Explanation facility for neural networks

This paper give a methodology, PROLOG code, as well as an example of an explanation facility EN applicable to most neural networks. It involves How?, Why? and TRACE facilities, and is based on a general explanation degree calculation in a multilayer neural network, as well as on input node characterization grammars for synthesis of explanation text.     

Generating probabilistic Boolean networks from a prescribed stationary distribution

Modeling gene regulation is an important problem in genomic research. Boolean networks (BN) and its generalization probabilistic Boolean networks (PBNs) have been proposed to model genetic regulatory interactions. BN is a deterministic model while PBN is a stochastic model. In a PBN, on one hand, its stationary distribution gives important information about the long-run behavior of the network. On the other hand, one may be interested in system synthesis which requires the construction of networks from the observed stationary distribution. This results in an inverse problem which is ill-posed and challenging. Because there may be many networks or no network having the given properties and the size of the inverse problem is huge. In this paper, we consider the problem of constructing PBNs from a given stationary distribution and a set of given Boolean Networks (BNs). We first formulate the inverse problem as a constrained least squares problem. We then propose a heuristic method based on Conjugate Gradient (CG) algorithm, an iterative method, to solve the resulting least squares problem. We also introduce an estimation method for the parameters of the PBNs. Numerical examples are then given to demonstrate the effectiveness of the proposed methods.     

基于分段语义比较的语句相似度计算方法

为了能更为准确地计算两个语句之间的语义相似度;提出了一种基于分段语义比较的语句相似度计算方法.将语句分成主干部分和修饰部分,并根据设定的语法规则,对修饰部分进行更为细致的划分;计算语义相似度时,将计算过程分为两部分:主干部分计算、修饰部分计算,并赋以不同权值;能够按照语法和语义结构,更为合理、细致地计算出两个语句之间的语义相似程度.实验结果表明了该方法的能够取得较好的计算结果.     

Adding equality to semantic tableaux

An extension to the system of semantic tableaux to deal with first-order logic with equality is introduced and proved sound and complete. This involves the use of partial unification, an operation which is based on unification without the presence of variables. We show, further, that semantic tableaux with partial unification provide a sound and complete proof method without needing the functionally reflexive axioms. We also give an example of an ordering rule which allows us to provide less complex proofs in the ground case.     

一种基于知网语义相似度计算的应用研究

随着常见问题集(FAQ)的积累,问句数量日益增大,逐页浏览的知识获取方式很难满足用户的实际需求,也浪费用户大量的精力和时间,而基于FAQ的自动问答系统是解决该问题的一种有效途径。本文设计了QuickFaq自动问答系统框架,采用基于知网的语义相似度计算模块,找出相似问句并返回答案。实验表明,该方法提高了问答系统的召回率和不匹配率。     

语义对等覆盖网中社区结构的发现和评价

最近,通过建立语义覆盖网络来提高大规模分布式网络环境中信息检索服务的性能成为对等计算领域的研究热点.目前,研究者们在语义覆盖协议和搜索算法方面已经做了大量研究,证明了语义覆盖在基于对等网络模型的内容定位应用方面极为有效.然而,分析和评价语义覆盖网络特征的研究工作确非常有限.文中通过建立数学模型和设计启发式回溯-贪婪混合算法、确认了语义覆盖网络的一种主要内在特性——社区结构特性.利用评价模型比较了SemreX语义覆盖网络和Gnutella网络的性能,实验结果显示SemreX覆盖网具有显著的社区结构特征,而Gnutella网络却没有这样的特征.另外,通过分别在两种覆盖网中仿真洪泛协议发现具有显著社区结构特征的覆盖网在内容定位方面效率更高.     

Generating instance models from meta models

Meta modeling is a wide-spread technique to define visual languages, with the UML being the most prominent one. Despite several advantages of meta modeling such as ease of use, the meta modeling approach has one disadvantage: it is not constructive, i.e., it does not offer a direct means of generating instances of the language. This disadvantage poses a severe limitation for certain applications. For example, when developing model transformations, it is desirable to have enough valid instance models available for large-scale testing. Producing such a large set by hand is tedious. In the related problem of compiler testing, a string grammar together with a simple generation algorithm is typically used to produce words of the language automatically. In this paper, we introduce instance-generating graph grammars for creating instances of meta models, thereby overcoming the main deficit of the meta modeling approach for defining languages.

一种基于实时联想的自然语言句子生成方法

在某一特定的领域中,组成自然语言句子的成分之间存在着很强的相互联系,根据这种相互联系可以由句子的中心语联想产生出句子的其它组成成分,由此提出了一种基于实时联想的自然语言句子的生成方法。     

Parallel architectures for AI semantic network processing

Artificial intelligence (AI) applications are growing in several fields, and in many such applications knowledge bases must be manipulated. This activity is usually performed by an external agent such as a central processor, but often this cannot supply the speed required. Knowledge-oriented architectures provide an efficient execution of knowledge manipulations. This paper provides an introduction to a particular subset of knowledge-oriented architectures, the semantic network approach, which is one of the most commonly used methods of representing and manipulating knowledge in the AI field. A brief overview of the semantic network components is presented in order to provide a background to the topic. The purpose of this paper is to review the proposed, implemented and/or simulated architectures for semantic network processing, and to discuss the capabilities and limitations of such architectures.     

