Building a Research Cluster Model Activities of the University of Aizu Cluster

Birth of the CAIST and it's purpose

"Thirteen years ago, on April 1, 2009, the Research Center for Advanced Information Science and Technology (CAIST) was established at the University of Aizu." Professor Cho, with his gentle smile, told us about the project. Professor Zhao, a doctor of engineering and vice president of the University of Aizu, set the following three goals for the establishment of CAIST.

  1. Promoting advanced research in information science and research that responds to social needs.
  2. Actively promote collaboration with other universities, companies, research institutions, etc.
  3. To create a center for research that contributes to the promotion of local industry and the creation of new industries.

Professor Zhao said, "With these goals in mind, we have spent more than a decade working on a variety of research projects within CAIST, and a new research cluster concept has been developing."

What is a cluster in the first place?

Professor ZHAO Qiangfu

Now in the year 2022, we are facing the disaster of the COVID-19 pandemic. Although media reports have tended to give the impression that clusters are far removed from the original meaning of "mass infection," the word "cluster" is widely used in various fields to mean "group," "aggregate," or "agglomeration". Therefore, we asked Professor Zhao to tell us again about the cluster activities at the University of Aizu, and what kind of initiatives the "research cluster model" mentioned in the title above refers to.

From "Ten", "Chi" and "Jin" to the creation of the "Research Cluster Model"

Originally, the University of Aizu had its first cluster, "Ten-Chi-Jin" with "Ten" classified as space exploration, "Chi" as meteorology, and "Jin" as medicine. Then, as mentioned above, CAIST was established in April 2009. In October 2020, we established the "Research Cluster Model" as a university-wide research organization, with the aim of further strengthening our research capabilities and stimulating a greater variety of research fields. This "research cluster model" means that existing research teams will be subdivided into smaller groups, while professors from various fields will brush up on each other's work to create a new collective body, which will be divided into three major phases. The first is the "Basic Cluster" (commonly known as the "B Cluster"), which consists of up-and-coming research teams. The second is the "Chirst Cluster" (commonly known as the "A Cluster"), which is positioned as the upper class of the B Cluster. The Research Centers are the highest level of the Research Cluster model. The research promotion body consisting of these three clusters has six groups in the B cluster, three groups in the A cluster, and three groups in the research center. Two groups have established their own unique research problems and are working daily to develop solutions.

Cluster Growth Model Overview

Creating a brand for the University of Aizu through mutual collaboration among research clusters

Professor Zhao says.The University of Aizu is building on its current "research cluster model" of Research Center, Cluster A, and Cluster B, while continuing to create new clusters to further strengthen its research capabilities and contribute to society. We will create cutting-edge clusters that can respond to social changes by increasing the variety of research. As Prime Minister Kishida has asserted, "Research is the foundation of a nation," the government has positioned project research to promote AI and data science as a way to enhance national power in the world. We aim to become a top-class research institution. The world's trends are changing rapidly and drastically, and solving today's problems may require a different kind of solution tomorrow. With an eye on the rapidly changing world of today, the University of Aizu will focus on what it can do, what it can do because it is the University of Aizu, and what it can do only if it is the University of Aizu. We will disseminate our cutting-edge research activities in information science from Aizu to the world, gather many colleagues from around the world for joint research, build clusters that create unknown values, and lead to one-stop solutions. Expectations are even higher for the future evolution of the "research cluster model" that the University of Aizu is aiming for.

IoT cluster to promote research and development of devices indispensable to daily life

Here are some specific examples of cluster activities. The "IoT Cluster (ARC-IoT)," established a little more than a year ago, is engaged in research on the development of AI-equipped devices. It is a high-profile research project that was promoted to the A Cluster at an exceptionally fast pace in recognition of its research to date. The Internet of Things (IoT), or the Internet of Things, which is the subject of this research, is a system that allows all kinds of information in the world to function with interactive information via the Internet. We are working on research that makes full use of the IoT in devices such as smartphones and tablets to make effective use of the IoT. For example, batteries are necessary when using sensors outdoors where there is no electricity, such as when measuring gas usage or checking the availability of parking spaces. Our goal is to develop compact IoT devices (chips that control the on/off switching of sensors) that save energy and are suitable for outdoor use, in order to prolong the life of the batteries as much as possible. Dr. Saito, who spoke thus, is working to develop compact, energy-saving IoT devices that integrate both the hardware research as well as the software to operate the devices.

Senior Associate Professor Saito Hiroshi

Research and development of AI-based wildlife detection systems

Recently, the development of smarter devices that incorporate a camera and AI processing of captured images has become mainstream, and Dr. Saito is also researching the configuration and design of smart devices that incorporate AI as much as possible. As an application of this research, he has begun working on the development of a wildlife detection system. We have been working on this effort for four years," he says. At the time, there were reports of bears appearing in the town of Aizu Wakamatsu, and as a hardware development researcher, I thought that if we could use AI to detect bears and quickly disseminate the detection information, we could prevent damage before it happens," he said. However, university research is generally about methodology, and "I had no particular experience in creating a detection device that could detect moving objects, so I had a lot of difficulties," he said. The same seems to be true for AI. Even though detection was possible in test images using AI models provided as open source, detection was not possible in field demonstration tests. So, they repeated the process of reprogramming the AI over and over again, and finally became able to detect bears around July of last year. "The AI model needs to learn to extract the characteristics of bears from their images. Unless the model is trained using the images obtained in the demonstration tests, rather than just images downloaded from the Internet, it will not be able to successfully detect bears in the demonstration tests."

