Offline policy evaluation,evaluating and selecting complex policies for decision-making by only using offline datasets is important in reinforcement learning.At present,the model-based offline policy evaluation(MBOPE)...Offline policy evaluation,evaluating and selecting complex policies for decision-making by only using offline datasets is important in reinforcement learning.At present,the model-based offline policy evaluation(MBOPE)is widely welcomed because of its easy to implement and good performance.MBOPE directly approximates the unknown value of a given policy using the Monte Carlo method given the estimated transition and reward functions of the environment.Usually,multiple models are trained,and then one of them is selected to be used.However,a challenge remains in selecting an appropriate model from those trained for further use.The authors first analyse the upper bound of the difference between the approximated value and the unknown true value.Theoretical results show that this difference is related to the trajectories generated by the given policy on the learnt model and the prediction error of the transition and reward functions at these generated data points.Based on the theoretical results,a new criterion is proposed to tell which trained model is better suited for evaluating the given policy.At last,the effectiveness of the proposed criterion is demonstrated on both benchmark and synthetic offline datasets.展开更多
To address the challenges of low accuracy and insufficient real-time performance in dynamic object detection for UAV surveillance,this paper proposes a novel tracking framework that integrates a lightweight improved Y...To address the challenges of low accuracy and insufficient real-time performance in dynamic object detection for UAV surveillance,this paper proposes a novel tracking framework that integrates a lightweight improved YOLOv5s model with adaptive motion compensation.A UAV-view dynamic feature enhancement strategy is innovatively introduced,and a lightweight detection network combining attention mechanisms and multi-scale fusion is constructed.The robustness of tracking under motion blur scenarios is also optimized.Experimental results demonstrate that the proposed method achieves a mAP@0.5 of 68.2%on the VisDrone dataset and reaches an inference speed of 32 FPS on the NVIDIA Jetson TX2 platform.This significantly improves the balance between accuracy and efficiency in complex scenes,offering reliable technical support for real-time applications such as emergency response.展开更多
With the large-scale promotion of distributed photovoltaics,new challenges have emerged in the photovoltaic consumptionwithin distribution networks.Traditional photovoltaic consumption schemes have primarily focused o...With the large-scale promotion of distributed photovoltaics,new challenges have emerged in the photovoltaic consumptionwithin distribution networks.Traditional photovoltaic consumption schemes have primarily focused on static analysis.However,as the scale of photovoltaic power generation devices grows and the methods of integration diversify,a single consumption scheme is no longer sufficient to meet the actual needs of current distribution networks.Therefore,this paper proposes an optimal evaluation method for photovoltaic consumption schemes based on BASS model predictions of installed capacity,aiming to provide an effective tool for generating and evaluating photovoltaic consumption schemes in distribution networks.First,the BASS diffusion model,combined with existing photovoltaic capacity data and roof area information,is used to predict the trends in photovoltaic installed capacity for each substation area,providing a scientific basis for consumption evaluation.Secondly,an improved random scenario simulation method is proposed for assessing the photovoltaic consumption capacity in distribution networks.This method generates photovoltaic integration schemes based on the diffusion probabilities of different regions and evaluates the consumption capacity of each scheme.Finally,the Technique for Order Preference by Similarity to an Ideal Solution(TOPSIS)is used to comprehensively evaluate the generated schemes,ensuring that the selected scheme not only meets the consumption requirements but also offers high economic benefits and reliability.The effectiveness and feasibility of the proposedmethod are validated through simulations of the IEEE 33-node system,providing strong support for optimizing photovoltaic consumption schemes in distribution networks.展开更多
A large number of 3D models are created on computers and available for networks. Some content-based retrieval technologies are indispensable to find out particular data from such anonymous datasets. Though several sha...A large number of 3D models are created on computers and available for networks. Some content-based retrieval technologies are indispensable to find out particular data from such anonymous datasets. Though several shape retrieval technologies have been developed, little attention has been given to the points on human＇s sense and impression （as known as Kansei） in the conventional techniques, In this paper, the authors propose a novel method of shape retrieval based on shape impression of human＇s Kansei. The key to the method is the Gaussian curvature distribution from 3D models as features for shape retrieval. Then it classifies the 3D models by extracted feature and measures similarity among models in storage.展开更多
This paper proposes a method to recognize human-object interactions by modeling context between human actions and interacted objects.Human-object interaction recognition is a challenging task due to severe occlusion b...This paper proposes a method to recognize human-object interactions by modeling context between human actions and interacted objects.Human-object interaction recognition is a challenging task due to severe occlusion between human and objects during the interacting process.Since that human actions and interacted objects provide strong context information,i.e.some actions are usually related to some specific objects,the accuracy of recognition is significantly improved for both of them.Through the proposed method,both global and local temporal features from skeleton sequences are extracted to model human actions.In the meantime,kernel features are utilized to describe interacted objects.Finally,all possible solutions from actions and objects are optimized by modeling the context between them.The results of experiments demonstrate the effectiveness of our method.展开更多
