相关参考文献
注意：仅列出部分参考文献，下载原文获取全部文献信息。
Article
Computer Science, Artificial Intelligence
Lingfei Wu et al.
Summary: Deep learning is widely used in Natural Language Processing (NLP) and there has been a growing interest in applying graph neural networks (GNNs) to NLP tasks. This survey provides a comprehensive overview of GNNs for NLP, including a new taxonomy and organization of existing research in this field. The survey also presents various NLP applications that utilize GNNs, along with benchmark datasets, evaluation metrics, and open-source codes. Additionally, it discusses challenges and future research directions for maximizing the potential of GNNs in NLP.
FOUNDATIONS AND TRENDS IN MACHINE LEARNING
(2023)
Article
Computer Science, Artificial Intelligence
Jianfeng Deng et al.
Summary: Knowledge graph technology plays a crucial role in efficient supply chain management in manufacturing enterprises. A top-down construction method is proposed to address problems in coarse concept granularity, entity recognition, and lack of annotated training samples. Experimental results demonstrate that the proposed method improves entity recognition accuracy and achieves high accuracy even with limited manual annotation data.
ADVANCED ENGINEERING INFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Bogyeong Lee et al.
Summary: Despite automation reducing the number of workers in construction, worker safety remains a crucial issue. Efforts have been made to monitor safety behaviors with additional personnel, but existing methods struggle to capture workers' compliance. This study proposes an approach based on OpenPose and a spatio-temporal graph convolutional network to evaluate workers' compliance with safety regulations and provide behavior-based feedback.
ADVANCED ENGINEERING INFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Xin Huang et al.
Summary: In this paper, a novel intelligent diagnosis framework is proposed for accurately identifying the crack severity of turbine blades. The framework combines multiscale sparse filtering (MSF)-based unsupervised sparse feature learning and multi-kernel support vector machine for information fusion (MKSVMIF). Signal processing methods, including the enhanced EEMD-based multiwavelet packet energy entropy (EEMD-WPEE), are used to eliminate interference and retain fault-related characteristics. Extensive experiments on a blade-rotor simulation rig validate the effectiveness of the proposed framework in quantitatively detecting different blade crack severities.
ADVANCED ENGINEERING INFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Junhao Hou et al.
Summary: Performing classification, retrieval, and reuse of 3D CAD models is crucial in industrial manufacturing for reducing development costs and shortening the engineering development cycle. Existing methods for mesh or point cloud data cannot be directly applied to 3D CAD models represented by B-rep, so a graph structure descriptor called B-rep graph is proposed to preprocess the data. A novel neural network called FuS-GCN, utilizing graph convolutional networks, is designed to handle the B-rep graph data. Experimental results show that FuS-GCN outperforms alternative 3D shape descriptors and effectively extracts features from 3D CAD models.
ADVANCED ENGINEERING INFORMATICS
(2023)
Article
Engineering, Industrial
Yu He et al.
Summary: Additive-subtractive hybrid manufacturing (ASHM) combines the advantages of both additive manufacturing (AM) and subtractive manufacturing (SM) to achieve high efficiency, quality, and precision in the production of complex structural parts. A novel process planning approach for ASHM is proposed, which decomposes the geometrical structure of complex parts, makes decisions on AM, SM, and their alternation based on multi-dimensional manufacturability evaluation, and optimizes manufacturing resources, time, and cost. Manufacturing constraint rules and manufacturability indexes are established for process selection, and a Comprehensive Hybrid Manufacturing Complexity Index (CHCI) is developed for decision support. Experimental case studies demonstrate the efficacy of the proposed method.
JOURNAL OF MANUFACTURING SYSTEMS
(2023)
Article
Engineering, Industrial
Weiye Li et al.
Summary: This paper proposes a multi-agent evolutionary reinforcement learning method (MAERL) to optimize the machining parameters for high-quality and high-efficiency machining. It combines the graph neural network and reinforcement learning. Experimental results on the commutator production line show that the proposed method improves the prediction effect by about 25% in the case of small samples, and the MAERL-based optimization method can better deal with the coupling problem of the reward function in the optimization process. Compared with the classical optimization method, the optimization effect is improved by 13% and a lot of optimization time is saved.
JOURNAL OF MANUFACTURING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Asmaa Rassil et al.
Summary: Deep generative models have made significant advances in medicine and drug research by efficiently predicting new molecular drug candidates with valid structures. However, sampling candidates from the chemical space using only one reinforcement learning agent can be challenging. To overcome this, we propose the Deep Fusion Q-Network (DFQN), a multi-agent RL-based DNN that allows better exploration of the environment by exploiting multiple RL agents to generate candidate graph structures. By using a heterogeneous graph representation and attention-based fusion network, DFQN promotes coordination and communication between agents to generate graphs with the required properties. We also incorporate adversarial training and a new approach to check the chemical validity of the designed molecules, achieving remarkable results compared to state-of-the-art models.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Computer Science, Interdisciplinary Applications
Fengyi Lu et al.
