The Science of Computation and the Science of Knowledge
-The Department of Information Science -
Professor Takeo Igarashi
(Graduate School of Information Science and Technology
/Department of Information Science)
Computers are able to beat human opponents in quiz games and Igo (a Japanese board game), and even their performance on traditionally difficult tasks such as image and voice recognition, as well as translation, has become equal to that of human beings. It is believed that in the near future, self-driving cars will be widespread and AI may cause people to lose their jobs.
Information science studies the fundamental principles and methods behind these information systems. It is particularly important to note that the Department of Information Science differs from information-related departments in the Department of Engineering as it possesses both an engineering aspect in terms of constructing highly intelligent information systems and a scientific aspect that searches for the essence of information and knowledge. In the following section, I will introduce a selection of some of the many research activities within the Department of Information Science.
As a field in the natural sciences, information science tries to abstract and model the fundamental principles of nature in the form of "operations." A major difference between information science and other natural sciences is that information science not only tries to analytically uncover mechanisms but also uses synthetic methodology to elucidate principles by creating things that "function." As an example of research in basic fundamental theory that is related to information science, I will now introduce research that is related to the study of matroids, which is a branch of discrete mathematics.
At first glance, computations that appear in individual fields of natural science seem completely unrelated but in many cases they have fundamentally similar structures. A prime example of a ubiquitous discrete model behind computational problems in various fields such as physics, chemistry, and economics is a bundle structure called a matriod. When using computers to do calculations, larger volumes of data cause an exponential increase in computation time. On the other hand, in regards to problems with matroid structures, we can develop algorithms that utilize ideal properties of matroids such as obtaining optimal solutions by repeated local searches without having to look at the entire data, or by high-speed computation that converts large amounts of data into small-sized problems. The Imai Laboratory is developing fast algorithms using matriod structures and promoting research on what kind of problems are able to be expressed as matroids.
Modern computers are a hierarchical combination of running complex systems consisting of various technologies. The Department of Information Science is conducting a broad range of research and education on fundamental technologies that support the operation of these kinds of computers, such as research on computer hardware, the base software that runs above that hardware, programming languages and verification technologies, and algorithms for efficient numerical calculations. Below are examples of research related to those kinds of computational systems, which includes research on the base software for the real-time processing of autonomous systems like self-driving cars.
Autonomous systems like self-driving cars need to recognize the surrounding environment, make a decision based on what was recognized, and operate by processing and executing that decision in real time. There are also various processing properties, from those emphasizing processing efficiency to rapid responses. The entire system also consists of various processes and groups that are mutually dependent. In terms of theory, research topics include the execution time analysis of directed graphs or scheduling algorithm, and in terms of practice, parallel distributed data processing and high-performance computing. Hereafter, mobility such as in cars and robots will have rapidly increasing connectivity (i.e., connection to the Internet) and cloud computing and security will also become important research themes. The Kato Laboratory is applying these theories and practices to their research on basic software.
Information systems contribute to our society through operational application systems and the Department of Information Science is conducting a variety of research in that area. Specifically, research topics include bioinformatics, which attempts to utilize information technology to approach the principles behind biological phenomena, natural language processing in regards to languages used by human beings, Human-Computer Interaction (HCI), which examines new ways to use computers and their ease of use, and as mentioned in the introduction, machine-learning and artificial learning technologies. Next I will introduce an example of research on applied technology, specifically shape modeling methods that use high-speed physical simulation.
Physical simulation analysis has long been widely used to design automobiles and airplanes. However, in terms of everyday use, it is used to primarily check the strength and efficiency of a shape after it has been designed. The Igarashi Laboratory is developing methods to implement real-time physical simulations into initial design systems in order to check the shape of a design while determining how it will move. For instance, they are developing methods to design custom-made objects in real-time such as metallophones by listening to the tones that are being made, garments by referring to the silhouette of a piece of clothing that is being worn, and checking the trajectory of an airplane while it is flying.
Information technology works as an infrastructure indispensable for our daily lives. It is also a rapidly developing cutting edge technology that makes impossible things yesterday possible today. Furthermore, information science’s pursuit of fundamental principles is also rapidly expanding and deepening. It is still a young field but it is not uncommon for the technology and methodology developed by one person to be used worldwide and for our view of something to change based on a new principle discovered by a single scientist. The members of the Department of Information Science engage in research and education to open up a new world, keeping such a situation in mind.
▶︎ For more information, please visit the Department of Information Science homepage:
― This is a translation of an article from the "Departmental Overviews in the School of Science" series in The Rigakubu News ―
― Office of Communication ―