A DISTRIBUTED NEURAL APPROACH FOR CAUSAL REASONING USING COOPERATIVE AND COMPETITIVE NEURAL COMPUTATIONS

ABSTRACT

In this work, we develop a neural model to solve causal reasoning problems (also called abduction) in the open, independent, and incompatibility classes. We model the reasoning process by a single and global energy function using cooperative and competitive neural computation. The update rules of the distinct connections of the network are derived from its energy function, using gradient descent techniques. Simulation results reveal a good performance of the model.     

Semantics in analogical reasoning

Abstract

To have broad application or approach the capacity of ordinary human thinking, analogical reasoning programs must become more complex or their semantics must become richer, or both. Little research is being done to discover what a broad collection of semantics can contribute to general-purpose analogical reasoning. From an analysis of a small collection of words, two classes containing over 300 categories are determined representing the semantics for understanding metaphors. A broad-based collection of simple target is source metaphors is sampled with each target and source represented in terms of these categories without known influence from the respective metaphor. A computational model is developed drawing on several disciplines while using rules for the recognition and elaboration phases of reasoning. Recognition primarily involves finding and reorganizing relevant schemata within the source. Elaboration involves reorganizing and possibly adding to the target. Given each metaphor, the TisS computer program creates a new representation of each target and source including several meanings for each metaphor. A test with human subjects making aptness and agreement judgements on the generated statements is discussed, suggesting a methodology for using a rich semantic base in analogical reasoning.     

行为理解的认知推理方法

目的 人类行为理解是机器智能研究中最富有挑战性的领域。其根本问题是语义获取,即从动作推理得到人的行为,需要跨越两者之间的语义鸿沟,为此提出一种人关于日常行为知识与人体动作行为、环境信息之间的建模方法,以及可扩展的开放式结构环境—行为关系模型,基于该模型提出一种新的行为理解的渐进式认知推理方法。方法 首先根据知识,建立多种特征、复合特征和行为之间的关系模型。系统根据当前的输入流,处理得到当前的特征与复合特征集,推理得到当前的可能行为集。该行为集指导处理模块,更新特征集,得到新的行为集。结果 应用本文渐进式连续推理方法,系统可以把人关于日常行为的知识与人体运动、环境变化等传感器数据处理获取到的信息动态绑定,实现知识辅助的行为理解。结论 提出的推理方法能连续处理长时间、同时发生的行为。     

提取全局语义信息的场景图生成算法

目的 场景图能够简洁且结构化地描述图像。现有场景图生成方法重点关注图像的视觉特征,忽视了数据集中丰富的语义信息。同时,受到数据集长尾分布的影响,大多数方法不能很好地对出现概率较小的三元组进行推理,而是趋于得到高频三元组。另外,现有大多数方法都采用相同的网络结构来推理目标和关系类别,不具有针对性。为了解决上述问题,本文提出一种提取全局语义信息的场景图生成算法。方法 网络由语义编码、特征编码、目标推断以及关系推理等4个模块组成。语义编码模块从图像区域描述中提取语义信息并计算全局统计知识,融合得到鲁棒的全局语义信息来辅助不常见三元组的推理。目标编码模块提取图像的视觉特征。目标推断和关系推理模块采用不同的特征融合方法,分别利用门控图神经网络和门控循环单元进行特征学习。在此基础上,在全局统计知识的辅助下进行目标类别和关系类别推理。最后利用解析器构造场景图,进而结构化地描述图像。结果 在公开的视觉基因组数据集上与其他10种方法进行比较,分别实现关系分类、场景图元素分类和场景图生成这3个任务,在限制和不限制每对目标只有一种关系的条件下,平均召回率分别达到了44.2%和55.3%。在可视化实验中,相比...     

ASEMIQUAUTATIVE APPROACH TO REASONING IN PROBABILISTIC NETWORKS

Abstract

This paper proposes the use of semiqualitative modeling for reasoning in probabilistic networks. Semiqualitative modeling is a generalization of qualitative modeling that refines the set of intervals in which values may be expressed. The advantage of semiqualitative modeling of probabilistic reasoning over more traditional methods is that a semiqualitative model can cope with incomplete and imprecise information that would prevent a more traditional model from functioning. The semi-qualitative analysis of a well-known example from the literature is presented, and conclusions about the general use of semiqualitative modeling in reasoning under uncertainty is discussed.     

