openlab

AI, machine learning, and applications

u-aizu.ac.jp/files/All-Posters2022JosephZhao-01.pdf

In this lab, students are conducting researches related to the following topics (not limited):
1) Deep learning-based image classfication and its application to automatic meter reading
2) Deep learning-based mul-label classification and its application to driving assistance
3) Generative neural networks and their application to defect detection

Through experiments students will become familiar with various machine learning tools and know how to solve practical problems.

ZHAO Qiangfu
Professor

We are conducting joint researches with several companies in Japan. Students will learn how to solve real problems by combining both high-tech (e.g. deep learning) and low-tech (convetional image/signal processing methods).

Cybersecurity

While the development of information and communication technology has made people's lives more convenient, cyber attacks and cyber crimes have emerged as a negative aspect.

With the widespread use of personal computers and smartphones in all aspects of our society and daily lives, it is becoming more and more important to have security technologies that can keep them safe and secure.

In our laboratory, we are conducting practical research on security measures for information systems.

We have developed technologies for detecting cyber attacks, analyzing system flaws and risks, and verification methods for security measures using simulated attacks.

We are also working on issues that are familiar to the general public when they use the Internet, such as risk analysis of information transmitted on social networking sites, countermeasures against phishing scams, and technologies to protect online privacy.

Web page of research contents:
uoanlab.wordpress.com/

NAKAMURA Akihito
Professor

Cognitive Science
Signal Processing

Everyone has a family of friends named "Signal". They can be sensed everyday in our life. When you listen to the music, it is a 1D signal processed by ears. When you read a manga book, it is a 2D matrix signal processed by eyes. When you play games, the motion of the character in 3D world is connecting your feeling basically by the mixed signal system with sound, images and haptics. They are processed by our brain to build a fantastic experience that developers and artists hope to show us.

Our lab cares signal and how it is working very much. Beyond the ability of sensing, it is able to help us to make decisions for not only behaviors as operating a machine or getting profit from financial trading in physical world, but also for creating smart behaviors of NPC as interaction in virtual world.

Web page of research contents:
u-aizu.ac.jp/~xiangli/lcglab/index.html

LI Xiang
Associate Professor

Edge AI Research and Applications (Computer Engineering)

Efforts to use AI in the field (edge) (called edge AI) will increase in the future.
Examples include surveillance systems and robots.
The inclusion of AI is expected to make these things more sophisticated.
Our lab focuses on edge AI research and applications.
In the application, we are researching a wildlife warning system that uses AI to detect bears and wild boars as part of measures to prevent damage by birds and beasts, and to notify people around them, as well as to drive them away with sound and light.
We are also planning to include AI functions in the robot to help patrol the area and remove snow.
Since these systems run on batteries, energy conservation is of paramount importance.
To this end, we are researching small, low energy-consuming edge AI devices.

Web page of research contents:
www-adl.u-aizu.ac.jp/researchtopics/

SAITO
Hiroshi
Professor

The wildlife warning system is being demonstrated in Aizuwakamatsu City, Kitakata City, and Aizu Misato Town.
Robots are an issue for the future.
We intend to continue our research to contribute to solving local issues.

Motion sensing, Sensing technology
Signal processing, Machine learning Communication networks

We are creating a digital internet world, which is independent from the real analog world we living in. But recently, as the development of the intelligent sensing technologies, the two worlds are merging together, so that in the near future, people could hardly to discriminate their difference. Our lab is focus on creating various new intelligent sensing technologies to facilitate such merging between the digital world and analog world.

Lab H P:u-aizu.ac.jp/~leijing/

Demos
1.WondeRing:
youtube.com/watch?v=MJhXZN6e5Ec

2.WonderSense:
youtube.com/watch?v=Q-snwKPvtok

3.Daily activity recognition:
youtube.com/watch?v=sm42yMv9z0E&t=18s

JING Lei
Senior Associate Professor

With our data gloves, people can directly touch the virtual world.

