Accurate Cost Estimation of Memory Systems Utilizing Machine Learning and Solutions from Computer Vision for Design Automation

Servadei, Lorenzo; Mosca, E.; Zennaro, Elena; Devarajegowda, Keerthikumara; Werner, Michael; Ecker, Wolfgang; Wille, Robert

doi:10.1109/tc.2020.2968888

Cited by 15 publications

(3 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this context, the application of artificial intelligence technology to mechanical design, manufacturing, and automation has a positive role in promoting the development of the entire industry [12]. Servadei, L. et al used machine learning principles to design a cost estimation strategy suitable for the field of computer vision, which has better accuracy than traditional methods and greatly reduces labor costs [13]. Chakraborty, S. K. et al conceived a fruit grading machine with a visual sorting function on the basis of modern intelligent sensing equipment, and the newly designed fruit grading machine showed good fruit sorting performance in practical feedback [14].…”

Section: Introductionmentioning

confidence: 99%

Research on the application of machine vision in the design of machine automation

Zhang,

2024

Applied Mathematics and Nonlinear Sciences

View full text Add to dashboard Cite

The application of machine vision technology to mechanical automation design is an inevitable measure to improve the comprehensive productivity of mechanical production manufacturing. In this paper, a mechanical automation system based on SOA architecture is designed to cover the image acquisition, processing, and recognition workflow of an image processing module based on machine vision. The binocular camera calibration is completed by the Zhang Zhengyou calibration method. The image is preprocessed by using algorithms such as the mean value method, segmented linear transformation and median filtering, and an algorithm is proposed to improve the ORB feature point extraction, which can extract and match the feature points quickly and efficiently. Meanwhile, a Harris corner point detection algorithm is proposed to improve the SIFT algorithm to enhance the accuracy of target recognition and localization. The designed mechanical automation system is applied to the lychee picking robot, for example, analysis, which shows that the overall recognition P-R value of the system reaches 0.953, and the classification accuracy is above 0.917. In the mechanical automation route localization, the lateral deviation is lower than 4.80cm. The maximum time for image processing and parameter transfer is 117.966ms, which indicates that the system involved in this paper is relatively stable in operation and has a better effect on the localization of the navigation line, which is of certain application value in the field of mechanical automation design.

show abstract

Section: Introductionmentioning

confidence: 99%

Research on the application of machine vision in the design of machine automation

Zhang,

2024

Applied Mathematics and Nonlinear Sciences

View full text Add to dashboard Cite

show abstract

“…In recent years, with the continuous development of society and the continuous progress of science and technology, machine learning [1][2] , artificial intelligence [3][4] and other technologies have gradually been applied in people's life, work, learning and other aspects, such as computer vision [5][6] , natural language processing [7][8] , recommendation technology [9][10] and other fields. The industry and application innovation related to artificial intelligence technology are constantly developing all over the world.…”

Section: Introductionmentioning

confidence: 99%

Personalized Federated Learning with Attention Mechanisms

2023

AJCIS

View full text Add to dashboard Cite

With the rapid development of science and technology, machine learning and deep learning technology have more and more applications in our daily life. At the same time, people pay more and more attention to the protection of their own privacy. And in recent years, various countries have introduced a series of laws and regulations to protect people's data privacy. In this case, the traditional method of pooling the data of each user for model training is no longer applicable, and the data of each user can only be saved in the hands of the user himself, which has a profound impact on the development of artificial intelligence technology. To solve this dilemma, the industry has proposed the concept of federated learning. Due to the data imbalance problem between each user, the prediction ability of the global aggregation model obtained by federated learning on the local specific data of each user needs to be improved. Therefore, personalization is an issue that federated learning needs to pay attention to.In order to improve the personalization ability, this paper proposes a federated transfer learning algorithm with attention mechanism. After each user obtains the global aggregation model of federated learning, the attention module is added to the local model, and then the parameters of the low-level neurons are frozen during training, and only the parameters of the high-level neurons and the attention module are updated. Finally, each user obtains a unique model that is more suitable for local data. In this paper, we conduct experiments to analyze the performance of this algorithm and the federated transfer algorithm and federated learning algorithm in terms of accuracy. The experimental results show that on the convolutional neural network model, the federated transfer learning algorithm applying the attention mechanism has improved the accuracy compared with the federated transfer algorithm and the federated learning algorithm.

show abstract

“…and computer vision analysis [11], Deep-Learning-based AI techniques are beginning to position themselves as the technological solution to the aforementioned problems. Deep learning is a specialized form of machine learning, characterized by the use of artificial neural network architectures composed of multiple processing stages.…”

Section: Additionallymentioning

confidence: 99%

Deep learning based methodologies applied to industrial electromechanical systems monitoring

