概率图模型研究进展综述

概率图模型作为一类有力的工具,能够简洁地表示复杂的概率分布,有效地(近似)计算边缘分布和条件分布,方便地学习概率模型中的参数和超参数.因此,它作为一种处理不确定性的形式化方法,被广泛应用于需要进行自动的概率推理的场合,例如计算机视觉、自然语言处理.回顾了有关概率图模型的表示、推理和学习的基本概念和主要结果,并详细介绍了这些方法在两种重要的概率模型中的应用.还回顾了在加速经典近似推理算法方面的新进展.最后讨论了相关方向的研究前景.     

Qualitative reasoning with imprecise probabilities

This paper investigates the possibility of performing automated reasoning in probabilistic knowledge bases when probabilities are expressed by means of linguistic quantifiers. Data are expressed in terms of ill-known conditional probabilities represented by linguistic terms. Each linguistic term is expressed as a prescribed interval of proportions. Then instead of propagating numbers, qualitative terms are propagated in accordance with the numerical interpretation of these terms. The quantified syllogism, modeling the chaining of probabilistic rules, is studied in this context. It is shown that a qualitative counterpart of this syllogism makes sense and is fairly independent of the thresholds defining the linguistically meaningful intervals, provided that these threshold values remain in accordance with the intuition. The inference power is less than a full-fledged probabilistic constraint propagation device but corresponds better to what could be thought of as commonsense probabilistic reasoning. Suggestions that may improve the inferencing power in the qualitative setting are proposed.This paper is an extended and revised version of a paper entitled A Symbolic Approach to Reasoning with Linguistic Quantifiers inProc. 8th Conf. Uncertainty in Artificial Intelligence Dubois, D., Wellman, M.P., D'Ambrosio, B., and Smets, Ph. (Eds.), Morgan Kaufmann, pp. 74–82, 1992.     

Context-specific sign-propagation in qualitative probabilistic networks

Qualitative probabilistic networks are qualitative abstractions of probabilistic networks, summarising probabilistic influences by qualitative signs. As qualitative networks model influences at the level of variables, knowledge about probabilistic influences that hold only for specific values cannot be expressed. The results computed from a qualitative network, as a consequence, can be weaker than strictly necessary and may in fact be rather uninformative. We extend the basic formalism of qualitative probabilistic networks by providing for the inclusion of context-specific information about influences and show that exploiting this information upon reasoning has the ability to forestall unnecessarily weak results.     

概率逻辑含多重原子交集分解模型的可靠性

文中给出含多重原子公式交集的概率逻辑分解计算模型的可靠性证明。若概率逻辑公式集可被划分为含多重原子交集的子集，且交集的原子个数较少，则分解算法将明显地缩小概率逻辑自动推理线性规划模型的规模。     

一种基于因子图模型的半监督社区发现方法

社区发现是社交网络分析中一个重要的研究方向.当前大部分的研究都聚焦在自动社区发现问题,但是在具有数据缺失或噪声的网络中,自动社区发现算法的性能会随着噪声数据的增加而迅速下降.通过在社区发现中融合先验信息,进行半监督的社区发现,有望为解决上述挑战提供一条可行的途径.本文基于因子图模型,通过融入先验信息到一个统一的概率框架中,提出了一种基于因子图模型的半监督社区发现方法,研究具有用户引导情况下的社交网络社区发现问题.在三个真实的社交网络数据（Zachary社会关系网、海豚社会网和DBLP协作网）上进行实验,证明通过融入先验信息可以有效地提高社区发现的精度,且将我们的方法与一种最新的半监督社区发现方法（半监督Spin-Glass模型）进行对比,在三个数据集中F-measure平均提升了6.34%、16.36%和12.13%.     

Annotated Logic Applications for Imperfect Information

Imperfect information is a very general term that comprises different types of information, such as uncertain, vague, fuzzy, inconsistent, possibilistic, probabilistic, partially or totally incomplete information [2]. In the literature of knowledge representation we find a different formal model for each one of these distinct types. For example, annotated logic is a formal model to represent inconsistent information.Annotated logics are non-classical logics introduced in [20] as a logic programming theory. They were proved to be paraconsistent. Based on [5], we present in this work the annotated logic programming theory and some of its applications in Artificial Intelligence (AI). We present it as a formalism to reason with inconsistent information and investigate its possibility to represent other types of imperfect information, such as possibilistic and non-monotonic reasoning. Our main goal is to verify and confirm the importance of annotated logics as a tool for developing knowledge-based and automated reasoning systems in AI.     