Realistic model for temporal reasoning in real-time patient monitoring

Time is a central factor in patient monitoring. Introduction of domain-dependent knowledge is essential to ensure efficiency of time managers, especially when embedded into systems that interact with the real world. We present a realistic temporal reasoning model based on two basic cognitive mechanisms: aggregation of similar observed situations and forgetting of non-relevant information. We describe in detail how we represented the proposed model and how, by refinement of domain-independent temporal entities and inferences, we added domain specific knowledge to manage a clinical therapy. The model allows clinical observations to be incrementally interpreted as they are acquired by an intelligent system, mainly reactive in its reasoning, for the management of patients receiving respiratory support.     

OSCAR: a general theory of rationality

Abstract

The enterprise is the construction of a general theory of rationality, and its implementation in an automated reasoning system named OSCAR. The paper describes a general architecture for rational thought. This includes both theoretical reasoning and practical reasoning, and builds in important interconnections between them. It is urged that a sophisticated reasoner must be an introspective reasoner, capable of monitoring its own reasoning and reasoning about it. An introspective reasoner is built on top of a non-introspective reasoner that represents the system's default reasoning strategies. The introspective reasoner engages in practical reasoning about reasoning in order to override these default strategies. The paper concludes with a discussion of some aspects of the default reasoner, including the manner in which reasoning is interest driven, and the structure of defeasible reasoning.     

Constructing models for predicting extract subclass refactoring opportunities using object-oriented quality metrics

ContextRefactoring is a maintenance task that refers to the process of restructuring software source code to enhance its quality without affecting its external behavior. Inspecting and analyzing the source code of the system under consideration to identify the classes in need of extract subclass refactoring (ESR) is a time consuming and costly process.ObjectiveThis paper explores the abilities of several quality metrics considered individually and in combination to predict the classes in need of ESR.MethodFor a given a class, this paper empirically investigates, using univariate logistic regression analysis, the abilities of 25 existing size, cohesion, and coupling metrics to predict whether the class is in need of restructuring by extracting a subclass from it. In addition, models of combined metrics based on multivariate logistic regression analysis were constructed and validated to predict the classes in need of ESR, and the best model is justifiably recommended. We explored the statistical relations between the values of the selected metrics and the decisions of the developers of six open source Java systems with respect to whether the classes require ESR.ResultsThe results indicate that there was a strong statistical relation between some of the quality metrics and the decision of whether ESR activity was required. From a statistical point of view, the recommended model of metrics has practical thresholds that lead to an outstanding classification of the classes into those that require ESR and those that do not.ConclusionThe proposed model can be applied to automatically predict the classes in need of ESR and present them as suggestions to developers working to enhance the system during the maintenance phase. In addition, the model is capable of ranking the classes of the system under consideration according to their degree of need of ESR.     

Automated reasoning on UML conceptual schemas with derived information and queries

ContextIt is critical to ensure the quality of a software system in the initial stages of development, and several approaches have been proposed to ensure that a conceptual schema correctly describes the user’s requirements.ObjectiveThe main goal of this paper is to perform automated reasoning on UML schemas containing arbitrary constraints, derived roles, derived attributes and queries, all of which must be specified by OCL expressions.MethodThe UML/OCL schema is encoded in a first order logic formalisation, and an existing reasoning procedure is used to check whether the schema satisfies a set of desirable properties. Due to the undecidability of reasoning in highly expressive schemas, such as those considered here, we also provide a set of conditions that, if satisfied by the schema, ensure that all properties can be checked in a finite period of time.ResultsThis paper extends our previous work on reasoning on UML conceptual schemas with OCL constraints by considering derived attributes and roles that can participate in the definition of other constraints, queries and derivation rules. Queries formalised in OCL can also be validated to check their satisfiability and to detect possible equivalences between them. We also provide a set of conditions that ensure finite reasoning when they are satisfied by the schema under consideration.ConclusionThis approach improves upon previous work by allowing automated reasoning for more expressive UML/OCL conceptual schemas than those considered so far.     

