Distinguished University Focusing on Computer Science and Engineering Education for Cultivation of Global IT Innovators

Zixue+Cheng+Qiangfu+Zhao+Shuxue+Ding+Abderazek+Ben+Abdallah+Wenxi+Chen

Abstract： The University of Aizu is one of the first group of universities in Japan that are selected in the Top Global University project. We introduce three characteristics of the university： internationalization， IT specialty education， and the technical innovational and business startup in new era. We will also discuss the related educational research directions， and introduce the methodology and curriculum to achieve the education goal for educating innovative IT talents.

Key words： Top Global University project； education for innovation； post graduate/student ventures； advanced ICT global program； all courses in English

1 Founding Philosophy and Its Features of the University of Aizu

Level of education in Aizu region was high from the beginning of 19th century. Aizu Clan School Nisshinkan was famous and one of the representatives of Aizu education. To cultivate samurai with strong mental power， it had a strong fundamental philosophy （Confucianism） and martial arts. Among the alumni of the school， some became the president of Tokyo University and Kyoto University， etc. However， after the civil war （in 1868）， as a famous battle happened in Aizu， education had not been supported by the government for many years. There was a great need from Aizu region to build a public university.

To make the long-cherished dream of Aizu come true， the University of Aizu （UoA in short） was established in 1993 in response to the request. Its founding philosophy is “To Advance Knowledge for Humanity”.

Along with the philosophy， the University of Aizu has the following three features.

（1）Internationalization： 40% of faculty members are from overseas， and education and administrative meetings are conducted in English.

（2）Specialized in computer science and engineering research and education deeply and widely， as well as rich expertise in ICT（Information Communications Technology） field.

（3）Encouraging students and graduates to launch the start-up companies or ventures and having a series of entrepreneurship courses. The number of ventures is ranked No. 1 in public universities in Japan.

2 The University of Aizu as a Top Global University

The Top Global University （also called Super Global University） project is a new funding scheme designed and supported by Ministry of Education， Culture， Sports， Science and Technology （MEXT） in Japan. It will last for 10 years aiming to achieve leading universities in the world. Only two public universities have been selected， and receiving the support from the MEXT from 86 public universities in Japan. The goal of the UoA Top Global University is to cultivate the ICT global innovators. The UoA Top Global University has been mainly working on the following three projects.

（1） Honors program which encourages and awards the challengers with high motivation.

We will provide challenging courses for the students in this program and spatial environment for students to implement their ideas， such as honors maker room. Support and services will be provided to such students selected based on their achievement and motivation

（2）Advanced ICT global program.

We will recruit overseas students from high schools or its graduates to join the program. All the education activities including lectures， exercises， seminars， and projects will be conducted in English as intermediate language. Students without Japanese language skills can enroll in the program and complete all the necessary credits. Japanese language and cultural courses are also offered to students to have opportunity to know Japan and find jobs in Japan.

（3）Spatial internship and entrepreneurship program.

A unique internship program consisting of three stages has been designed. The first stage is to experience at the local venture companies. The second stage is to do an internship at domestic companies in Tokyo， Osaka， and other cities. The third stage is to go overseas to do an internship， especially Silicon Valley， the most innovative place in ICT field. There are also a series of courses on entrepreneurship education and experiencing starting up ventures.

3 Advanced Research and Education in the University of Aizu

In the University of Aizu， research and education in computer science and engineering are widely and deeply， at 29 research laboratories. In this article， we have selected and introduced some activities of four laboratories as representative examples. They are heads of each laboratory that have a background and relation with China.

3.1 Machine learning and its applications

Research activities in the System Intelligence Laboratory headed by Prof. Qiangfu ZHAO are focused mainly on machine learning and its applications. So far they have proposed several machine learning models （e.g. neural network decision trees （NNTrees）， nearest neighbor classifier trees （NNC-Tree）， etc.）， and learning algorithms （e.g. the R4-rule， the decision boundary making （DBM） algorithm， etc.）. Students in this laboratory are encouraged to conduct free， creative， and original researches and publish their results in international conferences or academic journals. Current research topics of the students include， but not limited to， induction of compact and high performance aware agents for mobile/wearable devices； realization of secure and privacy-preserving cloud-based intelligent agents； biometric-based user authentication； automatic image morphing and morphing-based steganography； machine learning-based universal steganalysis； minor component-based image recognition； privacy-preserving activity identification based on homogeneous sensor array， and so on. Normally， each master student can publish at least one international conference paper during his/her 2-year master program. PhD students usually perform much better. During the last 20 years， several students have received the “President Award”， which goes only to the best student of the entire university each year.