The increasing global demand for sustainable agricultural practices and effective waste management has highlighted the potential of biochar as a multifaceted solution. This study evaluates the economic viability of su...The increasing global demand for sustainable agricultural practices and effective waste management has highlighted the potential of biochar as a multifaceted solution. This study evaluates the economic viability of sugarcane bagasse-based biochar in Brazil, focusing on its potential to enhance agricultural productivity and contribute to environmental sustainability. While existing literature predominantly explores the production, crop yield benefits, and carbon sequestration capabilities of biochar, there is a notable gap in comprehensive economic modeling and viability analysis for the region. This paper aims to fill this gap by employing a scenario-based economic modeling approach, incorporating relevant economic models. Findings include that biochar implementation can be economically viable for medium and large sugarcane farms (20,000 - 50,000 hectares) given the availability of funding, breaking even in about 7.5 years with an internal rate of return of 18% on average. For small farms, biochar can only be viable when applied biochar to the soil, which in all scenarios is found to be the more profitable practice by a large margin. Sensitivity analyses found that generally, biochar becomes economically feasible at biochar carbon credit prices above $120 USD/tCO2e, and at sugarcane bagasse availability percentages above 60%. While the economic models are well-grounded in existing literature, the production of biochar at the studied scales is not yet widespread, especially in Brazil and uncertainties can result. Reviewing the results, the land application scenario was found to be the most viable, and large farms saw the best results, highlighting the importance of scale in biochar operations. Small and medium farms with no land application were concluded to have no or questionable viability. Overall, sugarcane bagasse-based biochar can be economically viable, under the right circumstances, for agricultural and environmental advancement in Brazil.展开更多
Recently automotive nets are adopted to solve increasing problems in automotive electronic systems.Technologies of automotive local area network from CAN and LIN can solve the problems of the increasing of wire bunch ...Recently automotive nets are adopted to solve increasing problems in automotive electronic systems.Technologies of automotive local area network from CAN and LIN can solve the problems of the increasing of wire bunch weight and lack in module installation space.However,the multilayer automotive nets software becomes more and more complex,and the development expense is difficult to predict and to keep in check.In this paper,the modeling method of hierarchical automotive nets and the substitution operation based on object-oriented colored Petri net(OOCPN) are proposed.The OOCPN model which analyzes the software structure and validates the collision mechanism of CAN/LIN bus can speed the automobile system development.First,the subsystems are divided and modeled by object-oriented Petri net(OOPN).According to the sets of message sharing relations,the message ports among them are set and the communication gate transitions are defined.Second,the OOPN model is substituted step by step until the inner objects in the automotive body control modules(BCM) are indivisible and colored by colored Petri net(CPN).And the color subsets mark the node messages for the collision mechanism.Third,the OOCPN model of the automotive body CAN/LIN nets is assembled,which keeps the message sets and the system can be expanded.The proposed model is used to analyze features of information sharing among the objects,and it is also used to describe each subsystem real-time behavior of processing messages and implemental device controllers operating,and puts forward a reasonable software framework for the automotive body control subsystem.The research can help to design the communication model in the automotive body system effectively and provide a convenient and rapid way for developing the logical hierarchy software.展开更多
Modelica-based object-orient method is proved to be rapid, accurate and easy to modify, which is suitable for prototype modeling and simulation of rotor system, whose parameters need to be modified frequently. Classic...Modelica-based object-orient method is proved to be rapid, accurate and easy to modify, which is suitable for prototype modeling and simulation of rotor system, whose parameters need to be modified frequently. Classical non-object-orient method appears to be inefficient because the code is difficult to modify and reuse. An adequate library for object-orient modeling of rotor system with multi-faults is established, a comparison with non-object-orient method on Jeffcott rotor system and a case study on turbo expander with multi-faults are implemented. The relative tolerance between object-orient method and non-object-orient is less than 0.03%, which proves that these two methods are as accurate as each other. Object-orient modeling and simulation is implemented on turbo expander with crack, rub-impact, pedestal looseness and multi-faults simultaneously. It can be conclude from the case study that when acting on compress side of turbo expander separately, expand wheel is not influenced greatly by crack fault, the existence of rub-impact fault forces expand wheel into quasi-periodic motion and the orbit of expand wheel is deformed and enhanced almost 1.5 times due to pedestal looseness. When acting simultaneously, multi-faults cannot be totally decomposed but can be diagnosed from the feature of vibration. Object-orient method can enhance the efficiency of modeling and simulation of rotor system with multi-faults, which provides an efficient method on prototype modeling and simulation.展开更多
LIDAR point cloud-based 3D object detection aims to sense the surrounding environment by anchoring objects with the Bounding Box(BBox).However,under the three-dimensional space of autonomous driving scenes,the previou...LIDAR point cloud-based 3D object detection aims to sense the surrounding environment by anchoring objects with the Bounding Box(BBox).However,under the three-dimensional space of autonomous driving scenes,the previous object detection methods,due to the pre-processing of the original LIDAR point cloud into voxels or pillars,lose the coordinate information of the original point cloud,slow detection speed,and gain inaccurate bounding box positioning.To address the issues above,this study proposes a new two-stage network structure to extract point cloud features directly by PointNet++,which effectively preserves the original point cloud coordinate information.To improve the detection accuracy,a shell-based modeling method is proposed.It roughly determines which spherical shell the coordinates belong to.Then,the results are refined to ground truth,thereby narrowing the localization range and improving the detection accuracy.To improve the recall of 3D object detection with bounding boxes,this paper designs a self-attention module for 3D object detection with a skip connection structure.Some of these features are highlighted by weighting them on the feature dimensions.After training,it makes the feature weights that are favorable for object detection get larger.Thus,the extracted features are more adapted to the object detection task.Extensive comparison experiments and ablation experiments conducted on the KITTI dataset verify the effectiveness of our proposed method in improving recall and precision.展开更多