Summary: This paper proposes a novel multi-pass parametric optimization method based on deep reinforcement learning (DRL) to improve energy efficiency. By allowing parameters to vary, and transforming the model into a Markov Decision Process, the proposed method significantly improves material removal rate and specific cutting energy while meeting deformation tolerance, which substantiates the benefits of the energy-efficient parametric optimization.
ROBOTICS AND COMPUTER-INTEGRATED MANUFACTURING
(2023)
Article
Computer Science, Artificial Intelligence
Caihua Hao et al.
Summary: With the increasing demand for high precision machining of aluminum alloy parts in the field of intelligent electric vehicles and 3C, intelligent high-accuracy wear prediction of aluminum alloy high precision machining tools has significant industrial application value. This paper presents a novel TCM model (Conv-PhyFormer) that captures short-term and long-term dependencies from nonlinear cutting time series data with few training samples. The embedded soft and hard physical constraints in the model contribute to its superior prediction accuracy and robustness compared to current deep learning models for TCM.
ADVANCED ENGINEERING INFORMATICS
(2023)
Article
Engineering, Industrial
Xu Liu et al.
Summary: This research presents a new data-driven method for constructing machining regions by learning from historical data, eliminating the need for domain-specific expert-defined rules or strategies. Experimental results confirm the practicality of the proposed method in automatic machining region construction and suggest its potential for extension to parts in other domains.
JOURNAL OF MANUFACTURING SYSTEMS
(2022)
Article
Computer Science, Interdisciplinary Applications
Peizhi Shi et al.
Summary: In the field of intelligent manufacturing, recognizing interacting features on a CAD model is a critical yet challenging task. Some learning methods struggle with highly interacting features and require a large number of 3D models for training. The proposed method RDetNet can effectively recognize highly interacting features with small training samples.
ROBOTICS AND COMPUTER-INTEGRATED MANUFACTURING
(2022)
Article
Computer Science, Artificial Intelligence
Pin Lyu et al.
Summary: This paper proposes a novel deep learning-based method for bearing fault diagnosis, called RSG, which achieves effective noise reduction and feature extraction by integrating the working mechanisms of soft threshold and global context. Comparative analysis demonstrates the advantages of the proposed method, and experimental results show significant fault diagnosis accuracy in a real-world industrial environment.
ADVANCED ENGINEERING INFORMATICS
(2022)
Article
Engineering, Industrial
Jingjing Li et al.
Summary: The paper proposes a general framework for twin data and knowledge-driven intelligent process planning of aviation parts, supported by four standard procedures to optimize process plans and continuously improve the quality of process planning.
INTERNATIONAL JOURNAL OF PRODUCTION RESEARCH
(2022)
Review
Automation & Control Systems
Xiaojian Wen et al.
Summary: This article discusses the importance of 3D machining technology in modern manufacturing and highlights the key role of process design based on 3D models in ensuring machining quality. By conducting a statistical analysis of relevant literature, it summarizes the key technologies and development trends in 3D process design.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2022)
Article
Automation & Control Systems
Chunhua Feng et al.
Summary: This paper focuses on the strategy of optimizing energy consumption in the processing of parts with multiple features in CNC machines. The energy consumption of the cutting process is established using unit volume cutting energy, while the impact of feature sequencing on non-cutting energy consumption is analyzed. An energy model is established considering different moving axes and feed paths. Multi-objective optimization is carried out to minimize cutting energy consumption, improve machining quality, and reduce machining time. Experimental data is used to establish specific energy consumption models for each processing stage, and a case study demonstrates the effectiveness of the proposed method.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2022)
Article
Computer Science, Artificial Intelligence
Tong Zhao et al.
Summary: The study proposes a novel synergistic approach that combines pattern mining and feature learning for graph anomaly detection, showing significantly better performance than existing methods.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Chemistry, Physical
Magdalena Zawada-Michalowska et al.
Summary: This paper aims to analyze the effect of selected geometric properties on post-machining deformations of thin-walled structures. Through experiments on specially designed thin-walled sample elements, it was found that walls arranged in a semi-open structure had higher deformation values, and the use of high-speed cutting helped minimize the deformations of thin-walled elements. Moreover, the EN AW-7075 T651 alloy exhibited greater deformations.
Article
Computer Science, Interdisciplinary Applications
Yajun Zhang et al.
Summary: This paper proposes a macro process decision-making approach combining knowledge graph and deep learning technology, which can better plan the machining process, improve machining quality, and reduce costs.
COMPUTERS IN INDUSTRY
(2022)
Article
Automation & Control Systems
Xingyu Fu et al.
Summary: This article presents a novel 3-D descriptor called improved dexel representation (IDR), which enables the input of holistic information from an engineering CAD model to CNN-based manufacturing applications. The IDR carries position, size, and surface information of the model, providing high resolution for local features and potential for model reconstruction. The use of IDR as input for CNNs significantly improves prediction accuracy.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Engineering, Industrial
Hang Zhang et al.
