Tetsuya Maeshiro

An evolutionary system for the prediction of gene regulatory networks in biological cells

Estimation or prediction of gene regulatory networks is an important problem for the elucidation of biological mechanisms. This paper presents a method to predict gene regulatory networks from gene expression time course data, based on a ultra high speed gene network simulator Starpack and evolutionary mechanisms. The proposed method predicts gene regulatory network from gene expression time course data. The prediction is a combination of two stage loop, each stage using evolutionary mechanism. Networks are simulated with Starpack, and those producing time course data similar to the target gene expression data are selected as candidates. The simulation of the second stage has higher precision than the first stage, serving as local optimization process. Five synthetic networks were tested, and the performance of the proposed method was higher than conventional methods.

Hypernetwork Model to Represent Similarity Details Applied to Musical Instrument Performance

This paper treats the quantification and description of similarities among entities being represented as a network. The proposed representation model, hypernetwork model, allows more specific description of relationships among represented entities than conventional knowledge representation models. Musical instruments performance is represented with hypernetwork model. Detailed description of similarity relationships provided by the hypernetwork model enables the discrimination of various types and degrees of similarity. A method to compare similar relationships is also discussed, which leads to the analogical reasoning, associative search and retrieval.

A System Description Model with Fuzzy Boundaries

Describing phenomena of interest as a system is valuable to analyze using system science methodologies. The boundary is considered as the necessary component of a system, through which the system interacts with its environment. Although system based analysis is applicable, not all phenomena seem to present boundaries. We discuss boundary description of two phenomena, namely the lifestyle disease and the music composition process. The hypernetwork model homogenizes boundaries and relationships, and boundaries can be treated as an instance of relationships.

Representation of decision making process in music composition based on hypernetwork model

Music composition is treated as a repetitive decision making process, and represented using the hypernetwork model, the proposed model. The hypernetwork model allows more specific description of relationships among represented entities than conventional knowledge representation models such as semantic network. Music composition of a musical piece of 100 measures (performance duration of 6-8 minutes) by a professional composer is analyzed based on the description of decisions involved. Musical piece represented as musical scores are represented with hypernetwork model where the decisions are the representational units. Single or multiple decisions are related with other decisions, and quantitative similarity based on relationality among decision sequences provided by the hyperlink model enables the discrimination of various types and degrees of similarity.

Analysis of small-world networks with high broadcast speed

Given a network where each element has a small number of connected elements, the randomization of connections with small probabilities results in networks with a high broadcast speed. Such networks belong to the class of so-called small-world networks. Analyses indicate that broadcast to all elements of the network is fundamentally different from the average transmission speed. Networks with 10% randomness give the fastest broadcast speeds with no dependence on the position of the initial transmission source.

A System Description Model to Integrate Multiple Facets with Quantitative Relationships Among Elements

We propose a framework of system description model to represent directly the details of relationships among elements, for the quantitative analysis of individual relationships and the whole described system. The proposed model also enables analyses of dynamic aspects of the system integrating the specifications of relationships described in the system. Multiple types of relationships can coexist in the same representation.

Prediction of Standing Ovation of TED Technology Talks

This research aims at the prediction of whether speeches of TED talk can cause audience standing ovation after the end of the talk. The phenomenon of audience standing ovation that we can see in TED talk is one of the objective evidence of the effect that speeches give to audience. We gathered TED talk data that we used as data to experiment the prediction. The methods of this present research consist of quantitative analysis according to speech content and machine learning technique by convolutional neural network. As a result, we achieved 77.11% accuracy and 0.63 F-measure from the prediction using TED talks of Technology topic. Our method used in this study is useful to predict occurrences of standing ovations, although improvement is necessary. Compared to other studies, our contribution, on the one hand, is that we focused on speech content as the effect of standing ovation. On the other hand, we incorporated quantitative analysis especially in terms of what features are effective to standing ovation and eventually apply those features to machine learning technique.

A Mechanism to Control Aggressive Comments in Pseudonym Type Computer Mediated Communications

We propose a mechanism to alleviate the aggression on Computer Mediated Communications. A title reflecting the aggressivity of comments posted by the user is displayed on screen. The effects of the proposed mechanism are verified with laboratory experiment, where participants post comments after reading the topics of discussion and other participants’ comments. The results indicate the validity of the proposed mechanism.

A System Description Model Without Hierarchical Structure

In order to simulate and analyze the properties of a phenomena from systems point of view, a proper description is necessary, otherwise these objectives cannot be accomplished. We are currently describing two superficially distinctive phenomena, the feeding process of living organisms and music composition process by composers, but they share some fundamental properties from the system description point of view. Here we discuss these common characteristics and the requirements for the system description model.

Computer System for Musicians and Composers to Analyze Music Composition Process

This paper presents a computer system and its interface for musicians and composers to analyze musical pieces described as a sequence of decision making process during the composition of musical pieces. Representation of musical pieces from the viewpoint of creation process is valuable for both composers and musicians. For composers, it is valuable to verify her own composition techniques and creative process. And for musicians, it offers different viewpoints to understand the musical piece that results in better execution of the musical piece.