In the developing course of the information system, the modeling method of information system and model expressing problem are very important. This paper through discussing the facing object of Petri network, expatiat...In the developing course of the information system, the modeling method of information system and model expressing problem are very important. This paper through discussing the facing object of Petri network, expatiates the basic problem of how to set up the business procedure model on the basis of Petri network, and shows the expressing and storing methods of the model in computer way.展开更多
Use of features in order to achieve the integration of design and manufacture has been considered to be a key factor recent years. Features such as manufacturing properties form the workpiece. Features are structured ...Use of features in order to achieve the integration of design and manufacture has been considered to be a key factor recent years. Features such as manufacturing properties form the workpiece. Features are structured systematically through object oriented modeling. This article explains an object coding method developed for prismatic workpieces and the use of that method in process planning. Features have been determined and modeled as objects. Features have been coded according to their types and locations on the workpiece in this given method. Feature codings have been seen to be very advantageous in process planning.展开更多
Named Entity Recognition(NER)stands as a fundamental task within the field of biomedical text mining,aiming to extract specific types of entities such as genes,proteins,and diseases from complex biomedical texts and c...Named Entity Recognition(NER)stands as a fundamental task within the field of biomedical text mining,aiming to extract specific types of entities such as genes,proteins,and diseases from complex biomedical texts and categorize them into predefined entity types.This process can provide basic support for the automatic construction of knowledge bases.In contrast to general texts,biomedical texts frequently contain numerous nested entities and local dependencies among these entities,presenting significant challenges to prevailing NER models.To address these issues,we propose a novel Chinese nested biomedical NER model based on RoBERTa and Global Pointer(RoBGP).Our model initially utilizes the RoBERTa-wwm-ext-large pretrained language model to dynamically generate word-level initial vectors.It then incorporates a Bidirectional Long Short-Term Memory network for capturing bidirectional semantic information,effectively addressing the issue of long-distance dependencies.Furthermore,the Global Pointer model is employed to comprehensively recognize all nested entities in the text.We conduct extensive experiments on the Chinese medical dataset CMeEE and the results demonstrate the superior performance of RoBGP over several baseline models.This research confirms the effectiveness of RoBGP in Chinese biomedical NER,providing reliable technical support for biomedical information extraction and knowledge base construction.展开更多
Aiming at the problem of low surface defect detection accuracy of industrial products, an object detection method based on simplified spatial pyramid pooling fast(Sim SPPF) hybrid pooling improved you only look once v...Aiming at the problem of low surface defect detection accuracy of industrial products, an object detection method based on simplified spatial pyramid pooling fast(Sim SPPF) hybrid pooling improved you only look once version 5s(YOLOV5s) model is proposed. The algorithm introduces channel attention(CA) module, simplified SPPF feature vector pyramid and efficient intersection over union(EIOU) loss function. Feature vector pyramids fuse high-dimensional and low-dimensional features, which makes semantic information richer. The CA mechanism performs maximum pooling and average pooling operations on the feature map. Hybrid pooling comprehensively improves detection computing efficiency and accurate deployment ability. The results show that the improved YOLOV5s model is better than the original YOLOV5s model. The average test accuracy(mAP) can reach 91.8%, which can be increased by 17.4%, and the detection speed can reach 108 FPS, which can be increased by 18 FPS. The improved model is practicable, and the overall performance is better than other conventional models.展开更多
In the context of power generation companies, vast amounts of specialized data and expert knowledge have been accumulated. However, challenges such as data silos and fragmented knowledge hinder the effective utilizati...In the context of power generation companies, vast amounts of specialized data and expert knowledge have been accumulated. However, challenges such as data silos and fragmented knowledge hinder the effective utilization of this information. This study proposes a novel framework for intelligent Question-and-Answer (Q&A) systems based on Retrieval-Augmented Generation (RAG) to address these issues. The system efficiently acquires domain-specific knowledge by leveraging external databases, including Relational Databases (RDBs) and graph databases, without additional fine-tuning for Large Language Models (LLMs). Crucially, the framework integrates a Dynamic Knowledge Base Updating Mechanism (DKBUM) and a Weighted Context-Aware Similarity (WCAS) method to enhance retrieval accuracy and mitigate inherent limitations of LLMs, such as hallucinations and lack of specialization. Additionally, the proposed DKBUM dynamically adjusts knowledge weights within the database, ensuring that the most recent and relevant information is utilized, while WCAS refines the alignment between queries and knowledge items by enhanced context understanding. Experimental validation demonstrates that the system can generate timely, accurate, and context-sensitive responses, making it a robust solution for managing complex business logic in specialized industries.展开更多