Summary: This article presents a deep learning-based approach for estimating manufacturing costs, with a focus on the precision information of parts. The approach defines an attribute graph to represent the CAD model of a part and constructs a ConvGNN framework called Cost Estimation Network (CEN) that combines spectral-based and spatial-based convolutional layers. The trained CEN can accurately estimate manufacturing costs, and a modified Grad-CAM process is developed to explain the rationale behind cost decisions. Experimental studies using CNC machined rotary parts validate the feasibility and effectiveness of the approach.
JOURNAL OF MANUFACTURING SYSTEMS
(2022)
Article
Computer Science, Interdisciplinary Applications
Hang Zhang et al.
Summary: This paper proposes a novel multi-task network named ASIN based on point cloud data for machining feature recognition. By combining the tasks of segmentation, identification, and bottom face identification, ASIN can simultaneously achieve feature segmentation, identification, and bottom face identification. Experimental results demonstrate that the proposed method effectively segments machining features and performs well in recognizing intersecting machining features.
ROBOTICS AND COMPUTER-INTEGRATED MANUFACTURING
(2022)
Article
Chemistry, Analytical
Tae-Won Jung et al.
Summary: The paper proposes a one-stage GCN approach for 3D object detection and poses estimation by structuring non-linearly distributed points of a graph. Experimental results show that the method improves memory usage and efficiency, and achieves comparable performance against state-of-the-art systems.
Article
Engineering, Industrial
Wenbo Wu et al.
Summary: This paper addresses the uncertainty brought by mass customized production to CAPP using deep reinforcement learning, simplifying the decision procedure by integrating environmental states and agent behaviors, and efficiently solving planning problems by screening out inexecutable operations with a masking algorithm.
JOURNAL OF MANUFACTURING SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Soyoung Yoo et al.
Summary: Studies have recently focused on deep learning for manufacturing cost prediction, but the lack of explanation due to using models as black boxes remains a challenge. This study proposes a process using explainable artificial intelligence to predict manufacturing costs for 3D CAD models, allowing for visualization of machining features that impact cost. The proposed approach demonstrates high predictability for CNC machined parts and can offer guidance to engineering designers and real-time quotations to online manufacturing platform customers.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Interdisciplinary Applications
Tianchi Deng et al.
Summary: This study proposes a data-driven method for machining parameter planning by learning from high-quality historical processing files. By constructing attribute graphs and utilizing graph neural networks, human interactions can be greatly reduced and model performance can be improved.
ROBOTICS AND COMPUTER-INTEGRATED MANUFACTURING
(2021)
Article
Automation & Control Systems
Yajun Zhang et al.
Summary: This paper discusses a novel approach for process route generation based on deep learning. By using a fourth-order tensor model and encoder-decoder neural architecture, the automatic generation of machining process route for parts is achieved. The feasibility and effectiveness of the proposed method are demonstrated through experiments on slot cavity parts.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2021)
Article
Computer Science, Interdisciplinary Applications
Liang Guo et al.
Summary: The study introduces Knowledge Graph (KG) to address issues in the automatic machining process decision-making system. By establishing an information model and reasoning framework, as well as applying a hybrid reasoning algorithm, the problem of heterogeneity among process knowledge is successfully solved, and a prototype system is developed for verification.
INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING
(2021)
Review
Engineering, Industrial
S. P. Leo Kumar
INTERNATIONAL JOURNAL OF PRODUCTION RESEARCH
(2019)
Article
Computer Science, Interdisciplinary Applications
Guanghui Zhou et al.
ADVANCES IN ENGINEERING SOFTWARE
(2019)
Article
Automation & Control Systems
Yingxin Ye et al.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2018)
Article
Computer Science, Interdisciplinary Applications
Wei Ji et al.
INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING
(2018)
Article
Computer Science, Artificial Intelligence
Jiewu Leng et al.
KNOWLEDGE-BASED SYSTEMS
(2018)
Article
Automation & Control Systems
Chunlei Li et al.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2016)
Article
Automation & Control Systems
Danchen Zhou et al.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2015)
Article
Automation & Control Systems
Rui Huang et al.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2014)
Article
Computer Science, Artificial Intelligence
Xiao-jun Liu et al.
JOURNAL OF INTELLIGENT MANUFACTURING
(2013)
Article
Computer Science, Artificial Intelligence
Mariusz Deja et al.
JOURNAL OF INTELLIGENT MANUFACTURING
(2013)
Article
Engineering, Industrial
Cheol-Soo Lee et al.
JOURNAL OF MANUFACTURING SYSTEMS
(2013)
Article
Computer Science, Interdisciplinary Applications
Kriangkrai Waiyagan et al.
COMPUTERS IN INDUSTRY
(2009)
Article
Automation & Control Systems
Guang-ru Hua et al.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2007)
Article
Computer Science, Artificial Intelligence
Sankha Deb et al.
JOURNAL OF INTELLIGENT MANUFACTURING
(2006)