The problem of finding a strategy to control binary paramodulation

This article is the fifth of a series of articles discussing various open research problems in automated reasoning. Here we focus on the equality relation which is so vital to many applications of automated reasoning. The prolem proposed for research asks one to find a strategy for controlling the application of binary paramodulation. For evaluating a proposed solution to this research problem, we include suggestions concerning possible test problems.This work was supported by the Applied Mathematical Sciences subprogram of the Office of Energy Research, U.S. Department of Energy, under contract W-31-109-Eng-38.     

Probabilistic Reasoning Under Coherence in System P

In this paper we apply a probabilistic reasoning under coherence to System P. We consider a notion of strict probabilistic consistency, we show its equivalence to Adams' probabilistic consistency, and we give a necessary and sufficient condition for probabilistic entailment. We consider the inference rules of System P in the framework of coherent imprecise probabilistic assessments. Exploiting our coherence-based approach, we propagate the lower and upper probability bounds associated with the conditional assertions of a given knowledge base, obtaining the precise probability bounds for the derived conclusions of the inference rules. This allows a more flexible and realistic use of System P in default reasoning and provides an exact illustration of the degradation of the inference rules when interpreted in probabilistic terms. We also examine the disjunctive Weak Rational Monotony rule of System P⁺ proposed by Adams in his extended probabilistic logic. Finally, we examine the propagation of lower bounds with real -values and, to illustrate our probabilistic reasoning, we consider an example.     

A Bayesian Method for the Induction of Probabilistic Networks from Data

This paper presents a Bayesian method for constructing probabilistic networks from databases. In particular, we focus on constructing Bayesian belief networks. Potential applications include computer-assisted hypothesis testing, automated scientific discovery, and automated construction of probabilistic expert systems. We extend the basic method to handle missing data and hidden (latent) variables. We show how to perform probabilistic inference by averaging over the inferences of multiple belief networks. Results are presented of a preliminary evaluation of an algorithm for constructing a belief network from a database of cases. Finally, we relate the methods in this paper to previous work, and we discuss open problems.     

A virtual prototyping approach to product disassembly reasoning

An important aspect of Design for the Life Cycle is assessing the disassemblability of products. After an artifact has completed its useful life, it must be disassembled then recycled, remanufactured or scrapped. Disassemblability of a product can be evaluated by performing disassembly activities on prototypes. Virtual prototyping (VP) is an alternative to hardware prototyping in which analysis of designs can be done without manufacturing physical samples. In recent years, disassembly processes have been generated either by using interactive or automated approaches, but these approaches have limitations. Interactive approaches require extensive user input usually in the form of answering questions, whereas automated approaches can only be used to generate disassembly processes for products with simple component configuration and geometry. In this paper automated and interactive techniques are combined, using VP, to generate complete disassembly processes of a product design. To support generation of disassembly processes of a product, a virtual environment and VP method were developed that will support disassembly activities performed by a designer. The product model of the virtual prototype is generated from the CAD model. The disassembly process model for the prototype is generated using automated reasoning techniques and is completed by interactively disassembling the product in the virtual environment. Extensions to automatic reasoning techniques to compute ranges of feasible directions of component removal were developed to facilitate the generation of the disassembly process. A scheme to represent the disassembly process for disassemblability evaluation was developed and implemented. In this paper a Chrysler LHS center console has been used to illustrate our approach of generating disassembly processes via VP.     

Conception of complex probabilistic neural network system for classification of partial discharge patterns using multifarious inputs

Pattern recognition has a long history within electrical engineering but has recently become much more widespread as the automated capture of signal and images has been cheaper. Very many of the application of neural networks are to classification, and so are within the field of pattern recognition and classification. In this paper, we explore how probabilistic neural networks fit into the earlier framework of pattern recognition of partial discharge patterns since the PD patterns are an important tool for diagnosis of HV insulation systems. Skilled humans can identify the possible insulation defects in various representations of partial discharge (PD) data. One of the most widely used representation is phase resolved PD (PRPD) patterns. Also this paper describes a method for the automated recognition of PRPD patterns using a novel complex probabilistic neural network system for the actual classification task. The efficacy of composite neural network developed using probabilistic neural network is examined.     

The problem of choosing between logic programming and general-purpose automated reasoning

This article is the thirteenth of a series of articles discussing various open research problems in automated reasoning. Here we focus on finding criteria for correctly choosing between using logic programming and a more general automated reasoning approach to attack a given assignment. The problem proposed for research asks one to find criteria that classify problems as solvable with a well-focused algorithm or as requiring a more general search for new information. We include suggestions for evaluating a proposed solution to this research problem.This work was supported by the Applied Mathematical Sciences subprogram of the Office of Energy Research, U.S. Department of Energy, under Contract W-31-109-Eng-38.     