End-to-end object detection Transformer(DETR)successfully established the paradigm of the Transformer architecture in the field of object detection.Its end-to-end detection process and the idea of set prediction have ...End-to-end object detection Transformer(DETR)successfully established the paradigm of the Transformer architecture in the field of object detection.Its end-to-end detection process and the idea of set prediction have become one of the hottest network architectures in recent years.There has been an abundance of work improving upon DETR.However,DETR and its variants require a substantial amount of memory resources and computational costs,and the vast number of parameters in these networks is unfavorable for model deployment.To address this issue,a greedy pruning(GP)algorithm is proposed,applied to a variant denoising-DETR(DN-DETR),which can eliminate redundant parameters in the Transformer architecture of DN-DETR.Considering the different roles of the multi-head attention(MHA)module and the feed-forward network(FFN)module in the Transformer architecture,a modular greedy pruning(MGP)algorithm is proposed.This algorithm separates the two modules and applies their respective optimal strategies and parameters.The effectiveness of the proposed algorithm is validated on the COCO 2017 dataset.The model obtained through the MGP algorithm reduces the parameters by 49%and the number of floating point operations(FLOPs)by 44%compared to the Transformer architecture of DN-DETR.At the same time,the mean average precision(mAP)of the model increases from 44.1%to 45.3%.展开更多
The cooling rate of the center and edge of vacuum induction melting(VIM)or vacuum arc remelting(VAR)ingots exhibit substantial difference,leading to markedly distinct dendritic structures and precipitates.The current ...The cooling rate of the center and edge of vacuum induction melting(VIM)or vacuum arc remelting(VAR)ingots exhibit substantial difference,leading to markedly distinct dendritic structures and precipitates.The current lack of precise predictions for dendritic segregation and the distribution of precipitates in ingot makes it difficult to determine the annealing and homogenization heat treatment process.Thus,clarifying the impact of cooling rate on the solidification behavior of alloy is significantly important.The dendritic structure and precipitation characteristics of as-cast C-HRA-3 Ni–Cr–Co–Mo-based heat-resistant alloy were investigated using Thermo-Calc thermodynamic calculations,scanning electron microscopy observations,and electron probe microanalyzer.Based on high temperature observation system,the effects of cooling rate on the dendritic structure,dendritic segregation,and precipitation in this alloy were explored.The results showed that the precipitates in the as-cast C-HRA-3 alloy primarily consist of blocky Ti(C,N)phases,large-sized Ti(C,N)–M_(6)C–M_(23)C_(6) symbiotic phases and M_(6)C–M_(23)C_(6) carbides,and small-sized dispersed M_(6)C and M_(23)C_(6) carbides surronding these symbiotic phases.The primary constituent elements of these precipitates are Mo,Cr,C,and Ti,which predominantly concentrate in the interdendritic regions of the as-cast alloy.There is a clear power-law relationship between the secondary dendrite arm spacing and the cooling rate.The dendritic segregation ratio of Mo,Cr,and Ti exhibits a piecewise functional relationship with the cooling rate,under equiaxed dendritic solidification condition.These predictive models and theoretical analyses were validated using numerical simulations and experimental results from the 200 kg grade VIM electrode.展开更多
Automated detection of suspended anomalous objects on high-speed railway catenary systems using computer vision-based technology is a critical task for ensuring railway transportation safety. Despite the critical impo...Automated detection of suspended anomalous objects on high-speed railway catenary systems using computer vision-based technology is a critical task for ensuring railway transportation safety. Despite the critical importance of this task, conventional vision-based foreign object detection methodologies have predominantly concentrated on image data, neglecting the exploration and integration of textual information. The currently popular multimodal model Contrastive Language-Image Pre-training (CLIP) employs contrastive learning to enable simultaneous understanding of both visual and textual modalities. Drawing inspiration from CLIP’s capabilities, this paper introduces a novel CLIP-based multimodal foreign object detection model tailored for railway applications, referred to as Railway-CLIP. This model leverages CLIP’s robust generalization capabilities to enhance performance in the context of catenary foreign object detection. The Railway-CLIP model is primarily composed of an image encoder and a text encoder. Initially, the Segment Anything Model (SAM) is employed to preprocess raw images, identifying candidate bounding boxes that may contain foreign objects. Both the original images and the detected candidate bounding boxes are subsequently fed into the image encoder to extract their respective visual features. In parallel, distinct prompt templates are crafted for both the original images and the candidate bounding boxes to serve as textual inputs. These prompts are then processed by the text encoder to derive textual features. The image and text encoders collaboratively project the multimodal features into a shared semantic space, facilitating the computation of similarity scores between visual and textual representations. The final detection results are determined based on these similarity scores, ensuring a robust and accurate identification of anomalous objects. Extensive experiments on our collected Railway Anomaly Dataset (RAD) demonstrate that the proposed Railway-CLIP outperforms previous state-of-the-art methods, achieving 97.25% AUROC and 92.66% F1-score, thereby validating the effectiveness and superiority of the proposed approach in real-world high-speed railway anomaly detection scenarios.展开更多
The rise in construction activities within mountainous regions has significantly increased the frequency of rockfalls.Statistical models for rockfall hazard assessment often struggle to achieve high precision on a lar...The rise in construction activities within mountainous regions has significantly increased the frequency of rockfalls.Statistical models for rockfall hazard assessment often struggle to achieve high precision on a large scale.This limitation arises primarily from the scarcity of historical rockfall data and the inadequacy of conventional assessment indicators in capturing the physical and structural characteristics of rockfalls.This study proposes a physically based deterministic model designed to accurately quantify rockfall hazards at a large scale.The model accounts for multiple rockfall failure modes and incorporates the key physical and structural parameters of the rock mass.Rockfall hazard is defined as the product of three factors:the rockfall failure probability,the probability of reaching a specific position,and the corresponding impact intensity.The failure probability includes probabilities of formation and instability of rock blocks under different failure modes,modeled based on the combination patterns of slope surfaces and rock discontinuities.The Monte Carlo method is employed to account for the randomness of mechanical and geometric parameters when quantifying instability probabilities.Additionally,the rock trajectories and impact energies simulated using Flow-R software are combined with rockfall failure probability to enable regional rockfall hazard zoning.A case study was conducted in Tiefeng,Chongqing,China,considering four types of rockfall failure modes.Hazard zoning results identified the steep and elevated terrains of the northern and southern anaclinal slopes as areas of highest rockfall hazard.These findings align with observed conditions,providing detailed hazard zoning and validating the effectiveness and potential of the proposed model.展开更多
