MKEAH:Multimodal knowledge extraction and accumulation based on hyperplane embedding for knowledge-based visual question answering

Background External knowledge representations play an essential role in knowledge-based visual question and answering to better understand complex scenarios in the open world.Recent entity-relationship embedding approaches are deficient in representing some complex relations,resulting in a lack of topic-related knowledge and redundancy in topic-irrelevant information.Methods To this end,we propose MKEAH:Multimodal Knowledge Extraction and Accumulation on Hyperplanes.To ensure that the lengths of the feature vectors projected onto the hyperplane compare equally and to filter out sufficient topic-irrelevant information,two losses are proposed to learn the triplet representations from the complementary views:range loss and orthogonal loss.To interpret the capability of extracting topic-related knowledge,we present the Topic Similarity(TS)between topic and entity-relations.Results Experimental results demonstrate the effectiveness of hyperplane embedding for knowledge representation in knowledge-based visual question answering.Our model outperformed state-of-the-art methods by 2.12%and 3.24%on two challenging knowledge-request datasets:OK-VQA and KRVQA,respectively.Conclusions The obvious advantages of our model in TS show that using hyperplane embedding to represent multimodal knowledge can improve its ability to extract topic-related knowledge.

Knowledge-Based Efficient N-1 Analysis Calculation Method for Urban Distribution Networks with CIM File Data

The N-1 criterion is a critical factor for ensuring the reliable and resilient operation of electric power distribution networks.However,the increasing complexity of distribution networks and the associated growth in data size have created a significant challenge for distribution network planners.To address this issue,we propose a fast N-1 verification procedure for urban distribution networks that combines CIM file data analysis with MILP-based mathematical modeling.Our proposed method leverages the principles of CIM file analysis for distribution network N-1 analysis.We develop a mathematical model of distribution networks based on CIM data and transfer it into MILP.We also take into account the characteristics of medium voltage distribution networks after a line failure and select the feeder section at the exit of each substation with a high load rate to improve the efficiency of N-1 analysis.We validate our approach through a series of case studies and demonstrate its scalability and superiority over traditional N-1 analysis and heuristic optimization algorithms.By enabling online N-1 analysis,our approach significantly improves the work efficiency of distribution network planners.In summary,our proposed method provides a valuable tool for distribution network planners to enhance the accuracy and efficiency of their N-1 analyses.By leveraging the advantages of CIM file data analysis and MILP-based mathematical modeling,our approach contributes to the development of more resilient and reliable electric power distribution networks.

Knowledge-Based Classification in Automated Soil Mapping

A machine-learning approach was developed for automated building of knowledgebases for soil resources mapping by using a classification tree to generate knowledge from trainingdata. With this method, building a knowledge base for automated soil mapping was easier than usingthe conventional knowledge acquisition approach. The knowledge base built by classification tree wasused by the knowledge classifier to perform the soil type classification of Longyou County,Zhejiang Province, China using Landsat TM bi-temporal images and CIS data. To evaluate theperformance of the resultant knowledge bases, the classification results were compared to existingsoil map based on a field survey. The accuracy assessment and analysis of the resultant soil mapssuggested that the knowledge bases built by the machine-learning method was of good quality formapping distribution model of soil classes over the study area.

Knowledge-based bridge detection from SAR images

Automatic bridge detection is an important application of SAR images. Differed from the classical CFAR method, a new knowledge-based bridge detection approach is proposed. The method not only uses the backscattering intensity difference between targets and background but also applies the contextual information and spatial relationship between objects. According to bridges＇ special characteristics and scattering properties in SAR images, the new knowledge-based method includes three processes： river segmentation, potential bridge areas detection and bridge discrimination. The application to AIRSAR data shows that the new method is not sensitive to rivers＇ shape. Moreover, this method can detect bridges successfully when river segmentation is not very exact and is more robust than the radius projection method.