Probabilistic 3D Object Recognition

A probabilistic 3D object recognition algorithm is presented. In order to guide the recognition process the probability that match hypotheses between image features and model features are correct is computed. A model is developed which uses the probabilistic peaking effect of measured angles and ratios of lengths by tracing iso-angle and iso-ratio curves on the viewing sphere. The model also accounts for various types of uncertainty in the input such as incomplete and inexact edge detection. For each match hypothesis the pose of the object and the pose uncertainty which is due to the uncertainty in vertex position are recovered. This is used to find sets of hypotheses which reinforce each other by matching features of the same object with compatible uncertainty regions. A probabilistic expression is used to rank these hypothesis sets. The hypothesis sets with the highest rank are output. The algorithm has been fully implemented, and tested on real images.     

A new approach to hybrid probabilistic logic programs

This paper presents a novel revision of the framework of Hybrid Probabilistic Logic Programming, along with a complete semantics characterization, to enable the encoding of and reasoning about real-world applications. The language of Hybrid Probabilistic Logic Programs framework is extended to allow the use of non-monotonic negation, and two alternative semantical characterizations are defined: stable probabilistic model semantics and probabilistic well-founded semantics. These semantics generalize the stable model semantics and well-founded semantics of traditional normal logic programs, and they reduce to the semantics of Hybrid Probabilistic Logic programs for programs without negation. It is the first time that two different semantics for Hybrid Probabilistic Programs with non-monotonic negation as well as their relationships are described. This proposal provides the foundational grounds for developing computational methods for implementing the proposed semantics. Furthermore, it makes it clearer how to characterize non-monotonic negation in probabilistic logic programming frameworks for commonsense reasoning. An erratum to this article can be found at     

The problem of definition expansion and contraction

This article is the fourth of a series of articles discussing various open research problems in automated reasoning. The problem proposed for research focuses on finding criteria that would enable an automated reasoning program to expand or contract definitions wisely. For evaluating a proposed solution to this research problem, we include suggestions concerning possible test problems.This work was supported by the Applied Mathematical Sciences subprogram of the Office of Energy Research, U.S. Department of Energy, under contract W-31-109-Eng-38.     

自然语言中时间信息的模型化

在自然语言理解中，时间一个重要的语境因素，本文提出一种独立于句子表层形式的多层次时间语义结构，浅层语义结构是时间描述的最化语义表示，深层语义结构描述事件的动态属性和存在特征，该时间语义模型表示时刻和时段，将时间基点明确区分为物理时间基点和说话者时间基点，并提供通用的时间语义计算方法，将各语言的时间描述映射到时间轴上，在该模型基础上，从语义观点出发，建立时，体的可计算模型及各事件时间相关性计算模型。     

A multi-viewpoint system to support abductive reasoning

The paper proposes a multi-viewpoint system to support human abductive reasoning for diagnosis, prognosis and trial-and-error activities for supervising automated systems. This multi-viewpoint approach interprets the same set of events from the different viewpoints in this set. The algorithms for managing these viewpoints and the set of events are related to hypothetical reasoning, and they use several main functions to (1) select or reject certain events, (2) cancel or recover these events, and (3) manage the consistency of the viewpoints. This approach is applied to diagnosis and trial-and-error activities related to the phone troubleshooting problem.     

Probabilistic reasoning in an augmented fact-proposition space and its applications

A simple and versatile probabilistic reasoning scheme is presented. Based on an augmentation of a multi-dimensional inference space indexed by a Cartesian product of the fact and proposition sets, the scheme simplifies the processes involved in the representation and computation of a probabilistic reasoning system. In the augmented space, a set of auxiliary fields is utilized in addition to the fact-proposition relations to manipulate the uncertainty and incompleteness of the information presented. The scheme enhances the functionality of a probabilistic reasoning and facilitates the building of practical reasoning systems. The utilization of the augmented space in reasoning is illustrated by two problems in computer-vision applications.     

The problem of self-analytically choosing the weights

This article is the eighth of a series of articles discussing various open research problems in automated reasoning. Here we focus on self-analysis and the weighting strategy. The problem proposed for research asks one to find a means for an automated reasoning program to self-analytically determine which weights or priorities to give to clauses at various points during the program's attempt to complete a given assignment. For evaluating a proposed solution to this research problem, we include suggestions concerning possible test problems.This work was supported by the Applied Mathematical Sciences subprogram of the Office of Energy Research, U.S. Department of Energy, under contract W-31-109-Eng-38.     

The problem of self-analytically choosing the set of support

This article is the seventh of a series of articles discussing various open research problems in automated reasoning. Here we focus on self-analysis and the set of support strategy. The problem proposed for research asks for one to find a means for an automated reasoning program to self-analytically determine which clauses to be in the set of support at various points during the program's attempt to complete a given assignment. For evaluating a proposed solution to this research problem, we include suggestions concerning possible test problems.This work was supported by the Applied Mathematical Sciences subprogram of the Office of Energy Research, U.S. Department of Energy, under contract W-31-109-Eng-38.