game AI
human-like AI
machine learning
case-based reasoning
believability

AI (artificial intelligence) is the "brain" of a computer. As a general rule, we are always striving to make AI smarter. For example, in the same way that the facial recognition function of a smartphone is capable of recognizing a face even with glasses on, a robot vacuum cleaner can avoid socks that have been left on the floor. Sometimes, however, this cleverness can also cause undesirable results. For example, clever AI can easily defeat humans in games such as chess or shogi without making any silly mistakes. Therefore, simply "creating clever AI" is not always the goal. We are using games to research various AI systems. Games need to be fun, so we need to understand what type of AI will serve as a suitable teammate or an intriguing enemy character. In our research, we focus primarily on "realistic" or human-like AI. For example, in sports games such as soccer and fighting, it is desirable to have various opponents with diverse qualities and habits, and it is important to ensure that the characters in the game do not make robot-like movements at that time. Studying human behavior is essential for creating such "realistic" characters. First, we record the behavior of characters and analyze what behavior they have in common. This is a difficult task, and our approach will depend on the genre of the game. Currently, we are studying the relatively simple games of tennis, fighting, and soccer. Soccer is particularly interesting because it requires understanding the movements of the team as a whole rather than individual plays. Games provide a simple, enjoyable experimental environment using AI systems. At the same time, we believe that our research results can be applied to areas other than games. "Fun AI" and "human-like AI systems" have significant potential for education and medical care. By studying AI, we can understand why we feel something is "fun" or "real," and therefore better understand ourselves.

Web page of research contents:
mmozgovoy.dev/

MOZGOVOY Maxim
Associate Professor

Pattern Recognition
Machine Learning
Human-computer Interaction

Our laboratory focuses primarily on human-computer interaction, pattern processing, and recognition based on signal and image analysis. As part of our pattern processing research, we have done lots of research in hand written character recognition, signature verification, font generation, human identification, human activity recognition and so on.
We are conducting research on handwriting and sensor-based hand movements to automatically detect various types of disease, such as Parkinson's disease, ADHD, and ASD disorders. We are also doing research for making a non-contact interface with computer and computer-like devices by using 2D camera and hand gestures.

Web page of research contents:
u-aizu.ac.jp/labs/is-pp/pplab/

SHIN Jungpil
Professor

game design
human-computer interaction interactive networked multimedia
LBE (location-based entertainment)
mobile-ambient systems featuring personal control of public display
stereography (3D graphics)
spatial sound
LBE (location-based entertainment)

Multimodal interaction refers to the integration of control (input) and display (output) of media, including visual, auditory, and haptic modalities: graphics, sound, and touch. The Spatial Media Group in the Computer Arts Lab conducts research on practical and creative interfaces to enhance communication and expression: 3D graphics and panoramic ("360°") imagery; spatial sound, audio, and computer music; digital typography, hypermedia and electronic publishing; smartphone & mobile computing; XR (extended reality): VR (virtual reality), MR (mixed reality) & AR (augmented reality) for immersive sensation. We investigate various kinds of stereoscopic displays ("3D" imagery) that express not only width and height but also depth,and multichannel (polyphonic) spatial audio ("3D sound") systems with rich, dynamic soundscapes, realtime applications that close the feedback loop between input and output for immediate reaction and live, online experience. We explore user interfaces that incorporate vision, hearing, and proprioception, including 3D printing, physical rigging (connection of physical controllers to virtual objects), IoT (internet of things) and "ubicomp" (ubiquitous computing) for cyberphysical systems. Groupware is software for groups, allowing teamwork and collaboration, including conferencing, team design, and musical ensemble performance. Our focus is exploration of networked multiuser interfaces with realtime multimodal interactivity, including visual musical systems, games & toys, simulations, and story-telling.

Web page of research contents:
u-aizu.ac.jp/~mcohen/

COHEN Michael
Professor

Biomedical Engineering

• Develop diversified modalities to acquire physiological information using physical and chemical principles, as well as mathematical means.
• Perform theoretical modeling, numerical simulation and data mining to reveal statistical causalities and their interaction among diseases and various temporal and spatial factors.
• Construct a "SHIP" (Scalable Healthcare Integrated Platform) to promote health and cover care cycles from womb to tomb by utilizing the latest ICT achievements such as AI, IoT, and Big Data analytics comprehensively.
• Foster a new discipline "Metrology of Health" or "Healthology".

Web page of research contents:
bitlab.u-aizu.ac.jp/

CHEN Wenxi
Professor

To contribute to people's disease prevention and health promotion by seamlessly measuring and comprehensively analyzing various biological information anytime and anywhere without affecting daily life

Visually appealing computer visualization
Computer Visualization

We are researching computer visualization, a technology that helps people understand information by converting various types of data into visual information that can be seen by the human eye.
The data that surrounds us these days is getting larger and more complex, and understanding it properly and fully to make decisions is an important skill to survive in the modern world.
And since such decisions are ultimately made by people, we need effective ways to help people understand the information.
In our laboratory, we are working on computer algorithms to visualize such data through conversion into visual information such as images, and to assist people to see and understand the data.
In particular, our research focuses on the construction of universal mathematical models that are not affected by the scale or complexity of the data, and we make maximum use of computer graphics (CG) technology as a tool for generating visual information.
In addition to an understanding of the scientific knowledge inherent in the data itself, statistics for feature extraction, and learning theory, this field also requires an understanding of a variety of related fields, such as psychology and cognitive science, in order to provide people with easy-to-understand visual information.
In this sense, computer visualization itself can be seen as an interdisciplinary field of study that involves the collaborative work of experts from various academic disciplines.