Prof. Qiangfu Zhao is a leading researcher in computational awareness （CA）. He was the founder and now a co-chair of the technical committee on awareness computing under the umbrella of the IEEE Systems， Man， and Cybernetics Society. CA is a discipline for studying the awareness behavior and mechanism inside a brain. The purpose of CA is to realize aware systems that can detect novel events in a timely manner， inform the human user， and/or help the human user to make correct decisions. Equipped with various aware agents， we human beings can become more aware， and can avoid risks or capture chances. We can understand aware agents or systems using models developed in the context of machine learning （e.g. various types of neural networks）. One question we must answer is how we can interpret and justify the decisions made by a machine. This is extremely important if the decisions are related to our properties， health， or even lives. To answer the above question， Zhao is now trying to model aware systems using three-valued logic， and has proposed a way to convert deep structured aware systems into expert systems. Using expert system， we can track， verify， and assign meaning to the reasoning process. We believe that this research is a correct direction for us to realize human-like and human-friendly artificial intelligence.

3.2 Signal and information processing， neuro-computing

Research activities in the Cognitive Science Laboratory led by Prof. Shuxue Ding include mainly the signal and information processing， neuro-computing， machine learning， intelligence， functional logic programming， etc. He and his group had many contributions， especially on compressive sensing， sparse representation and dictionary learning， nonnegative matrix factorization， blind source separation， and independent component analysis， and had been involved in the initialization and foundation of these fields. Prof. Ding has also served for many years as visiting researcher of Brain Science Institute， RIKEN， which is a famous institute in the world for the related fields. Students in the laboratory， including many Chinese students， are very active and progressive for their researches and achieved a lot. A student won the Best Paper Award of 2015 IEEE International Conference on Digital Signal Processing （DSP 2015， Jul. 21-24， 2015， Singapore）； another student won the Best Student Paper Award in 2015 IEEE International Conference on Awareness Science and Technology （iCAST 2015， Sept. 22-24， 2015， Qinhuangdao， China）. Some PhD students were awarded by JSPS fellowship， even though the fellowship is usually for post-doctors.

Currently， Prof. Ding and his group are engaging themselves in the research of sparse representation and dictionary learning， and compressive sensing， which have many applications or potential applications and are considered as an extension of the conventional information theory， so that they become very hot research topics in the scope of the world and may be considered as representative founded since this century.

The sparse representation says that most samples of the signal or image are zero or near zero valued. This property can be enforced by introducing a dictionary for representation of the signals， which includes a group of basic “bases” or "words"， i.e.， "atoms". A very attractive feature is that， if a sparse signal is corrupted with noise or mixed with other signals， it will lose its sparsity or become less sparse. The sparse representation is just to find a piece of signal or image processing that can recover， or reconstruct the pure signal， or image by requiring the output with a higher sparsity. Compressive sensing， on the other hand， tries to take advantage of signal sparsity to make more efficient communication but with less bits， which will definitely become a standardized technology for the next generation of telecommunications.

Amusingly and surprisingly， the sparse representation and dictionary learning have some intimate relationships with the deep leaning， another hot research area. In fact， the trained deep nets， in each layer， can be arranged as a dictionary for sparse representation. Prof. Ding is trying to use his dictionary leaning results to directly apply to deep net layers so that a training is no more necessary. This may greatly improve the deep learning in its training phase， since it usually needed a huge amount of data and extraordinary computer resources. Or， conversely， the deep learning can be used as a method for learning dictionary.

3.3 Adaptive computing systems

Research activities in the Adaptive Systems Laboratory headed by Prof. Abderazek Ben Abdallah are focusing on computer and system architectures， which are expanding beyond the usual focus on performance to have other quantitative and qualitative criteria. The quantitative criteria mainly include energy/power consumption. The qualitative criteria include adaptability， efficiency （e.g. performance/area or performance/energy）， and reliability which is important in harsh environments， such as an environment with high radiation （i.e. space missions）. There are several enabling techniques or paradigms， such as reconfigurable fabric， ASIP， and neuro-inspired， that allow a computing system to perform such an adaptation to achieve the criteria mentioned above. This is achieved if such adaptive computing systems can monitor themselves， analyze their behavior， learn and adapt to several execution environments while keeping the systems complexity invisible to the user.