Tau plays a crucial role in several neurodegenerative diseases,collectively referred to as tauopathies.Therefore,targeting potential pathological changes in tau could enable useful therapeutic interventions.However,ta...Tau plays a crucial role in several neurodegenerative diseases,collectively referred to as tauopathies.Therefore,targeting potential pathological changes in tau could enable useful therapeutic interventions.However,tau is not an easy target because it dynamically interacts with microtubules and other cellular components,which presents a challenge for tau-targeted drugs.New cellular models could aid the development of mechanism-based tau-targeted therapies.展开更多
文摘Offline policy evaluation,evaluating and selecting complex policies for decision-making by only using offline datasets is important in reinforcement learning.At present,the model-based offline policy evaluation(MBOPE)is widely welcomed because of its easy to implement and good performance.MBOPE directly approximates the unknown value of a given policy using the Monte Carlo method given the estimated transition and reward functions of the environment.Usually,multiple models are trained,and then one of them is selected to be used.However,a challenge remains in selecting an appropriate model from those trained for further use.The authors first analyse the upper bound of the difference between the approximated value and the unknown true value.Theoretical results show that this difference is related to the trajectories generated by the given policy on the learnt model and the prediction error of the transition and reward functions at these generated data points.Based on the theoretical results,a new criterion is proposed to tell which trained model is better suited for evaluating the given policy.At last,the effectiveness of the proposed criterion is demonstrated on both benchmark and synthetic offline datasets.
文摘To address the challenges of low accuracy and insufficient real-time performance in dynamic object detection for UAV surveillance,this paper proposes a novel tracking framework that integrates a lightweight improved YOLOv5s model with adaptive motion compensation.A UAV-view dynamic feature enhancement strategy is innovatively introduced,and a lightweight detection network combining attention mechanisms and multi-scale fusion is constructed.The robustness of tracking under motion blur scenarios is also optimized.Experimental results demonstrate that the proposed method achieves a mAP@0.5 of 68.2%on the VisDrone dataset and reaches an inference speed of 32 FPS on the NVIDIA Jetson TX2 platform.This significantly improves the balance between accuracy and efficiency in complex scenes,offering reliable technical support for real-time applications such as emergency response.
基金supported in part by theThe Planning Subject Project of Guangdong Power Grid Co.,Ltd.(62273104).
文摘With the large-scale promotion of distributed photovoltaics,new challenges have emerged in the photovoltaic consumptionwithin distribution networks.Traditional photovoltaic consumption schemes have primarily focused on static analysis.However,as the scale of photovoltaic power generation devices grows and the methods of integration diversify,a single consumption scheme is no longer sufficient to meet the actual needs of current distribution networks.Therefore,this paper proposes an optimal evaluation method for photovoltaic consumption schemes based on BASS model predictions of installed capacity,aiming to provide an effective tool for generating and evaluating photovoltaic consumption schemes in distribution networks.First,the BASS diffusion model,combined with existing photovoltaic capacity data and roof area information,is used to predict the trends in photovoltaic installed capacity for each substation area,providing a scientific basis for consumption evaluation.Secondly,an improved random scenario simulation method is proposed for assessing the photovoltaic consumption capacity in distribution networks.This method generates photovoltaic integration schemes based on the diffusion probabilities of different regions and evaluates the consumption capacity of each scheme.Finally,the Technique for Order Preference by Similarity to an Ideal Solution(TOPSIS)is used to comprehensively evaluate the generated schemes,ensuring that the selected scheme not only meets the consumption requirements but also offers high economic benefits and reliability.The effectiveness and feasibility of the proposedmethod are validated through simulations of the IEEE 33-node system,providing strong support for optimizing photovoltaic consumption schemes in distribution networks.
文摘A large number of 3D models are created on computers and available for networks. Some content-based retrieval technologies are indispensable to find out particular data from such anonymous datasets. Though several shape retrieval technologies have been developed, little attention has been given to the points on human＇s sense and impression （as known as Kansei） in the conventional techniques, In this paper, the authors propose a novel method of shape retrieval based on shape impression of human＇s Kansei. The key to the method is the Gaussian curvature distribution from 3D models as features for shape retrieval. Then it classifies the 3D models by extracted feature and measures similarity among models in storage.
文摘This paper proposes a method to recognize human-object interactions by modeling context between human actions and interacted objects.Human-object interaction recognition is a challenging task due to severe occlusion between human and objects during the interacting process.Since that human actions and interacted objects provide strong context information,i.e.some actions are usually related to some specific objects,the accuracy of recognition is significantly improved for both of them.Through the proposed method,both global and local temporal features from skeleton sequences are extracted to model human actions.In the meantime,kernel features are utilized to describe interacted objects.Finally,all possible solutions from actions and objects are optimized by modeling the context between them.The results of experiments demonstrate the effectiveness of our method.