NEW METHODOLOGY TO FUZZY REASONING FOR RULE-BASED EXPERT SYSTEMS

Abstract

Much knowledge residing in the knowledge base of an expert system involves fuzzy concepts. A powerful expert system must have the capability of fuzzy reasoning. This paper presents a new methodology for dealing with fuzzy reasoning based on the matching function S. The single-input, single-output (SISO) fuzzy reasoning scheme and the multi-input, single-output (MISO) fuzzy reasoning schemes are discussed in detail. The proposed fuzzy reasoning methodology is conceptually clearer than the compositional rule of inference approach. It can provide an useful way for rule-based systems to deal with fuzzy reasoning.     

Model-based remotely-sensed imagery interpretation

Abstract

We present a model-based remotely-sensed image interpretation expert system embeded in a knowledge-based geographic information system (K. BIS). The KBIS consists of four sub-systems: a pictorial data base system, an image interpretation expert system, a computer-aided planning system and a computer-aided cartographic system. The image interpretation expert system represents ecological knowledge and other expert knowledge by frames. Its reasoning process consists of a forward reasoning based on the Bayes classification of Landsat imagery, a backward reasoning using frame knowledge and reasoning using a spatial consistency model. A forest inventory study was conducted in Shaxian county, in the southern part of China, using this expert system. The results have shown a significant improvement. Building image interpretation expert systems within knowledge-based pictorial systems is very convenient and efficient because there are well-organized data, knowledge and procedures available.     

问题引导的空间关系图推理视觉问答模型

目的 现有视觉问答模型的研究主要从注意力机制和多模态融合角度出发,未能对图像场景中对象之间的语义联系显式建模,且较少突出对象的空间位置关系,导致空间关系推理能力欠佳。对此,本文针对需要空间关系推理的视觉问答问题,提出利用视觉对象之间空间关系属性结构化建模图像,构建问题引导的空间关系图推理视觉问答模型。方法 利用显著性注意力,用Faster R-CNN(region-based convolutional neural network)提取图像中显著的视觉对象和视觉特征;对图像中的视觉对象及其空间关系结构化建模为空间关系图;利用问题引导的聚焦式注意力进行基于问题的空间关系推理。聚焦式注意力分为节点注意力和边注意力,分别用于发现与问题相关的视觉对象和空间关系;利用节点注意力和边注意力权重构造门控图推理网络,通过门控图推理网络的信息传递机制和控制特征信息的聚合,获得节点的深度交互信息,学习得到具有空间感知的视觉特征表示,达到基于问题的空间关系推理;将具有空间关系感知的图像特征和问题特征进行多模态融合,预测出正确答案。结果 模型在VQA(visual question answering)v2...     

Organizing human factors knowledge for the evaluation and design of interfaces

This paper presents research conducted in the area of human factors knowledge collection and organization. Providing that a large amountof human factors engineering data is available now concerning user interface aspects, this research started from the following question: Can this knowledge be transferred to user interface designers, and by which means can this be done? An important part of human factors knowledge is included in recommendations and guidelines. It is the problems with that type of knowledge that this paper focuses on. A first attempt to tackle some of these problems is made by offering a generic model for deciphering recommendations into rules. This model stresses the importance of attributes such as human factors criteria, classes of premises, and conclusions, as well as interface objects. Definitions of criteria for organizing recommendations are offered, and a few examples of rule‐based reasoning are provided. The remaining problems with a rule approach are then discussed according to two dimensions: intrinsic problems and usage problems. In addition, it appears that an approach based on literature recommendations is not sufficient. Other areas that need further work are described, particularly concerning task and interface modelling, and human factors expertise modelling. The paper concludes on the limits and benefits of such an approach.     

A New Characterization of NP, P, and PSPACE with Accepting Hybrid Networks of Evolutionary Processors

We consider three complexity classes defined on Accepting Hybrid Networks of Evolutionary Processors (AHNEP) and compare them with the classical complexity classes defined on the standard computing model of Turing machine. By definition, AHNEPs are deterministic. We prove that the classical complexity class NP equals the family of languages decided by AHNEPs in polynomial time. A language is in P if and only if it is decided by an AHNEP in polynomial time and space. We also show that PSPACE equals the family of languages decided by AHNEPs in polynomial length.     

