Our proposed cognitive distance learning problem solver generates sequence of actions from initial state to goal states in problem state space. This problem solver learns cognitive distance (path cost) of arbitrary combination of two states. Action generation at each state is selection of next state that has minimum cognitive distance to the goal, like Q-learning agent. In this paper, first, we show that our proposed method reduces search cost than conventional search method by analytical simulation in spherical state space. Second, we show that an average search cost is more reduced more the prior learning term is long and our problem solver is familiar to the environment, by a computer simulation in a tile world state space. Third, we showed that proposed problem solver is superior to the reinforcement learning techniques when goal is changed by a computer simulation. Forth, we found that our simulation result consist with psychological experimental results.
We would like to see many documents in order to get useful information. Though summaries are useful pieces of documents, a document has various viewpoints to be summarized. Therefore, if a viewpoint of a summary is different from user’s, a user cannot grasp the contents of the document correctly, and the user has to see through the documents in the end. In this paper, we present a system which makes a summary based on a user’s viewpoint as user’s search keywords.
It is important to increase computation efficiency of a system at a low cost while holding the system correct. For this realization, addition of new and efficient equivalent transformation (ET) rules, whose correctness is assured, is useful. As long as correct ET rules are added to a correct system, the computation result of the system is always correct. Improvement of rules is promoted further by improvement of data structure.In this paper, we improve data structure by introducing interval variables into the usual term domain and add two correct and efficient ET rules, which are promoted by introduction of interval variables, for member constraints on interval variables. These rules are the candidate elimination rule and the common pattern specialization rule. We show by an experiment that computation efficiency is increased by using these rules.
We present an overview of a legal negotiation support system, ANS (Argumentation based Negotiation support System). ANS consists of a user interface, three inference engines, a database of old cases, and two decision support modules. The ANS users negotiates or disputes with others via a computer network. The negotiation status is managed in the form of the negotiation diagram. The negotiation diagram is an extension of Toulmin’s argument diagram, and it contains all arguments insisted by participants. The negotiation protocols are defined as operations to the negotiation diagram. By exchanging counter arguments each other, the negotiation diagram grows up. Nonmonotonic reasoning using rule priorities are applied to the negotiation diagram.
A decision tree is one of the machine learning techniques and also one of the major knowledge representations of data mining results.This is because it is easy to understand its meaning for human analysts.Even ID3, the representative algorithm, is known to exhibit remarkable performance deterioration under certain circumstances, particularly due to strong correlation between attributes representing the class of examples. One of the approaches to get more preferable decision trees is pre-processing the training data to extend its description, such as attributes generation and attribute selection. There is also the idea of decision trees with a region rule. In this paper, we consider two approaches, i.e., decision trees with a region rule allowing multiple attributes, and a pre-processing method of a region rule to enabling any suitable number of attributes to correspond to branch nodes, where an optimal division condition with arbitrarily multiple attributes is acquired. By using this method, we propose a new decision tree generation algorithm guaranteeing to select effective compound attributes with each branch node, where an MDL-based new evaluation criterion is also defined for determining the optimal number of compound attributes specified to each node.This algorithm is applied to datasets containing only nominal values. It consists of three processes: compound attributes selection, parent node integration, and pruning. We call this new decision trees DTMACC (Decision Trees with Multiple Attributes Concept Clustering). The effectiveness and comprehensiveness of the proposed algorithm are confirmed through experiments comparing to the ordinary decision trees and an effective pre-processing method.
Much of product design is executed concurrently these days. For such concurrent design, the method which can share and ueuse varioud kind of design information among designers is needed. However, complete understanding of the design information among designers have been a difficult issue. In this paper, design process model with use of designers’ intention is proposed. A method to combine the design process information and the design object information is also proposed. We introduce how to describe designers’ intention by providing some databases. Keyword Database consists of ontological data related to design object/activities. Designers select suitable keyword(s) from Keyword Database and explain the reason/ideas for their design activities by the description with use of keyword(s). We also developed the integration design information management system architecture by using a method of integrated description with designers’ intension. This system realizes connections between the information related to design process and that related to design object through designers’ intention. Designers can communicate with each other to understand how others make decision in design through that. Designers also can re-use both design process information data and design object information data through detabase management sub-system.