文摘The increasing global demand for sustainable agricultural practices and effective waste management has highlighted the potential of biochar as a multifaceted solution. This study evaluates the economic viability of sugarcane bagasse-based biochar in Brazil, focusing on its potential to enhance agricultural productivity and contribute to environmental sustainability. While existing literature predominantly explores the production, crop yield benefits, and carbon sequestration capabilities of biochar, there is a notable gap in comprehensive economic modeling and viability analysis for the region. This paper aims to fill this gap by employing a scenario-based economic modeling approach, incorporating relevant economic models. Findings include that biochar implementation can be economically viable for medium and large sugarcane farms (20,000 - 50,000 hectares) given the availability of funding, breaking even in about 7.5 years with an internal rate of return of 18% on average. For small farms, biochar can only be viable when applied biochar to the soil, which in all scenarios is found to be the more profitable practice by a large margin. Sensitivity analyses found that generally, biochar becomes economically feasible at biochar carbon credit prices above $120 USD/tCO2e, and at sugarcane bagasse availability percentages above 60%. While the economic models are well-grounded in existing literature, the production of biochar at the studied scales is not yet widespread, especially in Brazil and uncertainties can result. Reviewing the results, the land application scenario was found to be the most viable, and large farms saw the best results, highlighting the importance of scale in biochar operations. Small and medium farms with no land application were concluded to have no or questionable viability. Overall, sugarcane bagasse-based biochar can be economically viable, under the right circumstances, for agricultural and environmental advancement in Brazil.
基金supported by National Natural Science Foundation of China (Grant No. 60873003)
文摘Recently automotive nets are adopted to solve increasing problems in automotive electronic systems.Technologies of automotive local area network from CAN and LIN can solve the problems of the increasing of wire bunch weight and lack in module installation space.However,the multilayer automotive nets software becomes more and more complex,and the development expense is difficult to predict and to keep in check.In this paper,the modeling method of hierarchical automotive nets and the substitution operation based on object-oriented colored Petri net(OOCPN) are proposed.The OOCPN model which analyzes the software structure and validates the collision mechanism of CAN/LIN bus can speed the automobile system development.First,the subsystems are divided and modeled by object-oriented Petri net(OOPN).According to the sets of message sharing relations,the message ports among them are set and the communication gate transitions are defined.Second,the OOPN model is substituted step by step until the inner objects in the automotive body control modules(BCM) are indivisible and colored by colored Petri net(CPN).And the color subsets mark the node messages for the collision mechanism.Third,the OOCPN model of the automotive body CAN/LIN nets is assembled,which keeps the message sets and the system can be expanded.The proposed model is used to analyze features of information sharing among the objects,and it is also used to describe each subsystem real-time behavior of processing messages and implemental device controllers operating,and puts forward a reasonable software framework for the automotive body control subsystem.The research can help to design the communication model in the automotive body system effectively and provide a convenient and rapid way for developing the logical hierarchy software.
基金supported by National Basic Research Program of China(973 Program,Grant No.2011CB706502)
文摘Modelica-based object-orient method is proved to be rapid, accurate and easy to modify, which is suitable for prototype modeling and simulation of rotor system, whose parameters need to be modified frequently. Classical non-object-orient method appears to be inefficient because the code is difficult to modify and reuse. An adequate library for object-orient modeling of rotor system with multi-faults is established, a comparison with non-object-orient method on Jeffcott rotor system and a case study on turbo expander with multi-faults are implemented. The relative tolerance between object-orient method and non-object-orient is less than 0.03%, which proves that these two methods are as accurate as each other. Object-orient modeling and simulation is implemented on turbo expander with crack, rub-impact, pedestal looseness and multi-faults simultaneously. It can be conclude from the case study that when acting on compress side of turbo expander separately, expand wheel is not influenced greatly by crack fault, the existence of rub-impact fault forces expand wheel into quasi-periodic motion and the orbit of expand wheel is deformed and enhanced almost 1.5 times due to pedestal looseness. When acting simultaneously, multi-faults cannot be totally decomposed but can be diagnosed from the feature of vibration. Object-orient method can enhance the efficiency of modeling and simulation of rotor system with multi-faults, which provides an efficient method on prototype modeling and simulation.
基金This work was supported,in part,by the National Nature Science Foundation of China under grant numbers 62272236in part,by the Natural Science Foundation of Jiangsu Province under grant numbers BK20201136,BK20191401in part,by the Priority Academic Program Development of Jiangsu Higher Education Institutions(PAPD)fund.
文摘LIDAR point cloud-based 3D object detection aims to sense the surrounding environment by anchoring objects with the Bounding Box(BBox).However,under the three-dimensional space of autonomous driving scenes,the previous object detection methods,due to the pre-processing of the original LIDAR point cloud into voxels or pillars,lose the coordinate information of the original point cloud,slow detection speed,and gain inaccurate bounding box positioning.To address the issues above,this study proposes a new two-stage network structure to extract point cloud features directly by PointNet++,which effectively preserves the original point cloud coordinate information.To improve the detection accuracy,a shell-based modeling method is proposed.It roughly determines which spherical shell the coordinates belong to.Then,the results are refined to ground truth,thereby narrowing the localization range and improving the detection accuracy.To improve the recall of 3D object detection with bounding boxes,this paper designs a self-attention module for 3D object detection with a skip connection structure.Some of these features are highlighted by weighting them on the feature dimensions.After training,it makes the feature weights that are favorable for object detection get larger.Thus,the extracted features are more adapted to the object detection task.Extensive comparison experiments and ablation experiments conducted on the KITTI dataset verify the effectiveness of our proposed method in improving recall and precision.
