Hierarchical multilabel classification based on path evaluation

Multi-label classification assigns more than one label for each instance; when the labels are ordered in a predefined structure, the task is called Hierarchical Multi-label Classification (HMC). In HMC there are global and local approaches. Global approaches treat the problem as a whole but tend to explode with large datasets. Local approaches divide the problem into local subproblems, but usually do not exploit the information of the hierarchy. This paper addresses the problem of HMC for both tree and Direct Acyclic Graph (DAG) structures whose labels do not necessarily reach a leaf node. A local classifier per parent node is trained incorporating the prediction of the parent(s) node(s) as an additional attribute to include the relations between classes. In the classification phase, the branches with low probability to occur are pruned, performing non-mandatory leaf node prediction. Our method evaluates each possible path from the root of the hierarchy, taking into account the prediction value and the level of the nodes; selecting the path (or paths in the case of DAGs) with the highest score. We tested our method with 20 datasets with tree and DAG structured hierarchies against a number of state-of-the-art methods. Our method proved to obtain superior results when dealing with deep and populated hierarchies.     

CONVERGENCE OF BACKPROPAGATION WITH MOMENTUM FOR NETWORK ARCHITECTURES WITH SKIP CONNECTIONS

We study a class of deep neural networks with architectures that form a directed acyclic graph(DAG).For backpropagation defined by gradient descent with adaptive momentum,we show weights converge for a large class of nonlinear activation functions.'The proof generalizes the results of Wu et al.(2008)who showed convergence for a feed-forward network with one hidden layer.For an example of the effectiveness of DAG architectures,we describe an example of compression through an AutoEncoder,and compare against sequential feed-forward networks under several metrics.     

提高软件Agent策略生成的语义互操作能力

软件Agent遇到语义二义性时无法正确地为用户解决矛盾问题.利用本体支持语义互操作的特点,在软件Agent策略生成机制中引入复合元本体,加入Agent智能引导的人机交互方法,实现了具有语义歧义消除能力的可拓策略生成系统.当用户输入的问题信息语义模糊时,Agent根据本体中的知识与用户交互,逐步理解用户的信息语义,直到能为用户生成解决矛盾问题的策略.以旅游与购物问题为例进行了实验,结果表明了软件Agent策略生成的语义互操作能力得到了提高.     

A discrete dynamics approach to sparse calculation and applied in ontology science

ABSTRACT

In the era of big data, with the increase of data processing information and the increase of data complexity, higher requirements are put on the tools and algorithms of data processing. As a tool for structured information representation, ontology has been used in engineering fields such as chemistry, biology, pharmacy, and materials. As a dynamic structure, the increasing concepts contributes to a gradual increase of a single ontology. In order to solve the problem of computational complexity decreasing in the procedure of similarity calculating, the techniques of dimensionality reduction and sparse computing are applied to ontology learning. This article presents discrete dynamics approach showing several tricks on applying the sparse computing method to ontology learning, and verify its efficiency through experiments.     

Unifying ontological similarity measures: A theoretical and empirical investigation

This paper theoretically and empirically investigates ontological similarity. Tversky’s parameterized ratio model of similarity [3] is shown as a unifying basis for many of the well-known ontological similarity measures. A new family of ontological similarity measures is proposed that allows parameterizing the characteristic set used to represent an ontological concept. The three subontologies of the prominent Gene Ontology (GO) are used in an empirical investigation of several ontological similarity measures. Another study using well known semantic similarity within two different anatomy ontologies, the NCIT anatomy and the mouse anatomy, is also presented for comparison to several of the GO results. A discussion of the correlation among the measures is presented as well as a comparison of the effects of two different methods of determining a concept’s information content, corpus-based and ontology-based.     