Web page of research contents:
u-aizu.ac.jp/~shigeo/research/index-e.html

Demonstration Video:
youtube.com/user/TheTakLab

Research introduction video (2020 version):
u-aizu.ac.jp/~shigeo/research/project-video.html

TAKAHASHI Shigeo
Professor

System frameworks for robots and sensors

For robots acting autonomously, we need many functions such as object recognition, localization, motion planning, control, learning, etc., and computer systems for them. For realizing the autonomous robots, we research and develop the following topics:

• Target robots: Disaster response robots, service robots for offices and buildings, factory automation and manufacturing robots, agricultural robots
• Research topics: Object recognition and pose estimation in 3D measurements, motion planning and control, robot software development with robot simulators, cyber-physical system such as modeling of physical 3D space and representation it as cyber space.

Web page of research content:
sites.google.com/view/caist-arc-robot/

NARUSE Keitaro
Professor

Our goal is to get robots closer to our daily life such as smartphones, which means multiple robots and sensors should be connected to networks, share data among them, and acts coordinately. We research and develop frameworks for the goal.

Big Data and AI

u-aizu.ac.jp/files/Lab-Intro-2022-IncheonPaik.pdf

1. automatic AI - Deep Learning (DL) service generation
Automatically creating Deep Learning (DL) services for artificial intelligence (AI) to build image recognition, signal/speech recognition, document recognition, etc. will be a new business model of the future with great potential and significance. To this end, we will conduct research on the automatic synthesis of DL services using semantic services computing technology, and on the construction of automatic Deep Learning generation services.
2. big data and Deep Learning applications
We are building and improving Deep Learning (DL) Architectures, book classification, ontology generation, intelligent chatbots, natural language translation, medical healthcare, and other deep research, as well as situational awareness, question and answer systems, chatbots, and improve them.

Web page of research contents:
ebiz.u-aizu.ac.jp/

PAIK Incheon
Professor

Building various artificial intelligence systems: translation, QA systems, medical, monitoring factory systems, etc.

An Invitation to Deep Space　
Science and Engineering for Deep Space Explorations, Lunar and Planetary Archived Data Science, Remote Sensing, Machine Learning

INTERVIEW_HIROHIDE DEMURA【FUKUSHIMA INDEX】
youtube.com/watch?v=rP53ZJEpGo4&list=PLNfX42QTjf-kaIGKW6sLtIUcDi4zDCypE&index=3

Aizu > Research Center & Cluster Introduction & Cluster Introduction & Cluster Introduction for Space Informatics（ARC-Space）
u-aizu.ac.jp/en/research/introduction/arc-space/

DEMURA Hirohide
Professor

Let's go to space together from Fukushima!　We dream of a future where Fukushima-made robots are active on the Moon.

Data-oriented research on understanding the Moon, Mars and the Earth (Space and planetary informatics)

The Moon, Mars and the Earth are similar, but still significantly different. The Earth is very unique with oceans and life. The current environments on Mars and the Moon are both like desert on the Earth, no ocean and no life. The 3 planetary bodies originated in almost the same region within the solar system. Why are they so different? I am interested in that point. We have a huge amount of data of great variety on the 3 planets. We are using those data to study the surface environment and inner structure of each planet. We are analyzing the exploration data from the Moon and Mars. We also develop tools to analyze and visualize the data. We are also conducting studies to detect ground movement based on the satellite data and to monitor active volcanoes on the Earth.

OGAWA Yoshiko
Senior Associate Professor

The Moon and Mars are our Earth neighbors. However, lots of things are still veiled. Do we really understand the Moon and Mars? Why is the Earth so unique with diversity of life? Your curiosity is a great motivation. We look forward to having you join in our data-oriented planetary science!

We're Onkyo Lab. We are interested in sound and audio.

Find out more about our student projects

onkyo.u-aizu.ac.jp/#/Students

We use sound regularly to communicate with others, yet our understanding of it is so limited that there are many opportunities for new technologies to be discovered. We are interested in:
- Spatial sound: We want to find ways to convey information via spatial (3D) sound using loudspeakers or headphones.
- Applied psychoacoustics: The processing capabilities of the brain are sometimes exceeded by hardware. We explore new interfaces such as near ultrasound communication, bass enhancement, etc.
- Applied phonetics: We are studying the effects of noise on speech, multilingualism, articulation, and phonation phenomena.

Check our website to learn mor about who we are and what we do:
onkyo.u-aizu.ac.jp/

JULIAN Villegas
Senior Associate Professor

We are interested in sound as a vehicle to transmit information between humans and machines. In our research, we often rely on machine learning methods and focus mainly on spatial sound, applied psychoacoustics, and applied phonetics.

