introduction 2 Development of Mixed Production Rules 3 Application of a Complex Approach to MPR for Diagnosing the Asynchronous Electric Motor 4 Conclusion References Ontology-Based Methodology for Knowledge Maps Design 1 Introduction 2 Knowledge Maps: Short Overview 3 Ontologies as a Conceptual Skeleton of a Knowledge Domain 4 Four Meta-steps to Create a Knowledge Map 5 Conclusion References Operating with Fuzzy Cases in Distributed Intelligent Systems 1 Introduction 2 Searching for a Decision Based on Cases 2.1 Ways of Representing Fuzzy Case 2.2 Extraction of Fuzzy Cases 3 Distributed Intelligent Systems 3.1 Distributed System Architecture 4 Conclusion References Methodology for Development Based on Ontological Models Intelligent Services with Explanation Generation 1 Introduction 2 Ontological Approach to Knowledge Representation 3 Intelligent Service Architecture 4 Service Ontological Components Formation 4.1 Ontological Knowledge Base Formation 4.2 Building an Ontological Solver 4.3 Intelligent Service User Interface 5 Discussion 6 Conclusion References Characterization of the Person’s leading interests in terms of RIASEC Scores 1 Introduction 2 Background Review 3 Problem Statement 4 Research Methodology 4.1 Description of the Data Set Under Study 4.2 Allocation of the Holland Scores to Leading Group Thematic 5 Discussion 6 Conclusion References Quantum Game Theory on Entangled Players 1 Introduction 1.1 Overview of Quantum Game Theory 1.2 Importance of Studying Entangled Players in Quantum Game Theory 1.3 Research Questions and Objectives 2 Background 3 Entangled Players in Quantum Game Theory 3.1 Types of Entangled States in Quantum Game Theory 3.2 Advantages of Entangled Players in Quantum Games 3.3 Disadvantages of Entangled Players in Quantum Games 3.4 Experimental Implementation of Entangled Players in Quantum Games 4 Analysis of Selected Quantum Games with Entangled Players 4.1 Quantum Prisoner's Dilemma Game with Entangled Players 4.2 Quantum Battle of the Sexes Game with Entangled Players 5 Experimental Implementation of Quantum Games with Entangled Players 5.1 Entanglement Swapping Game 6 Applications of Quantum Game Theory with Entangled Players 7 Conclusion References Cognitive Architecture of a System to Replicate Human Strategic Decision-Making 1 Introduction 2 Related Work 3 Cognitive Architecture of Human Strategic Decision-Making 4 System Description 4.1 Deployment Diagram 4.2 Decision-Making Modules 4.3 Support Modules 4.4 Mathematical Formulation of the Control Problem for the AEC 5 Conclusions References A Method for Modeling the Control Impact Strategy Based on the Mental Frame of References of the Decision-Maker 1 Introduction 2 Related Work 3 Methods 3.1 Predictive Institutional Model 3.2 Problem Statement 3.3 Modified AEC (MAEC) 4 Experiments 5 Discussion and Conclusions References Using an Intelligent Assistant for Aircraft Diagnostics and Maintenance 1 Introduction 2 Background 3 The Intelligent Assistant for Aircraft Diagnostics and Maintenance 4 Discussion 5 Conclusion References Approach to the Selection of Significant Parameters for Diagnostics of Complex Technical Systems Using Soft Computing 1 Introduction 2 Problem Statement and Algorithm for Solving the Problem of Significant Parameters Selection 3 Example of Selecting Significant Parameters 4 Conclusion References The Use of Machine Learning for Digital Shadowing in Thermal Spray Coating 1 Introduction 2 Literature Review 2.1 Data Collection 2.2 Data Ingestion 2.3 Data Contextualization - Digital Surface Manufacturing 3 Machine Learning Models 3.1 Linear Regression (LR) 3.2 Support Vector Regressor (SVR) 3.3 K Nearest Neighbor Regressor (KNNR) 3.4 Decision Tree Regression (DTR) 3.5 Case Based Reasoning (CBR) 4 Experiments and Results 5 Conclusion References Ontology-Based Explanations of Neural Networks for Collaborative Human-AI Decision Support Systems 1 Introduction 2 Ontology-Based Explanation Techniques for Neural Networks 2.1 Existing Approaches to Ontology-Based Explanations of Neural Networks 2.2 Using Ontologies for Building Explanations 3 Conceptual Interface and Interaction Mechanisms 4 Conclusion References Multiagent Systems Features of the Use of Multiagent Technology in the Management of Urban Parking Space 1 Introduction 2 Analysis of the Main Approaches to the Development of MAS Architecture in the Context of Implementing the Basic Requirements of the Industry 4.0 Concept 3 Concept-Functional Features of Intelligent Agents for Managing Urban Parking Space 3.1 Behavioral Model of DA-Type Agent (Driver's Tasks in On-Street or Off-Street Parking and Functions of DA Agent Acting in the Driver's Interests) 3.2 Behavioral Model of EA-Type Agent (Enforcement Tasks and Functions of an EA-Type Agent Acting in the Interests of the Enforcement Officer) 3.3 Behavioral Model of a PSDA Agent (Tasks of a Parking Space Dispatcher and Functions of a PSDA-Type Agent) 4 Features of Modeling the Interaction of Intelligent Agents Using the IMVIA System 5 Conclusion References Emergent intelligence of Digital Twins: From Concept to Applications 1 Introduction 2 Emergent Intelligence: The Concept 3 Methodology of EI Collective Problem Solving 4 Applications of EI Systems 5 Discussion of Results 6 Conclusion References Explainability in Artificial Intelligence Explanatory Capabilities of an Ontology-Based Solver with Linguistic and Visual Support 1 Introduction 2 Methodology 2.1 General 2.2 Ontology 3 Experiments 3.1 Problem 3.2 Ontology and Interactive Imaging 4 Discussion 5 Conclusion References Research on Neural Network Defense Problem Based on Random Noise Injection 1 Introduce 2 Related Works 3 Neural Network Defense Problem Based on Random Noise Injection 3.1 The Robustness of the Probability Mapping 3.2 Neural Network Defense Problem Based on Random Noise Injection 4 Experiments 4.1 Method Introduction 4.2 The Influence of Noise Injection on the Accuracy of Neural Networks 4.3 Effect of Noise Injection on Accuracy of Underfitted Neural Networks 4.4 Effect of Noise Injection on Accuracy of Convolutional Neural Networks 4.5 Heteroscedasticity Gaussian Noise Injection Defense Based on Layer-Wise Relevance Propagation on Underfitted Network 4.6 Heteroscedasticity Gaussian Noise Injection Defense Based on Layer-Wise Relevance Propagation on Neural Network 5 Conclusion References Explainable AI: Graph Based Sampling Approach for High Dimensional AI System 1 Introduction 2 Background and Related Work 2.1 Isolation Forest and Cancer Detection 2.2 Shapley Value 2.3 Shapley Sampling Approach 3 Relationship Map and Biased-Graph Based Sampling 3.1 Overview of Algorithm 3.2 Pearson Product-Moment Correlation Coefficient 3.3 Biased Random Path Searching Method 3.4 Measurements to Convergence 4 Simulation Result 4.1 Dataset Description 4.2 Relationship Map Generation and Configuration 4.3 Analysis of Results 5 Conclusion References Explainable Document Classification via Pattern Structures 1 Introduction 2 Related Work 3 Basic Definitions 4 Methodology 4.1 Document to Graph Conversion 4.2 Classification with an Aggregate Rule and Intermediate Explanation Generation 4.3 Final Explanation Generation 4.4 Complexity Analysis 5 Experiments and Results 6 Conclusion and Future Work References Author Index