Subsequently, the development of an automatic bear detection system began in earnest as a project commissioned by the Aizu Regional Promotion Bureau of Fukushima Prefecture. As soon as the accuracy of automatic bear detection began to improve, we were approached by people asking if we could scare them away. In fact, the system was confirmed to drive the bears away using sound and light, but they would reappear after a few weeks. Experts say that if you don't give it an aversive stimulus, it will appear again and again, but with the current AI, it may misidentify the bear because it recognizes that bear = black object, so it will be a little while before we are able to drive it away by inflicting pain. That is why we believe that learning by using more images is the key to solving the problem," said Dr. Saito. Four students from the cluster are involved in this bear detection system demonstration experiment in terms of maintenance, such as collecting images in the field (Aizu Misato-cho and Kitakata City) and dealing with equipment malfunctions, etc. In FY2021, they will conduct the experiment once every two weeks. They collected still images from nine units of the equipment. "In the beginning, it took more than an hour to retrieve 10,000 to 20,000 images. It was very hard during cold weather, such as in early spring. In response to these challenges, a graduate student spontaneously modified the image retrieval program, and now it takes only about 10 minutes to retrieve the images. I was also very encouraged by the student's efforts to improve the equipment to detect bears in multiple directions, not just in the front, while reducing the size of the existing detection equipment."

Image of a black bear captured by a wildlife detection system.

Image of a black bear captured by a wildlife detection system.

Some of the videos taken by the trail camera

Some of the videos taken by the trail camera
*These results were obtained through a demonstration project commissioned by the Aizu Regional Promotion Bureau of Fukushima Prefecture.

"As a future aspiration, I would like to improve the accuracy of AI and develop a system that can detect not only bears but also wild boars. I would also like to respond to a request from the Aizu Regional Promotion Bureau of Fukushima Prefecture to conduct a survey to identify individual black bears, and I would like to work on a system that can track bears by installing several cameras over a larger area. These activities are just one research project for the cluster, but by applying this research mechanism, we can shift our focus to a variety of issues. For example, I am convinced that the possibility will expand to solve more different social issues by installing the relevant sensors, such as finding illegal dumping, landslides, river flooding, and so on. The devices we are researching and developing are necessary for this purpose." The future connected by AI and IoT devices. How will our lives change in the future? We cannot take our eyes off of Dr. Saito and the other researchers at Cluster.

Research Center, Cluster A, and Cluster B groups

The two groups positioned in the research center are

1. Space Information Science Center
As of April 1, 2019, the Center was approved by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) as a "Lunar and Planetary Exploration Archive Science Center" for joint use and research, and was transferred from the A Cluster "Space Information Science Cluster (ARC-Space)" to a research center. Currently, it is also participating in JAXA's Hayabusa2 Project.
2. Robotics and Information Engineering Cluster
The center is currently an A-cluster, but also functions as a research center, as it has been recognized for its research efforts in serving as a national and prefectural robotics development center.

The three groups positioned in Cluster A are as follows

1. Bioinformatics Cluster
Research and development in a wide range of areas related to health, medicine, and nursing care by integrating the fields of medicine and engineering.
2. cloud cluster
Research and development to propose solutions to various security issues in systems such as sensor nets, robotics, and energy management.
3. cluster
Research and development of AI and devices that are compact and energy efficient while integrating hardware and software.

The following six groups are positioned in the B cluster

1. Intelligent Networking
Research and development of safe, secure, scalable, and smart next-generation network technology.
2. Satellite Data Utilization
Research on software and technology related to disaster prevention, reconstruction, and resource utilization using remote sensing data acquired by earth observation satellites.
3. Smart Design
Research and development aimed at "intelligence" of the knowledge creation process itself, overseeing the description, acquisition, and updating of knowledge related to "design" and other processes.
4. Smart Service
Research and development of cutting-edge technologies related to cloud-based e-learning, e-commerce, etc.
5. Vision Computing Platform
Research and development related to the realization of human "visual functions" using deep learning and other technologies.
6. Automated AI System Design
Research aimed at analyzing and solving difficult tasks in machine learning and deep learning, and automating a series of processes such as design, development, operation, and management of AI systems.

Each group, which conducts advanced and diverse research, is engaged in research and development according to the respective phases of the Research Center, Cluster A, and Cluster B.


ZHAO Qiangfu

Professor of Computer Science, School of Computer Science and Engineering. He is also the Vice President of the University of Aizu.

Profile Details
ZHAO Qiangfu | Faculty List

SAITO Hiroshi

Senior Associate Professor of Computer Science, School of Computer Science and Engineering. He is himself a graduate of the University of Aizu.

Profile Details
SAITO Hiroshi | Faculty List