Researcher List

Laboratories in Department of Data Science

Management Systems Laboratory
Prof. Masatsugu SHIIHARA

Management Systems Laboratory

Our research focuses on the optimal design and operation of various kinds of management systems, with particular attention given to production and education systems. The methodologies used in our research include Industrial Engineering (IE), Operations Research (OR), and computer simulations.

Main Research Topic

Improvement of production schedules by using batch splitting

It is commonly accepted that batch splitting can improve various evaluation measures, but as the number of batches is increased through splitting, shop floor control and scheduling problems tend to become more difficult. Thus, there is a great need for efficient batching and scheduling methods. We have developed a genetic algorithm (GA) that prevents searching efficiency from being reduced when the number of batches is increased through batch splitting, and we are evaluating its properties using computer simulations.

Advanced Software Laboratory
Prof. Seigo YASUTOME

Visualization of the progresss
Visualization of the progresss

We are working on system development using open source software (OSS), which is also used in smartphones and televisions. In particular, we are focusing on building web applications for educational purposes. We have developed a web application that monitors the progress of students using PCs in the exercise room, and we are operating a system that teachers and TAs to monitor the progress of students using tablet devices in real time while moving around the room, which is useful for teaching.

Main Research Topic

  • Research on educational support system

Manufacturing Management Laboratory
Prof. Kentaro MINAGAWA

Analysis of the workflow by video recording
Analysis of the workflow by video recording

In the manufacturing process, the 4M of Man, Machine, Material, and Method are input and the QCD of Quality, Cost, and Delivery are output. It is a simple system, but there is still a lot of waste and loss in the process.
Recently, advances in sensor technology have made it possible to easily measure this data. However, just being able to see them does not solve these problems, and Kaizen activities to make them better are important. As the trend toward smart manufacturing continues, expectations for cyber-physical systems (CPS) are also rising, and in this laboratory, we are working on research to promote Japanese smart manufacturing by integrating the wisdom and ingenuity of the workplace.

Main Research Topics

  • Approaches to productivity improvement using visualized data
  • Study of a real-time visualization system for the workplace using IoT sensors
  • Research on the relationship between SDGs activities and corporate operating profit margins

Distributed Information Processing Laboratory
Prof. Takayuki SUYAMA

Illustrative Image of Distributed Information Processing
Illustrative Image of Distributed Information Processing

Currently, there are many sensors in our living environment. We will carry out research and development on artificial intelligence (AI), which uses distributed information processing technology to efficiently process the information (images, temperature, etc.) obtained from these sensors and extracts meaningful knowledge for users.
In distributed information processing, we process the information using IoT devices with built-in sensors and processors which enable edge computing and the cloud (center server) which is capable of large-scale data processing.

Main Research Topics

  • Research on optimal operation technology through cooperation between edge nodes such as IoT devices and the cloud (center server)
  • Research on methods to reflect learning results in the cloud from information acquired from sensors in edge nodes to sensor nodes.
  • Research on methods for applying methods for sensing environmental information to real environments
  • Research on distributing and executing methods for necessary operations among multiple edge nodes and clouds (distributed artificial intelligence)
  • Research on methods for efficiently performing operations on edge nodes with limited resources.

Statistical Data Inquisitive Laboratory
Prof. Etsuo HAMADA

Linear regression model
Linear regression model

I am conducting research on various types of actual data to obtain new insights from the data by applying statistical models.

Examples of specific topics include:
-Multiple regression analysis for data with a large number of explanatory variables,
-Statistical inference and testing for data with outliers,
-Estimation of correlations for data that seem to be related.

Main Research Topics

  • General data analysis
  • Propensity scores and double robustness in statistical causal inference
  • Consistency of exponential families with probability distributions

Internet of Things System Laboratory
Prof. Hideo ARAKI

Image of embedded microcomputer system
Image of embedded microcomputer system

A specific application microcomputer system with network function is called IoT (Internet of Things) system. Our study scope is including an applicated embedded microcomputer system and a design method for it. Furthermore, the many application systems have some sensors, and the system design include controlling the sensors and sensor circuits. For example, a pressure sensor and a gyro sensor application are used for sensing human movement. Through our study, to make a safe and pleasant society is our goal.

Main Research Topics

  • Microcomputer applied embedded system and the design method
  • Sensors and Sensing system with a microcomputer
  • Human user interface device and system

Intelligence Modelling and Soft Computing Laboratory
Prof. Yoichi HIRASHIMA

Smooth logistics will support the continued development of the global economy.
Smooth logistics will support the continued development of the global economy.

Our research topic is focusing on design method of an autonomous learning system that extracts knowledge by deploying various data in the form of network. In order to obtain the most desirable results from the vast scope of information on the cyber-physical environment, the valuable challenges arise from realistic and detail constraints in the modern society. For example, our research interests include learning systems for autonomously finding the best work schedule for production and distribution processes, such as scheduling for port container remarshaling, scheduling for material handling operation in shunting railroad cargo, and route planning for auto-guided vehicles.

Main Research Topics

  • Scheduling for port Container premarshaling/remarshaling
  • Scheduling for railroad cargo in shunting train
  • Path planning for auto-guided vehicles
  • Augmented Reinforcement learning for fine scalable AI

Computational Social Science Laboratory
Associate Prof. Fumihiro SAKAHIRA

Schelling's segregation model  (by artisoc Cloud)
Schelling's segregation model (by artisoc Cloud)

Computational social science is a new research field that integrates the humanities and sciences, using big data and artificial intelligence technology to approach problems in social science that target human behavior and social phenomena.
In this laboratory, we use data analysis such as data mining and text mining, and agent-based social simulation to scientifically explore the mechanisms of human behavior and various phenomena that have occurred in human society from the past to the present.

Main Research Topics

  • Research on the elucidation of patterns and processes of cultural evolution (Related to Anthropology, Archaeology)
  • Research on the extraction of disaster-related events and their causal knowledge from text data (Related to Disaster prevention)
  • Research on the relationship between team management and performance in professional sports (Related to Management Science, Organizational Science)

Time Series Data Analysis Laboratory
Assistant Prof. Shoichi EGUCHI

Time Series Data Analysis Laboratory

The purpose of this research is to understand the laws (models) of time series data and to construct statistical inference and analysis methods. Specifically, we are planning the following activities.
Model construction and estimation: How can we construct a model that explains the data?
Evaluation and selection of models: When there are several possible models to explain the data, which one can explain the data best?
Software development: development of modules to calculate the above.

Main Research Topics

  • Data analysis and module development
  • Development of statistical inference methods