New Knowledge-based Genetic Algorithm for Excavator Boom Structural Optimization

Due to the insufficiency of utilizing knowledge to guide the complex optimal searching, existing genetic algorithms fail to effectively solve excavator boom structural optimization problem. To improve the optimization efficiency and quality, a new knowledge-based real-coded genetic algorithm is proposed. A dual evolution mechanism combining knowledge evolution with genetic algorithm is established to extract, handle and utilize the shallow and deep implicit constraint knowledge to guide the optimal searching of genetic algorithm circularly. Based on this dual evolution mechanism, knowledge evolution and population evolution can be connected by knowledge influence operators to improve the conflgurability of knowledge and genetic operators. Then, the new knowledge-based selection operator, crossover operator and mutation operator are proposed to integrate the optimal process knowledge and domain culture to guide the excavator boom structural optimization. Eight kinds of testing algorithms, which include different genetic operators, arc taken as examples to solve the structural optimization of a medium-sized excavator boom. By comparing the results of optimization, it is shown that the algorithm including all the new knowledge-based genetic operators can more remarkably improve the evolutionary rate and searching ability than other testing algorithms, which demonstrates the effectiveness of knowledge for guiding optimal searching. The proposed knowledge-based genetic algorithm by combining multi-level knowledge evolution with numerical optimization provides a new effective method for solving the complex engineering optimization problem.

Structural Topology Design of Container Ship Based on Knowledge-Based Engineering and Level Set Method

Knowledge-Based Engineering （KBE） is introduced into the ship structural design in this paper. From the implementation of KBE, the design solutions for both Rules Design Method （RDM） and Interpolation Design Method （IDM） are generated. The corresponding Finite Element （FE） models are generated. Topological design of the longitudinal structures is studied where the Gaussian Process （GP） is employed to build the surrogate model for FE analysis. Multi-objective optimization methods inspired by Pareto Front are used to reduce the design tank weight and outer surface area simultaneously. Additionally, an enhanced Level Set Method （LSM） which employs implicit algorithm is applied to the topological design of typical bracket plate which is used extensively in ship structures. Two different sets of boundary conditions are considered. The proposed methods show satisfactory efficiency and accuracy.

FAULT DIAGNOSIS OF ROTATING MACHINERY USING KNOWLEDGE-BASED FUZZY NEURAL NETWORK

A novel knowledge-based fuzzy neural network （KBFNN） for fault diagnosis is presented. Crude rules were extracted and the corresponding dependent factors and antecedent coverage factors were calculated firstly from the diagnostic sample based on rough sets theory. Then the number of rules was used to construct partially the structure of a fuzzy neural network and those factors were implemented as initial weights, with fuzzy output parameters being optimized by genetic algorithm. Such fuzzy neural network was called KBFNN. This KBFNN was utilized to identify typical faults of rotating machinery. Diagnostic results show that it has those merits of shorter training time and higher right diagnostic level compared to general fuzzy neural networks.

Knowledge-based potentials in bioinformatics: From a physicist's viewpoint

Biological raw data are growing exponentially, providing a large amount of information on what life is. It is believed that potential functions and the rules governing protein behaviors can be revealed from analysis on known native structures of proteins. Many knowledge-based potentials for proteins have been proposed. Contrary to most existing review articles which mainly describe technical details and applications of various potential models, the main foci for the discussion here are ideas and concepts involving the construction of potentials, including the relation between free energy and energy, the additivity of potentials of mean force and some key issues in potential construction. Sequence analysis is briefly viewed from an energetic viewpoint.

A Knowledge-based System for the Analysis of the Ability of Paying back Loans

This paper describes the development of a knowledgebased system (KBS) for determining whether or not, and under what conditions, a bank Ioan officer should grant a business loan to a company. The prototype system developed focuses on what is bank loans risks management, how to prevent risk by the analysis of the ability of paying back loans. The paper makes the structural analysis involved in the system's decision situation, the structured situation diagram or model, dependency diagram and the document needed by the KBS prototype system thus are developed. Through testing the samples from loan business, the quality for the analysis of the ability of paying back loans can be effectively evaluated by the KBS prototype system.

Integration of Decision-Support and Knowledge-Based Techniques in a Problem-Solving Strategy for Selection Problems

The limitations of traditional approaches to selection problems are examined. A problemsolving strategy is presented in which decision-support and knowledge-based techniques play complementary roles. An approach to the representation of knowledge to support the problem-solving strategy is presented which avoids commitment to a specific programming language or implementation environment. The problem of choosing a home is used to illustrate the representation of knowledge in a specific problem domain. Techniques for implementation of the problem-solving strategy are described. Knowledge elicitation techniques and their implementation in a development shell for application of the problem-solving strategy to any selection problem are also described.