Usage scenarios of a semantic-based digital shape workbench

Motivated by the undisputed emergence of 3D content, this paper presents cases where semantic technologies have the potential to be used to build robust real-life solutions out of the AIM@SHAPE Digital Shape Workbench both in its current and its future forms. We focus on understanding what makes semantic knowledge applications successful in operational environments which involve shapes and present existing and future scenarios on favoring ontology based integration that systematically resolves both syntactic and semantic heterogeneity, allowing scientists to concentrate more on their content. (© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)     

Carnap,Goguen, and the Hyperontologies: Logical Pluralism and Heterogeneous Structuring in Ontology Design

This paper addresses questions of universality related to ontological engineering, namely aims at substantiating (negative) answers to the following three basic questions: (i) Is there a ‘universal ontology’?, (ii) Is there a ‘universal formal ontology language’?, and (iii) Is there a universally applicable ‘mode of reasoning’ for formal ontologies? To support our answers in a principled way, we present a general framework for the design of formal ontologies resting on two main principles: firstly, we endorse Rudolf Carnap’s principle of logical tolerance by giving central stage to the concept of logical heterogeneity, i.e. the use of a plurality of logical languages within one ontology design. Secondly, to structure and combine heterogeneous ontologies in a semantically well-founded way, we base our work on abstract model theory in the form of institutional semantics, as forcefully put forward by Joseph Goguen and Rod Burstall. In particular, we employ the structuring mechanisms of the heterogeneous algebraic specification language HetCasl for defining a general concept of heterogeneous, distributed, highly modular and structured ontologies, called hyperontologies. Moreover, we distinguish, on a structural and semantic level, several different kinds of combining and aligning heterogeneous ontologies, namely integration, connection, and refinement. We show how the notion of heterogeneous refinement can be used to provide both a general notion of sub-ontology as well as a notion of heterogeneous equivalence of ontologies, and finally sketch how different modes of reasoning over ontologies are related to these different structuring aspects.     

Agent协助建立矛盾问题的可拓模型研究

针对普通计算机系统缺乏界定矛盾问题(尤其是核问题)的能力的缺点,探讨了如何利用Agent技术协助建立矛盾问题的可拓模型的问题.提出一种方案,利用Agent的通信能力与反应性进行人机交互,引导用户输入问题信息,减轻让计算机识别自然语言信息的困难;利用Agent的自主性和学习能力,将学习到的知识保存到协助进行语义分析的本体中,使本体不断进化,提高系统界定各种各样矛盾问题的能力.创新点在于结合了可拓学、Agent、本体等几种不同学科领域技术的优势,使计算机系统具有一定的语义分析能力而协助人们界定矛盾问题.     

Identification of graphical models for nonignorable nonresponse of binary outcomes in longitudinal studies

In this paper, we use directed acyclic graphs (DAGs) with temporal structure to describe models of nonignorable nonresponse mechanisms for binary outcomes in longitudinal studies, and we discuss identification of these models under an assumption that the sequence of variables has the first-order Markov dependence, that is, the future variables are independent of the past variables conditional on the present variables. We give a stepwise approach for checking identifiability of DAG models. For an unidentifiable model, we propose adding completely observed variables such that this model becomes identifiable.     

Acyclic directed graphs representing independence models

In this paper we study the problem of representing probabilistic independence models, in particular those closed under graphoid properties. We focus on acyclic directed graph (DAG): a new algorithm to build a DAG, given an ordering among random variables, is described and peculiarities and advantages of this approach are discussed. Moreover, we provide a necessary and sufficient condition for the existence of a perfect map representing an independence model and we describe an algorithm based on this characterization.     

A methodology for knowledge discovery to support product family design

This paper introduces a methodology for knowledge discovery related to product family design that integrates an ontology with data mining techniques. In the proposed methodology, the ontology represents attributes for the components of products in functional hierarchies. Fuzzy clustering is employed for data mining to first partition product functions into subsets for identifying modules in a given product family and then identify the similarity level of components in a module. Module categorization is introduced to support association rule mining for knowledge discovery related to platform design. We apply the proposed methodology to first develop and then utilize design knowledge for a family of power tools. Based on the developed design knowledge, a new platform is suggested to improve commonality in the power tool family.     