Planetary exploration, planetary science

Our laboratory is interested in using lunar and planetary exploration data to investigate the craters, rock masses, and lava flow trajectories visible on lunar and planetary surfaces.
Recently, we have been conducting research on detecting and mapping rock masses and craters from lunar orbiter LRO images using machine learning, and on planning a route by considering the "shaped objects" detected in this way as obstacles for future lunar landing and driving exploration.
I have often assisted in analyzing the images taken by the spacecraft before landing for consideration of lunar landing exploration, but I have done so by visual survey.
However, I am beginning to feel that my eyesight is reaching its limits.
The detection method I have been promoting as an alternative method has been able to detect more than my visual survey.
Of course, there are many false positives, and this improvement is our current goal.

HONDA Chikatoshi
Associate Professor

We are working with students to ensure that the results of our research and development will contribute to future lunar and planetary exploration.

liberal arts education

The University of Aizu expects its students to become rich and healthy in mind and body at the same time as attaining a high degree of expertise in computer science and technology, so they can graduate and take responsibility for the future.

The nine faculty members, responsible for the Center's humanities, social studies, physical education, and education courses, provide an education which builds a broader background for the University's specialist education and share the responsibility for realizing the University's goals.

In particular, in "Academic Skills 1," the entire faculty of the Center teaches first-year students the study skills (logical thinking, reading and writing skills) that are essential for university study.

u-aizu.ac.jp/fileslabs/ccrs/ol2022.pdf

KARIMAZAWA,Hayato
Professor

SEINO, Masaya Senior Associate Professor

NAKAZAWA, Ken Professor

AMITANI, Yuichi Senior Associate Professor

IKEMOTO, Junichi Senior Associate Professor

KOGURE, Katsuo Senior Associate Professor

EBINA, Shoji
Senior Associate Professor

OKI, Kazusa Associate Professor

OGAWA Olivia Chisato
Senior Associate Professor

Video Introduction to the CLR Phonetics Lab" (phonetics)
youtu.be/OR3LU6AgL-k

The Center for Language Research at the University of Aizu was established in 1993 to foster research in English for Specific Purposes, with special emphasis on the English needed for study and work in the fields of computer science and computer engineering. In addition to research on other areas of second language acquisition, the CLR has always had a strong commitment to research on pronunciation and phonetics. Prof. Ian Wilson joined the university in 2006, and with the help of Prof. Kazuaki Yamauchi, acquired funding for an ultrasound machine, and established a separate laboratory dedicated to speech research - the CLR Phonetics Lab. This led to a great increase in the number of students doing phonetics research.

Besides Profs. Wilson and Yamauchi, other professors in the lab include Prof. Kaneko and Prof. Perkins. Prof. Wilson's main research interest is speech production and L2 acquisition of pronunciation. Experimental phonetics - both articulatory and acoustic phonetics - underlies most of his work. Since 2000, he has specialized in using ultrasound as a tool to view and measure the tongue during speech. Prof. Yamauchi is a native speaker and researcher of the Aizu dialect of Japanese. Prof. Kaneko also does research on the Aizu dialect, as well as English pronunciation and elicited imitation/shadowing of speech. Prof. Perkins is focused on tone and phonation, and he does a lot of fieldwork on Asian languages.

To learn more about the professors and students in our lab, please access the People link on our lab homepage. Detailed research information and downloadable papers are available from the Research link. Newspaper articles and news reports about our lab are available from the Media link. This entire website is available in both English and Japanese.

CLR Phonetics Lab website:
clrlab1.u-aizu.ac.jp/index.html

WILSON Ian
Professor

This research informs pronunciation teachers and speech scientists about the differences in pronunciation between languages - especially Japanese and English.

Phonetics and Phonology

My main research interests involve sound systems in languages of the world.
Different languages have different sound systems and when we learn these as native speakers, this fixes our brain structure in ways that make other languages harder to learn.
Students in my lab may be interested in why certain sounds ("l" and "r" or English vowel sounds) may be difficult to hear and to make for Japanese learners of English.
In addition to English, students who are interested in learning more about sounds in other languages (including Japanese) are also welcome.
Students in my lab will learn about using software to analyze speech sounds and how to do experiments.
These experiments could be about listening (speech perception) or speaking (speech production).
Students who are interested in using R programming language to help run experiments and do statistics to analyze experimental results are particularly welcome to apply.
Finally, my specific research area is about tone in languages, which is the use of pitch in languages like Chinese to distinguish words from each other.

Web page of research contents:
u-aizu.ac.jp/~jperkins/index.html

PERKINS Jeremy
Senior Associate Professor