Yarn Quality Prediction and Diagnosis Based on Rough Set and Knowledge-Based Artificial Neural Network

In the spinning process, some key process parameters( i. e.,raw material index inputs) have very strong relationship with the quality of finished products. The abnormal changes of these process parameters could result in various categories of faulty products. In this paper, a hybrid learning-based model was developed for on-line intelligent monitoring and diagnosis of the spinning process. In the proposed model, a knowledge-based artificial neural network( KBANN) was developed for monitoring the spinning process and recognizing faulty quality categories of yarn. In addition,a rough set( RS)-based rule extraction approach named RSRule was developed to discover the causal relationship between textile parameters and yarn quality. These extracted rules were applied in diagnosis of the spinning process, provided guidelines on improving yarn quality,and were used to construct KBANN. Experiments show that the proposed model significantly improve the learning efficiency, and its prediction precision is improved by about 5. 4% compared with the BP neural network model.

MY PERSONAL STUDY ON KNOWLEDGE-BASED ECONOMY

The Financial Crisis in Asia is having a negative impacion the economic development of China, but it also enlightens us. It makes us consider and take measures to avoid such a crisis. I have put forward six measures, one of which is to promote the transformation of S&T knowledge into productive forces.

Research on Knowledge-based Connecting Elements Modeling

Steps of manipulation is required to complete the m od eling of the connection elements such as bolt, pin and the like in commerce CAD system. It leads to low efficiency, difficulty to assure the relative position, impossibility to express rules and knowledge. Based on the inner character analy sis of interpart, detail modification and assembly relation of mechanical connec ting element, the idea, which extends the feature modeling of part to the interp art feature modeling for assembly purpose, is presented, and virtual part based connecting element modeling is proposed. Virtual part is a complement set of lo cal modification of part to be connected. In assembly modeling, base part is mod ified by Boolean operation between base part and virtual part. The modeling and assembly is finished just in one operation, at the same time the rules and knowl edge of the connection elements are encapsulated through virtual part. According to this mechanism, a knowledge-based connecting elements rapid design module i s developed on commerce software package UG with satisfying results.

Optimizing the atom types of proteins through iterative knowledge-based potentials

Knowledge-based scoring functions have been widely used for protein structure prediction, protein-small molecule, and protein-nucleic acid interactions, in which one critical step is to find an appropriate representation of protein structures. A key issue is to determine the minimal protein representations, which is important not only for developing of scoring func- tions but also for understanding the physics of protein folding. Despite significant progresses in simplifying residues into alphabets, few studies have been done to address the optimal number of atom types for proteins. Here, we have investigated the atom typing issue by classifying the 167 heavy atoms of proteins through 11 schemes with 1 to 20 atom types based on their physicochemical and functional environments. For each atom typing scheme, a statistical mechanics-based iterative method was used to extract atomic distance-dependent potentials from protein structures. The atomic distance-dependent pair potentials for different schemes were illustrated by several typical atom pairs with different physicochemical proper- ties. The derived potentials were also evaluated on a high-resolution test set of 148 diverse proteins for native structure recognition. It was found that there was a crossover around the scheme of four atom types in terms of the success rate as a function of the number of atom types, which means that four atom types may be used when investigating the basic folding mechanism of proteins. However, it was revealed by a close examination of typical potentials that 14 atom types were needed to describe the protein interactions at atomic level. The present study will be beneficial for the development of protein related scoring functions and the understanding of folding mechanisms.

A knowledge-based decision support system for engineering design

A knowledge-based decision supporting system, used for engineering design is introduced by describing the architecture, function, workflow of the system and its way of implementation. Based upon information composed of knowledge, models, data, cases, methods, etc, the system is designed to use such methods as knowledge-based reasoning, case-based reasoning, and multi-criteria evaluation techniques to provide effective tools to support the decision-making process.