Epistemological Skepticism, Semantic Blindness, and Competence-Based Performance Errors

The semantic blindness objection to contextualism challenges the view that there is no incompatibility between (i) denials of external-world knowledge in contexts where radical-deception scenarios are salient, and (ii) affirmations of external-world knowledge in contexts where such scenarios are not salient. Contextualism allegedly attributes a gross and implausible form of semantic incompetence in the use of the concept of knowledge to people who are otherwise quite competent in its use; this blindness supposedly consists in wrongly judging that there is genuine conflict between claims of type (i) and type (ii). We distinguish two broad versions of contextualism: relativistic-content contextualism and categorical-content contextualism. We argue that although the semantic blindness objection evidently is applicable to the former, it does not apply to the latter. We describe a subtle form of conflict between claims of types (i) and (ii), which we call différance-based affirmatory conflict. We argue that people confronted with radical-deception scenarios are prone to experience a form of semantic myopia (as we call it): a failure to distinguish between différance-based affirmatory conflict and outright inconsistency. Attributing such semantic myopia to people who are otherwise competent with the concept of knowledge explains the bafflement about knowledge-claims that so often arises when radical-deception scenarios are made salient. Such myopia is not some crude form of semantic blindness at all; rather, it is an understandable mistake grounded in semantic competence itself: what we call a competence-based performance error.     

Parallel Information Algorithm with Local Tuning for Solving Multidimensional GO Problems

In this paper we propose a new parallel algorithm for solving global optimization (GO) multidimensional problems. The method unifies two powerful approaches for accelerating the search: parallel computations and local tuning on the behavior of the objective function. We establish convergence conditions for the algorithm and theoretically show that the usage of local information during the global search permits to accelerate solving the problem significantly. Results of numerical experiments executed with 100 test functions are also reported.     

Granular computing applied to ontologies

Granular Computing is an emerging conceptual and computing paradigm of information-processing. A central notion is an information-processing pyramid with different levels of clarifications. Each level is usually represented by ‘chunks’ of data or granules, also known as information granules. Rough Set Theory is one of the most widely used methodologies for handling or defining granules.Ontologies are used to represent the knowledge of a domain for specific applications. A challenge is to define semantic knowledge at different levels of human-depending detail.In this paper we propose four operations in order to have several granular perspectives for a specific ontological commitment. Then these operations are used to have various views of an ontology built with a rough-set approach. In particular, a rough methodology is introduced to construct a specific granular view of an ontology.     

Upward straight-line embeddings of directed graphs into point sets

In this paper we study the problem of computing an upward straight-line embedding of a planar DAG (directed acyclic graph) G into a point set S, i.e. a planar drawing of G such that each vertex is mapped to a point of S, each edge is drawn as a straight-line segment, and all the edges are oriented according to a common direction. In particular, we show that no biconnected DAG admits an upward straight-line embedding into every point set in convex position. We provide a characterization of the family of DAGs that admit an upward straight-line embedding into every convex point set such that the points with the largest and the smallest y-coordinate are consecutive in the convex hull of the point set. We characterize the family of DAGs that contain a Hamiltonian directed path and that admit an upward straight-line embedding into every point set in general position. We also prove that a DAG whose underlying graph is a tree does not always have an upward straight-line embedding into a point set in convex position and we describe how to construct such an embedding for a DAG whose underlying graph is a path. Finally, we give results about the embeddability of some sub-classes of DAGs whose underlying graphs are trees on point set in convex and in general position.     

基于Fuzzy-Grey关联分析的信息资源聚焦的数学模型

本文运用饱和信息维的增减技术对信息获取子空间进行结构上的剖分,用饱和信息维列作为因子列来构造灰色关联空间,将灰色理论中的关联分析用于模糊综合评判模型,用灰色关联系数来替代隶属度,用关联序来描述信息之间的相关联程度,得到信息资源聚焦的一种新的数学模型。     

Improving complexity of structured convex optimization problems using self-concordant barriers

The purpose of this paper is to provide improved complexity results for several classes of structured convex optimization problems using the theory of self-concordant functions developed by Nesterov and Nemirovski in SIAM Studies in Applied Mathematics, SIAM Publications, Philadelphia, 1994. We describe the classical short-step interior-point method and optimize its parameters in order to provide the best possible iteration bound. We also discuss the necessity of introducing two parameters in the definition of self-concordancy and which one is the best to fix. A lemma due to den Hertog et al. in Mathematical Programming Series B 69 (1) (1995) is improved, which allows us to review several classes of structured convex optimization problems and improve the corresponding complexity results.