The purpose of this research is to predict the subjects which will become the fashion in an electronic bulletin board in the near future. We proposed the technique which analyzes propagation of the subject based on link information. To extract the pattern of propagation, we proposed several criteria to measure the fashion degree of the subject based on link information which appears in contributed articles. We realized prediction method with unknown subject in fashion using the classification by Support Vector Machine. We conducted experiments to verify the validity of this technique with known collected fashion-subjects.
This paper introduces a method of representing in a network the thoughts of individual authors of dogmatic texts numerically and objectively by means of co-citation analysis and a method of distinguishing between the thoughts of various authors by clustering and analysis of clustered elements, generated by the clustering process. Using these methods, this paper creates and analyzes the co-citation networks for five authoritative Christian theologians through history (Augustine, Thomas Aquinas, Jean Calvin, Karl Barth, John Paul II). These analyses were able to extract the core element of Christian thought (Jn 1:14, Ph 2:6, Ph 2:7, Ph 2:8, Ga 4:4), as well as distinctions between the individual theologians in terms of their sect (Catholic or Protestant) and era (thinking about the importance of God's creation and the necessity of spreading the Gospel). By supplementing conventional literary methods in areas such as philosophy and theology, with these numerical and objective methods, it should be possible to compare the characteristics of various doctrines. The ability to numerically and objectively represent the characteristics of various thoughts opens up the possibilities of utilizing new information technology, such as web ontology and the Artificial Intelligence, in order to process information about ideological thoughts in the future.
It is well known that local search (LS) improves the performance of genetic algorithms (GA) in single objective optimization, and it has recently been reported that the hybridization of GA with LS is effective in multiobjective combinatorial optimization as well. In most studies of this kind, LS is applied to the solutions of each generation of GA, which is the scheme called ``GA with LS'' herein. Another scheme, in which LS is applied to the solutions obtained with GA, has also been studied, which is called ``GA then LS'' herein. It seems there is no consensus in the literature as to which scheme is better. The situation in the multibojective function optimization literature is even more unclear since the number of such studies in the field has been small. However, some argue that LS contributes marginally to improving the performance of GA in multiobjective function optimization.
This paper, assuming that objective functions are differentiable, reveals the reasons why GA is not necessarily effective in finding solutions of high precision, and hence hybridizing it with LS is indeed effective in multiobjective function optimization. It also suggests that the hybridization scheme which maximally exploits both GA and LS is GA then LS. Experiments confirmed that GA is not suitable for obtaining solutions of high precision, and GA then LS performs better than GA and GA with LS on many benchmark problems.
This paper presents a method of finding a specification page on the Web for a given object (e.g., ``Ch. d'Yquem'') and its class label (e.g., ``wine''). A specification page for an object is a Web page which gives concise attribute-value information about the object (e.g., ``county''-``Sauternes'') in well formatted structures. A simple unsupervised method using layout and symbolic decoration cues was applied to a large number of the Web pages to acquire candidate attributes for each class (e.g., ``county'' for a class ``wine''). We then filter out irrelevant words from the putative attributes through an author-aware scoring function that we called site frequency. We used the acquired attributes to select a representative specification page for a given object from the Web pages retrieved by a normal search engine. Experimental results revealed that our system greatly outperformed the normal search engine in terms of this specification retrieval.
As computer scientists, we have been trained in the methodology of natural science, which is analytic in its essence. Informatics, and particularly Artificial Intelligence, is not an analytic discipline. It is required to establish a constructive methodology.
Error-based Simulation (EBS) is a framework for assisting a learner to become aware of his error. It makes simulation based on his erroneous hypothesis to show what unreasonable phenomena would occur if the hypothesis were correct, which has been proved effective in causing cognitive conflict. In making EBS, it is necessary (1) to make simulation by dealing with a set of inconsistent constraints because erroneous hypotheses often contradict the correct knowledge, and (2) to estimate the 'unreasonableness' of phenomena in simulation because it must be recognized to be 'unreasonable' by a learner. Since the method used in previous EBS-systems was much domain-dependent, this paper describes a method for making EBS based on any inconsistent simultaneous equations/inequalities by using TMS (it is called 'Partial Constraint Analysis (PCA)'). It also describes a set of general heuristics to estimate the 'unreasonableness' of physical phenomena. By using PCA and the heuristics, a prototype of EBS-system for elementary mechanics and electric circuit problems was implemented in which a learner is asked to set up equations of the systems. A preliminary test proved our method useful in which most of the subjects agreed that the EBSs and explanations made by the prototype were effective in making a learner be aware of his error.
In this paper, we discuss guidelines for a reward design problem that defines when and what amount of reward should be given to the agent/s, within the context of reinforcement learning approach. We would like to take keepaway soccer as a standard task of the multiagent domain which requires skilled teamwork. The difficulties of designing reward for this task are due to its features as follows: i) since it belongs to the continuing task which has no explicit goal to achieve, it is hard to tell when reward should be given to the agent/s. ii) since it is a multiagent cooperative task, it is hard to decide what is a fair share of reward for each agent's contribution to achieve the goal. Through some experiments, we show that the reward design have a major effect on the agent's behavior, and introduce the successful reward function that makes agents perform keepaway better and more interesting than the conventional one does. Finally, we explore the relationship between `reward design' and `acquired behaviors' from the viewpoint of teamwork.
In this paper, we propose an extension of the Minimal Generation Gap (MGG) to reduce the number of fitness evaluation for the real-coded GAs (RCGA). When MGG is applied to actual engineering problems, for example applied to optimization of design parameters, the fitness calculating time is usually huge because MGG generates many children from one pair of parents and the fitness is calculated by repetitive simulation or analysis. The proposed method called Saving MGG reduces the number of fitness evaluation by estimating the promising degrees of children using individual distribution and fitness information of population, and selecting children based on the promising degree before evaluating the fitness. Experimental results show that RCGA with Saving MGG can provide large reducing effects on 20 or 30 dimensional Sphere functions, Rosenbrock functions, ill-scaled Rosenbrock functions, and Rastrigin function.