Prof. Abderazek Ben Abdallah and his team are engaged in research and development of such an autonomic computing. In particular， we develop adaptive computing systems for incremental learning and adaptation in dynamic and harsh environments targeted for a wide range of applications， including mobile robots， automotive， and motion analysis for activity recognition. This includes coordinated tasks on all system layers ranging from core level architecture to compiler and runtime systems all the way to the system level.

3.4 Biomedical engineering

Biomedical Engineering （BME） is the fastest growing multi-discipline across engineering and medicine domains. Biomedical Information Technology Laboratory （BIT Lab） is seeking to utilize the latest ICT achievements， such as internet of everything and data warehouse， data mining and deep learning， computer graphics and visualization， mathematical modeling and computer simulation， ubiquitous devices and embedded systems， to promote health and the whole care cycle from prediction to prognosis.

BIT Lab headed by Prof. Wenxi Chen is enhancing fundamental studies and clinical applications on several key aspects below：

（1）Measurement technology for physiological function monitoring. Diverse instrumentation for detecting biological signals conveniently in daily living and clinical environment and various algorithms will be developed to evaluate biological functions. A ubiquitous infrastructure will be built to track changes in biological functions at any time.

（2）Prevention technology for physiological function decline. Modern adaptive control theory and data mining technology will be applied to investigate biological regulatory systems and to elucidate the self-adaptive and self-healing mechanisms in human beings. Furthermore， next generation diagnostic and therapeutic technology will be developed for the prevention of sudden cardiac death and the improvement of quality of life.

（3）Support technology for declined physiological functions. Biomaterials and electro-mechanics， robotics， network technology， computer graphics & imaging technology， and augmented & virtual reality technology will be utilized to develop integrated systems to help physical and mental functional therapy and rehabilitation.

（4）Information technology for basic and clinical medicine. We are also promoting and strengthening research on bioinformatics， computer modeling and simulation， biostatics， and medical information technology in basic and clinical medicine.

BIT Lab commits to excellence in education and research， and encourages students to explore their professional potential， discover their life passion. BIT Lab enhances social contribution through casting academic outcomes into shape. BIT Lab is also endeavoring to strengthen the ties between our university and external collaborators. We conduct field tests and clinical evaluation through connections with hospitals， nursing homes， rehabilitation centers， medical equipment manufacturers， and healthcare suppliers. We demonstrate our achievements through diversified approaches such as technical forums， academic conferences， patents， industrial exhibitions， and academic exchange with domestic and oversea institutions.

4 International Collaboration with China and the World

The University of Aizu has been collaborating with many universities in the world. The relationship with China has been especially endeavored by its international faculty members. We have implemented the dual degree program at Graduate School level with Harbin Institute of Technologies in Heilongjiang and Huazhong University of Science and Technology in Hubei， and Tamkang University in Taiwan. We also accepted students from Dalian University of Technology in Liaoning for short-term program.

We have exchange students with Rose-Hulman Institute of Technology in USA for 10 years. It is a top university on undergraduate engineering education in US. We have established a triangle alliance with San Jose State University in Silicon Valley， Dalian Neusoft University of Information， to promote ICT innovators education， by making good use of potential environment， since all the three universities are strong at ICT fields and surrounded be many ventures and companies.

5 Cultivation of Global IT Innovators in New Era

Our goal is to train innovators to be active in the world. Our approach is to recruit students who have potential from Asia. And we provide an education curriculum to study the fundamentals of computer science and engineering compatible with ACM/IEEE international standardized

In addition， we provide a series of the courses and factories for venture entrepreneurial experiences and internships， in order for student to have a strong motivation and mindset. Especially， they will be good at gathering information， presentation， coordination， and leadership in multi-cultural environment. Moreover， they will have opportunities to visit the Silicon Valley， where they can experience at the top level of the contests and hackathons etc. in the world.In the future， they will become competitive to find a job anywhere in the world based on their education and experiences.

Acknowledgements

Thanks Prof. M. Kansen and T. Kawaguchi for proof-reading the article and giving revision comments.