文摘In the developing course of the information system, the modeling method of information system and model expressing problem are very important. This paper through discussing the facing object of Petri network, expatiates the basic problem of how to set up the business procedure model on the basis of Petri network, and shows the expressing and storing methods of the model in computer way.
文摘Use of features in order to achieve the integration of design and manufacture has been considered to be a key factor recent years. Features such as manufacturing properties form the workpiece. Features are structured systematically through object oriented modeling. This article explains an object coding method developed for prismatic workpieces and the use of that method in process planning. Features have been determined and modeled as objects. Features have been coded according to their types and locations on the workpiece in this given method. Feature codings have been seen to be very advantageous in process planning.
基金supported by the Outstanding Youth Team Project of Central Universities(QNTD202308)the Ant Group through CCF-Ant Research Fund(CCF-AFSG 769498 RF20220214).
文摘Named Entity Recognition(NER)stands as a fundamental task within the field of biomedical text mining,aiming to extract specific types of entities such as genes,proteins,and diseases from complex biomedical texts and categorize them into predefined entity types.This process can provide basic support for the automatic construction of knowledge bases.In contrast to general texts,biomedical texts frequently contain numerous nested entities and local dependencies among these entities,presenting significant challenges to prevailing NER models.To address these issues,we propose a novel Chinese nested biomedical NER model based on RoBERTa and Global Pointer(RoBGP).Our model initially utilizes the RoBERTa-wwm-ext-large pretrained language model to dynamically generate word-level initial vectors.It then incorporates a Bidirectional Long Short-Term Memory network for capturing bidirectional semantic information,effectively addressing the issue of long-distance dependencies.Furthermore,the Global Pointer model is employed to comprehensively recognize all nested entities in the text.We conduct extensive experiments on the Chinese medical dataset CMeEE and the results demonstrate the superior performance of RoBGP over several baseline models.This research confirms the effectiveness of RoBGP in Chinese biomedical NER,providing reliable technical support for biomedical information extraction and knowledge base construction.
基金supported by the Tianjin Postgraduate Research Innovation Project (No.2022SKY286)the National Science and the National Key Research and Development Program (No.2022YFF0706000)。
文摘Aiming at the problem of low surface defect detection accuracy of industrial products, an object detection method based on simplified spatial pyramid pooling fast(Sim SPPF) hybrid pooling improved you only look once version 5s(YOLOV5s) model is proposed. The algorithm introduces channel attention(CA) module, simplified SPPF feature vector pyramid and efficient intersection over union(EIOU) loss function. Feature vector pyramids fuse high-dimensional and low-dimensional features, which makes semantic information richer. The CA mechanism performs maximum pooling and average pooling operations on the feature map. Hybrid pooling comprehensively improves detection computing efficiency and accurate deployment ability. The results show that the improved YOLOV5s model is better than the original YOLOV5s model. The average test accuracy(mAP) can reach 91.8%, which can be increased by 17.4%, and the detection speed can reach 108 FPS, which can be increased by 18 FPS. The improved model is practicable, and the overall performance is better than other conventional models.
文摘In the context of power generation companies, vast amounts of specialized data and expert knowledge have been accumulated. However, challenges such as data silos and fragmented knowledge hinder the effective utilization of this information. This study proposes a novel framework for intelligent Question-and-Answer (Q&A) systems based on Retrieval-Augmented Generation (RAG) to address these issues. The system efficiently acquires domain-specific knowledge by leveraging external databases, including Relational Databases (RDBs) and graph databases, without additional fine-tuning for Large Language Models (LLMs). Crucially, the framework integrates a Dynamic Knowledge Base Updating Mechanism (DKBUM) and a Weighted Context-Aware Similarity (WCAS) method to enhance retrieval accuracy and mitigate inherent limitations of LLMs, such as hallucinations and lack of specialization. Additionally, the proposed DKBUM dynamically adjusts knowledge weights within the database, ensuring that the most recent and relevant information is utilized, while WCAS refines the alignment between queries and knowledge items by enhanced context understanding. Experimental validation demonstrates that the system can generate timely, accurate, and context-sensitive responses, making it a robust solution for managing complex business logic in specialized industries.
基金Shanghai Municipal Commission of Economy and Information Technology,China(No.202301054)。
文摘End-to-end object detection Transformer(DETR)successfully established the paradigm of the Transformer architecture in the field of object detection.Its end-to-end detection process and the idea of set prediction have become one of the hottest network architectures in recent years.There has been an abundance of work improving upon DETR.However,DETR and its variants require a substantial amount of memory resources and computational costs,and the vast number of parameters in these networks is unfavorable for model deployment.To address this issue,a greedy pruning(GP)algorithm is proposed,applied to a variant denoising-DETR(DN-DETR),which can eliminate redundant parameters in the Transformer architecture of DN-DETR.Considering the different roles of the multi-head attention(MHA)module and the feed-forward network(FFN)module in the Transformer architecture,a modular greedy pruning(MGP)algorithm is proposed.This algorithm separates the two modules and applies their respective optimal strategies and parameters.The effectiveness of the proposed algorithm is validated on the COCO 2017 dataset.The model obtained through the MGP algorithm reduces the parameters by 49%and the number of floating point operations(FLOPs)by 44%compared to the Transformer architecture of DN-DETR.At the same time,the mean average precision(mAP)of the model increases from 44.1%to 45.3%.
基金funded by the National Key R&D Program Funded Projects(No.2021YFB3704102).