Version spaces and the consistency problem

A version space is a collection of concepts consistent with a given set of positive and negative examples. Mitchell [Artificial Intelligence 18 (1982) 203-226] proposed representing a version space by its boundary sets: the maximally general (G) and maximally specific consistent concepts (S). For many simple concept classes, the size of G and S is known to grow exponentially in the number of positive and negative examples. This paper argues that previous work on alternative representations of version spaces has disguised the real question underlying version space reasoning. We instead show that tractable reasoning with version spaces turns out to depend on the consistency problem, i.e., determining if there is any concept consistent with a set of positive and negative examples. Indeed, we show that tractable version space reasoning is possible if and only if there is an efficient algorithm for the consistency problem. Our observations give rise to new concept classes for which tractable version space reasoning is now possible, e.g., 1-decision lists, monotone depth two formulas, and halfspaces.     

USERS AND SYSTEM ADAPTIVITY: A GSPS Perspective

Abstract

The relationship between systems and their users is an important aspect in the study of general systems. In order to enhance their functionality, systems should be adaptive to different types of human users. In this paper, we discuss how to apply inductive reasoning techniques to achieve this goal and we use knowledge-based systems as an example to illustrate the approach. A conceptual model is described. Our approach is interpreted from the perspective of the General System Problem Solving framework     

Future Event Prediction Based on Temporal Knowledge Graph Embedding

Accurate prediction of future events brings great benefits and reduces losses for society in many domains, such as civil unrest, pandemics, and crimes. Knowledge graph is a general language for describing and modeling complex systems. Different types of events continually occur, which are often related to historical and concurrent events. In this paper, we formalize the future event prediction as a temporal knowledge graph reasoning problem. Most existing studies either conduct reasoning on static knowledge graphs or assume knowledges graphs of all timestamps are available during the training process. As a result, they cannot effectively reason over temporal knowledge graphs and predict events happening in the future. To address this problem, some recent works learn to infer future events based on historical event-based temporal knowledge graphs. However, these methods do not comprehensively consider the latent patterns and influences behind historical events and concurrent events simultaneously. This paper proposes a new graph representation learning model, namely Recurrent Event Graph ATtention Network (RE-GAT), based on a novel historical and concurrent events attention-aware mechanism by modeling the event knowledge graph sequence recurrently. More specifically, our RE-GAT uses an attention-based historical events embedding module to encode past events, and employs an attention-based concurrent events embedding module to model the associations of events at the same timestamp. A translation-based decoder module and a learning objective are developed to optimize the embeddings of entities and relations. We evaluate our proposed method on four benchmark datasets. Extensive experimental results demonstrate the superiority of our RE-GAT model comparing to various baselines, which proves that our method can more accurately predict what events are going to happen.     

Exploring the No-Function-In-Structure principle

Abstract

Although much of past work in AI has focused on compiled knowledge systems, recent research shows renewed interest and advanced efforts both in model-based reasoning and in the integration of this deep knowledge with compiled problem solving structures. Device-based reasoning can only be as good as the model used; if the needed knowledge, correct detail, or proper theoretical background is not accessible, performance deteriorates. Much of the work on model-based reasoning references the ‘no-function-in-structure’ principle, which was introduced by de Kleer and Brown. Although they were well motivated in establishing the guideline, this paper explores the applicability and workability of the concept as a universal principle for model representation. This paper first describes the principle, its intent and the concerns it addresses. It then questions the feasibility and the practicality of the principle as a universal guideline for model representation.     

Rational model selection in large engineering knowledge bases

It is generally admitted that several models differing along various dimensions are needed for executing complex engineering tasks such as diagnosis and monitoring. A key problem is thus to decide what model to use in a particular situation in front of a specified problem-solving task and reasoning objectives. We address this problem within the Multimodeling framework for reasoning about physical systems that we proposed in a previous work. After having characterized the space of possible models in the Multimodeling approach, we formulate the selection problem using the conceptual tools offered by the economic theory of rationality. In this frame we illustrate a preference-based model selection method that is used to navigate in the universe of available models of a system searching for the model that best matches a given task and reasoning objectives. The method exploits the use of a model map that is a metalevel concept representing the ontology and teleology of each model and the transformational relations (abstractions and approximations) connecting each model to other models. The model map is used to compare models on the basis of their content and to understand what can be gained or lost when switching from one model to another. Finally, some implications of the foregoing selection method in developing action-based diagnostic systems are discussed.