Applications of Artificial Intelligence in Quality Assurance and Assurance of Productivity

Abstract

Probabilistic intelligence is vital in current management and technology. It is simpler to persuade readers when a management or engineer reports connected difficulties with objective statistical data. Statistical data support the evaluation of the true status, and cause and effect can be induced. The rationale is proven using deductive logic and statistical data verification and induction. Quality practitioners should develop statistical thinking skills and fully grasp the three quality principles: “essence of substance,” “process of business,” and “psychology.” Traditional quality data include variables, attributes, faults, internal and external failure costs, etc., obtained by data collection, data processing, statistical analysis, root cause analysis, etc. Quality practitioners used to rely on these so-called professional qualities to get a job. If quality practitioners do not keep up with the steps of times, quality data collection, organization, analysis, and monitoring will be confusing or challenging. Increasingly, precision tool machines are embedded in various IoTs, gathering machine operation data, component diagnostic and life estimation, consumables monitoring and utilization monitoring, and various data analyses. Data mining and forecasting have steadily been combined into Data Science, which is the future of quality field worth worrying about.

Modeling nonlinear dynamic objects using pre-trained time delay neural networks

“…To solve the contradiction that hinders the development of neural networks in the direction of identifying nonlinear dynamics, several approaches have been proposed depending on the properties of the object, the amount of a priori data, and the conditions of application [6]. Analysis and systematization of these approaches to the development of methods for accelerating the training of neural networks and rapid adaptation to new tasks on the basis of limited data allows us to identify several areas [1,7].…”

“…Analysis and systematization of these approaches to the development of methods for accelerating the training of neural networks and rapid adaptation to new tasks on the basis of limited data allows us to identify several areas [1,7]. Among them: improvement of activation functions [5,8], teaching methods [6,9] and the structure of neural networks [7].…”

“…This model consists of a set of basic pretrained neural networks fθSj that correspond to the existing typical characteristics of the object. Given the limitations (6) in the work, the parameters of the rough model θSi are determined from the composition of the corresponding parameters of the basic pretrained neural networks:…”

Interdisciplinary Nature of Computational Science Cases in Business Studies

“…Quantifying and managing this risk is paramount. Computational methods provide tools to simulate various market conditions and stress scenarios, helping financial institutions understand potential vulnerabilities in their positions (Chisty & Adusumalli, 2022). These simulations can test how portfolios respond to adverse events, allowing for preemptive strategies to mitigate potential losses (Rodriguez & Lewis, 2015).…”

Assessment of the quality of neural network models based on a multifactorial information criterion

“…The success of intelligent systems is determined by the simultaneous superposition of © Fomin O., Krykun V., 2024 many factors, including the following: a significant increase in computing resources, which is ensured by the increase in hardware performance and the development of cloud technologies; development of machine learning algorithms, primarily neural network (NN) architectures [3,4]. However, the quality of an intelligent system depends on the balanced development of all its components, including mathematical support [5,6]. Machine learning models, being the mathematical basis, play a crucial role in the functioning of intelligent systems.…”