文摘The cooling rate of the center and edge of vacuum induction melting(VIM)or vacuum arc remelting(VAR)ingots exhibit substantial difference,leading to markedly distinct dendritic structures and precipitates.The current lack of precise predictions for dendritic segregation and the distribution of precipitates in ingot makes it difficult to determine the annealing and homogenization heat treatment process.Thus,clarifying the impact of cooling rate on the solidification behavior of alloy is significantly important.The dendritic structure and precipitation characteristics of as-cast C-HRA-3 Ni–Cr–Co–Mo-based heat-resistant alloy were investigated using Thermo-Calc thermodynamic calculations,scanning electron microscopy observations,and electron probe microanalyzer.Based on high temperature observation system,the effects of cooling rate on the dendritic structure,dendritic segregation,and precipitation in this alloy were explored.The results showed that the precipitates in the as-cast C-HRA-3 alloy primarily consist of blocky Ti(C,N)phases,large-sized Ti(C,N)–M_(6)C–M_(23)C_(6) symbiotic phases and M_(6)C–M_(23)C_(6) carbides,and small-sized dispersed M_(6)C and M_(23)C_(6) carbides surronding these symbiotic phases.The primary constituent elements of these precipitates are Mo,Cr,C,and Ti,which predominantly concentrate in the interdendritic regions of the as-cast alloy.There is a clear power-law relationship between the secondary dendrite arm spacing and the cooling rate.The dendritic segregation ratio of Mo,Cr,and Ti exhibits a piecewise functional relationship with the cooling rate,under equiaxed dendritic solidification condition.These predictive models and theoretical analyses were validated using numerical simulations and experimental results from the 200 kg grade VIM electrode.
基金supported by the Technology Research and Development Program of China National Railway Group(Q2024T002)the Open Project Fund of National Engineering Research Center of Digital Construction and Evaluation Technology of Urban Rail Transit(2024023).
文摘Automated detection of suspended anomalous objects on high-speed railway catenary systems using computer vision-based technology is a critical task for ensuring railway transportation safety. Despite the critical importance of this task, conventional vision-based foreign object detection methodologies have predominantly concentrated on image data, neglecting the exploration and integration of textual information. The currently popular multimodal model Contrastive Language-Image Pre-training (CLIP) employs contrastive learning to enable simultaneous understanding of both visual and textual modalities. Drawing inspiration from CLIP’s capabilities, this paper introduces a novel CLIP-based multimodal foreign object detection model tailored for railway applications, referred to as Railway-CLIP. This model leverages CLIP’s robust generalization capabilities to enhance performance in the context of catenary foreign object detection. The Railway-CLIP model is primarily composed of an image encoder and a text encoder. Initially, the Segment Anything Model (SAM) is employed to preprocess raw images, identifying candidate bounding boxes that may contain foreign objects. Both the original images and the detected candidate bounding boxes are subsequently fed into the image encoder to extract their respective visual features. In parallel, distinct prompt templates are crafted for both the original images and the candidate bounding boxes to serve as textual inputs. These prompts are then processed by the text encoder to derive textual features. The image and text encoders collaboratively project the multimodal features into a shared semantic space, facilitating the computation of similarity scores between visual and textual representations. The final detection results are determined based on these similarity scores, ensuring a robust and accurate identification of anomalous objects. Extensive experiments on our collected Railway Anomaly Dataset (RAD) demonstrate that the proposed Railway-CLIP outperforms previous state-of-the-art methods, achieving 97.25% AUROC and 92.66% F1-score, thereby validating the effectiveness and superiority of the proposed approach in real-world high-speed railway anomaly detection scenarios.
基金supported by the National Natural Science Foundation of China(Grant Nos.42172318 and 42377186)the National Key R&D Program of China(Grant No.2023YFC3007201).
文摘The rise in construction activities within mountainous regions has significantly increased the frequency of rockfalls.Statistical models for rockfall hazard assessment often struggle to achieve high precision on a large scale.This limitation arises primarily from the scarcity of historical rockfall data and the inadequacy of conventional assessment indicators in capturing the physical and structural characteristics of rockfalls.This study proposes a physically based deterministic model designed to accurately quantify rockfall hazards at a large scale.The model accounts for multiple rockfall failure modes and incorporates the key physical and structural parameters of the rock mass.Rockfall hazard is defined as the product of three factors:the rockfall failure probability,the probability of reaching a specific position,and the corresponding impact intensity.The failure probability includes probabilities of formation and instability of rock blocks under different failure modes,modeled based on the combination patterns of slope surfaces and rock discontinuities.The Monte Carlo method is employed to account for the randomness of mechanical and geometric parameters when quantifying instability probabilities.Additionally,the rock trajectories and impact energies simulated using Flow-R software are combined with rockfall failure probability to enable regional rockfall hazard zoning.A case study was conducted in Tiefeng,Chongqing,China,considering four types of rockfall failure modes.Hazard zoning results identified the steep and elevated terrains of the northern and southern anaclinal slopes as areas of highest rockfall hazard.These findings align with observed conditions,providing detailed hazard zoning and validating the effectiveness and potential of the proposed model.
文摘Tau plays a crucial role in several neurodegenerative diseases,collectively referred to as tauopathies.Therefore,targeting potential pathological changes in tau could enable useful therapeutic interventions.However,tau is not an easy target because it dynamically interacts with microtubules and other cellular components,which presents a challenge for tau-targeted drugs.New cellular models could aid the development of mechanism-based tau-targeted therapies.