It has been recognized that functional knowledge used in conceptual design is scattered around technology and target domains. One of its reasons is that diffierent frameworks (viewpoints)for conceptualization are used when authors describe knowledge in diffierent domains. The other one is there are several functional concepts without clear definitions. Aiming at systematization of functional knowledge for synthesis, we discuss ontologies that guides conceptualization of artifacts from the functional point of view. We propose a device-centered ontology and a functional concept ontology. The former provides a device-centered viewpoint for capturing a target domain in order to make models or knowledge consistent. The latter provides concepts representing functions of devices which are used as a common vocabulary in functional knowledge. Some utilities of these ontologies are also mentioned.
In conceptual design, a designer decomposes a required function into sub-functions, so-called functional decomposition, using a kind of functional knowledge representing achievement relations among functions. Aimin at systematization of such functional knowledge, we proposed ontologies that guide conceptualization of artifacts from the functional point of view. This paper discusses its systematic description based on the functional ontologies. Firstly, we propose a new concept named “way of achievement” as a key concept for its systematization. Categorization of typical representations of the knowledge and organization as is-a hierarchies are also discussed. Such concept, categorization, and functional ontologies make the functional knowledge consistent and applicable to other domains. Next, the implementation of the functional ontologies and their utility on description of the knowledge are shown. Lastly, we discuss development of a knowledge-based system to help human designers redesign an existin artifact. The ontology of functional concepts and the systematic description of functional knowledge enable the supporting system to show designers a wide range of alternative ways and then to facilitate innovative redesign.
This paper proposes computerized methods of case-based design aid using similarity and relevance between design cases. First, we introduce methods of describing and recording design cases and design modification cases by physical quantities and their calculations to represent physical phenomena and terms to explain them. Then, we introduce a quantity dimension space defined by nine fundamental and supplementary quantities in SI. In quantity dimension space, a distance between physical quantities is mathematically defined based on city-block distance, and then physical quantity similarity by dimension is defined using the physical quantity distance. Then, similarity between physical quantities is defined by combining the similarity by dimension and similarity of quantity calculation structures. Also, similarity between terms is defined by combining literal similarity and cooccurrence statistics. Finally, similarity between design (modification) cases is defined based on physical quantity similarity and term similarity. By using design case similarity, designers can retrieve and consult the similar or relevant cases to a new design problem in a design case library. Also, design modification knowledge such as “which physical quantity or design parameter should be modified to solve specific trouble” can be extracted by analyzing recorded design modification cases and clustering them using design (modification) case similarity. The proposed methods are implemented as a Lisp program and are examined through some examples.
This paper describes a system that directly supports a design process in a mechanical domain. This system is based on a thinking process development diagram that draws distinctions between requirement, tasks, solutions, and implementation, which enables designers to expand and deepen their thoughts of design. The system provides five main functions that designers require in each phase of the proposed design process: (1) thinking process description support which enables designers to describe their thoughts, (2) creativity support by term association with thesauri, (3) timely display of design knowledge including know-how obtained through earlier failures, general design theories, standard-parts data, and past designs, (4) design problem solving support using 46 kinds of thinking operations, and (5) proper technology transfer support which accumulates not only design conclusions but also the design process. Though this system is applied to mechanical engineering as the first target domain, it can be easily expanded to many other domains such as architecture and electricity.
In this paper, we propose a fundamental idea of a new CAD mechanism to facilitate design knowledge management. This mechanism encourages a designer to externalise his/her knowledge during a design process and facilitates sharing and reuse of such externalised design knowledge in later stages. We also describe the implementation of this idea called DDMS (Design Documentation Management System). DDMS works as a front end to KIEF (Knowledge Intensive Engineering Framework), which we have been developing. We also illustrate an example of machining tool design to demonstrate the features of DDMS.