Understanding semantic relationships

To develop sophisticated database management systems, there is a need to incorporate more understanding of the real world in the information that is stored in a database. Semantic data models have been developed to try to capture some of the meaning, as well as the structure, of data using abstractions such as inclusion, aggregation, and association. Besides these well-known relationships, a number of additional semantic relationships have been identified by researchers in other disciplines such as linguistics, logic, and cognitive psychology. This article explores some of the lesser-recognized semantic relationships and discusses both how they could be captured, either manually or by using an automated tool, and their impact on database design. To demonstrate the feasibility of this research, a prototype system for analyzing semantic relationships, called the Semantic Relationship Analyzer, is presented.     

The use of semantic networks to support concurrent engineering in semiconductor product development

The design of semiconductor devices is an extremely complex and costly process. Numerous design and test iterations are typically necessary to finally complete a successful device. Competition in the industry has forced semiconductor manufacturers to reduce design cycle times and costs. One method now being used to accomplish these objectives is concurrent engineering. This paper will review how concurrent engineering is being integrated into semiconductor device development and how artificial intelligence-based models will support concurrent engineering implementation. Major changes are needed in design simulation, methods of knowledge sharing, and incorporating best practices. A semantic network is proposed that retains the knowledge of a product in a central repository as various engineers contribute to the product's development. The knowledge contained in this central repository can be referenced for applicability by engineers during product design, development, and production.     

基于内容的视频分层语义联想模型

提出一种视频的分层语义联想模型,构造三个层次的信息：概念层次树,场景网络和语义对象网络。利用概念层次树来适应不同的应用环境,场景网络表示视频的时间信息,而语义对象及其关系用来表示视频镜头的内容,通过分属不同镜头的语义对象的关系来表示镜头间的语义相关度。该模型采用基于时间和语义关系的检索方法,搜索结果是收敛的。     

A novel solution for maze traversal problems using artificial neural networks

In this paper we have addressed the problem of finding a path through a maze of a given size. The traditional ways of finding a path through a maze employ recursive algorithms in which unwanted or non-paths are eliminated in a recursive manner. Neural networks with their parallel and distributed nature of processing seem to provide a natural solution to this problem. We present a biologically inspired solution using a two level hierarchical neural network for the mapping of the maze as also the generation of the path if it exists. For a maze of size S the amount of time it takes would be a function of S (O(S)) and a shortest path (if more than one path exists) could be found in around S cycles where each cycle involves all the neurons doing their processing in a parallel manner. The solution presented in this paper finds all valid paths and a simple technique for finding the shortest path amongst them is also given. The results are very encouraging and more applications of the network setup used in this report are currently being investigated. These include synthetic modeling of biological neural mechanisms, traversal of decision trees, modeling of associative neural networks (as in relating visual and auditory stimuli of a given phenomenon) and surgical micro-robot trajectory planning and execution.     

Negotiating the semantic gap: from feature maps to semantic landscapes

In this paper, we present the results of a project that seeks to transform low-level features to a higher level of meaning. This project concerns a technique, latent semantic indexing (LSI), in conjunction with normalization and term weighting, which have been used for full-text retrieval for many years. In this environment, LSI determines clusters of co-occurring keywords, sometimes, called concepts, so that a query which uses a particular keyword can then retrieve documents perhaps not containing this keyword, but containing other keywords from the same cluster. In this paper, we examine the use of this technique for content-based image retrieval, using two different approaches to image feature representation. We also study the integration of visual features and textual keywords and the results show that it can help improve the retrieval performance significantly.     

Generating Segmented Quality Meshes from Images

Techniques devoted to generating triangular meshes from intensity images either take as input a segmented image or generate a mesh without distinguishing individual structures contained in the image. These facts may cause difficulties in using such techniques in some applications, such as numerical simulations. In this work we reformulate a previously developed technique for mesh generation from intensity images called Imesh. This reformulation makes Imesh more versatile due to an unified framework that allows an easy change of refinement metric, rendering it effective for constructing meshes for applications with varied requirements, such as numerical simulation and image modeling. Furthermore, a deeper study about the point insertion problem and the development of geometrical criterion for segmentation is also reported in this paper. Meshes with theoretical guarantee of quality can also be obtained for each individual image structure as a post-processing step, a characteristic not usually found in other methods. The tests demonstrate the flexibility and the effectiveness of the approach.

Advanced learning resource management system with ontology-based hierarchy semantic model

The requirement for advanced learning resource management system is a hot topic in e-learning, which should support sharing, reusability and interoperability of learning resource. The development of semantic technology has provided opportunities to solve the problems. In e-learning area, there have been many research efforts in this direction. However, such kinds of existing systems have many weaknesses on resources usability and system availability because of various requirements on learning resources sharing, and the complexity of semantic technology itself. This paper proposes a feasible architecture for constructing learning resource sharing system with high availability through adopting semantic technology, and ontology-based hierarchy semantic model and method to construct course ontology are proposed. The approach has been validated with prototype development.     

基于语义网技术的主题词自动标引

为了实现电子政务公文主题词语义意义上的自动标引，创建了电子政务领域的概念空间，包括使用主题词表本体形式描述基本的概念及其关系和使用语义网资源描述框架表示概念类别词及其关系，然后根据主题词表抽取出题名所包含的主题词，并消除歧义词，最后将所得到的主题词使用逻辑推理机RACER进行语义逻辑的推理，从语义的角度进行主题词标引，突破机械标引的局限。