Arellano Espitia

View full text Add to dashboard Cite

(English) In recent years, there has been significant attention from both the academic and industry sectors towards condition-based maintenance of rotating systems. This attention stems from the high relevance of these systems in the industrial sector, particularly in production plants and manufacturing processes. To ensure their high efficiency, safety, and proper performance, it is crucial to have monitoring systems, schemes, and methodologies capable of diagnosing and detecting malfunctions in the various components that comprise the electromechanical rotating system, including electrical motors, bearings, gearboxes, and shafts among others. However, monitoring schemes face considerable challenges in dealing with the complexity inherent in these systems. This complexity arises from the diverse operational conditions, configurations with multiple components, and the presence of faults of different natures (i.e. mechanical, electrical and/or electromagnetic), occurring in isolated or combined scenarios. In recent times, techniques based on artificial intelligence, particularly artificial neural networks, have gained significant popularity in monitoring schemes. These techniques offer great potential in terms of system characterization and feature extraction, leading to the exploration of methodologies based on them. Furthermore, the advent of deep neural networks, which are multi-layered architectures, represents a significant advancement in the field of pattern recognition. Nevertheless, the implementation of these schemes in industrial environments continues to present ongoing challenges due to the demanding requirements, such as high reliability, robustness, and scalability, particularly within the context of the new Industry 4.0 paradigm. In this context, this thesis focuses on the investigation and proposal of a series of complementary methodologies based on deep learning, aiming to enable the implementation of predictive maintenance schemes for the industrial sector. Initially, a data-driven monitoring methodology based on deep learning is proposed for fault diagnosis in complex electromechanical system environments. Subsequently, a methodological process for anomaly identification, comprising deep learning strategies and machine learning tools, is presented. Finally, an intelligent diagnostic model utilizing domain adaptation techniques is proposed, resulting in the creation of a scalable model with the ability to generalize to fault scenarios under different operating conditions and address the issue of performance degradation. The proposed methodologies are studied and validated using experimental databases of electromechanical systems laboratory based. Quantifiable performance metrics and qualitative validation procedures are employed for rigorous evaluation and validation purposes. (Español) En los últimos años, el campo de mantenimiento basado en la condición de los sistemas rotativos ha recibido atención tanto del sector académico como de la industria. Esto se debe a la gran relevancia de estos sistemas en el sector industrial, especialmente para los procesos de producción y manufactura. Para garantizar la alta eficiencia, seguridad y correcto desempeño de los diferentes componentes que pueden formar un sistema rotativo, tales como motores, rodamientos, cajas de engranes y ejes, es fundamental contar con sistemas de monitoreo, esquemas y metodologías con la capacidad de diagnosticar y detectar mal funcionamientos de las mismas. Sin embargo, los esquemas de monitoreo presentan grandes desafíos al tratar con la complejidad de estos sistemas, que implica un ambiente de trabajo sujeto a diferentes condiciones de operación, configuraciones con múltiples componentes y la presencia de fallas de diferente naturaleza, es decir, mecánicas, eléctricas, electromagnéticas, bajo condiciones aisladas o escenarios combinados. Recientemente, las técnicas basadas en inteligencia artificial se han utilizado ampliamente en esquemas de monitoreo, especialmente las redes neuronales artificiales. Debido a su gran capacidad de caracterización y de extracción de características, se ha hecho un esfuerzo por estudiar metodologías basadas en estas técnicas. Además, las arquitecturas multicapa, conocidas como redes neuronales profundas, representan un avance considerable en el campo del reconocimiento de patrones. Sin embargo, la implantación de estos esquemas en entornos industriales aún presenta un desafío continúo debido a los requerimientos, como la alta confiabilidad, robustez y escalabilidad que deben cumplirse, especialmente en el nuevo paradigma de Industria 4.0. En este sentido, esta tesis consiste en la investigación y propuesta de una serie de metodologías complementarias basadas en aprendizaje profundo que conducen a la implementación de esquemas de mantenimiento predictivo para el sector industrial. Inicialmente, se propone una metodología de monitoreo de datos basada en aprendizaje profundo para el diagnóstico de fallas para entornos de sistemas electromecánicos complejos. A continuación, se propone un proceso metodológico para la identificación de anomalías compuesto por estrategias de aprendizaje profundo y herramientas de aprendizaje automático. Finalmente, se propone un modelo de diagnóstico inteligente utilizando herramientas basadas en la adaptación de dominio contradictorio. Esto conduce a la creación de un modelo escalable con la capacidad de diagnosticar de manera general, escenarios de fallas bajo diferentes condiciones de operación y resolver el problema de degradación de rendimiento. El estudio y la validación de las metodologías propuestas se han llevado a cabo mediante bases de datos experimentales de sistemas electromecánicos producidos en laboratorio. Se implementaron métricas de rendimiento cuantificables y también procedimientos de validación cualitativa.

show abstract

Accurate Cost Estimation of Memory Systems Utilizing Machine Learning and Solutions from Computer Vision for Design Automation

Cited by 15 publications

References 18 publications

Research on the application of machine vision in the design of machine automation

Research on the application of machine vision in the design of machine automation

Personalized Federated Learning with Attention Mechanisms

Deep learning based methodologies applied to industrial electromechanical systems monitoring

Contact Info

Product

Resources

About