A novel knowledge-based potential for RNA 3D structure evaluation

Ribonucleic acids(RNAs)play a vital role in biology,and knowledge of their three-dimensional(3D)structure is required to understand their biological functions.Recently structural prediction methods have been developed to address this issue,but a series of RNA 3D structures are generally predicted by most existing methods.Therefore,the evaluation of the predicted structures is generally indispensable.Although several methods have been proposed to assess RNA 3D structures,the existing methods are not precise enough.In this work,a new all-atom knowledge-based potential is developed for more accurately evaluating RNA 3D structures.The potential not only includes local and nonlocal interactions but also fully considers the specificity of each RNA by introducing a retraining mechanism.Based on extensive test sets generated from independent methods,the proposed potential correctly distinguished the native state and ranked near-native conformations to effectively select the best.Furthermore,the proposed potential precisely captured RNA structural features such as base-stacking and base-pairing.Comparisons with existing potential methods show that the proposed potential is very reliable and accurate in RNA 3D structure evaluation.

A Cognitive Science Framework for the Analysis of Knowledge-Based Systems

The paper presents a cognitive science framework for the analysis of knowledge-based systems,including people, media. simulation and expert systems, resulting in a practical model for the procedures ofknowledge engineering. Starting with the construct of a social organization model driven by anticipationand thed differentiating this into pesonal scientists with diverse relations to people and their internal andexternal communication, it provides powerful and general model of society. people, and the roles of peoplein society. This model extends naturally ic the role of conventional media in the knowledge processes ofsociety and the new roles of computer-based simulation and expert systems. In particular it provides amodel of knowledge transfer that enables the processes of knowledge engineering to be analyzed andautomated.

Knowledge-Based Network Management System for Movable and Deployable ICT Resource Unit

When a disaster occurs, the demand for information and communication technology (ICT) services drastically increases. To meet such demands, a national project was undertaken in Japan to develop the Movable and Deployable ICT Resource Unit (MDRU). One challenge regarding the MDRU is securing operators to work the units in emergency situations. As ICT service users have diverse and frequently changing demands, strong technical skills and practical knowledge are required for the administration of MDRUs. In this paper, we propose a knowledge-based network management system to alleviate the burden on administrators. To deal with the structural changes to network systems that frequently occur with changes in ICT service demand, we introduce modularization techniques into our previous research. The proposed system can be easily reconfigured by join/disjoin modules corresponding to changes in the system configuration of the MDRU. The results of our experiments using the implemented experimental system confirm that the proposed system can be applied to MDRU operation and effectively supports administrators.

Knowledge-based engineering approach for the support of ship structure design

The paper presents a knowledge-based engineering （KBE） approach for ship node components design. In the ship design process, many design tasks need design experiences to support. Howev- er, a ship design process is a complicated process with many simultaneously repetitive and time-con- suming activities. In this research, the method combines KBE with Tribon system＇s built-in devel- opment language tools of Vitesse, captures and applies design knowledge for achieving standard com- ponents intelligent design modeling. A case study and industry implementation illustrate the feasibili- ty of the proposed methodology. The KBE technique can provide not only proper references, sug- gests and supports but also knowledge integrated in the ship structure design. Especially, these rules related to the design can avoid lots of design mistakes. During the ship design stage, getting more precise and better designs will not only reduce the time of rework and wasting resources but also shorten the construction time_ imnrov~ clilnl;hz ~nA nrnf;t

Using Rapid Prototyping Data to Enhance a Knowledge-Based Framework for Product Redesign

The particular characteristics of Rapid Prototyping technologies, both in terms of constrains and opportunities, often require the reconfiguration of the product model to obtain the best compliance with the product functionalities and performances. Within this field of research, a knowledge-based tool named Design GuideLines Collaborative Framework (DGLs-CF) was developed to support both the designers defining the product consistently with the manufacturing technologies and the manufacturers defining the building setup consistently with the product requirements. Present work is focused on enhancing the DGLs-CF knowledge base and on updating the DGLs-CF knowledge management by using the information gathered on some RP technologies. The added-value of this research is represented by an improvement in the Redesign/Reconfig- uration Package, the final result of the DGLs-CF adoption. This is a list of actions to be performed on the product model and on the process parameters to avoid the limitations of the technology and to exploit at best its opportunities.