In the Principles of Informatics Research Division we seek to discover new principles, theories and methods in informatics, and extend our goal to pioneering the frontiers to try and achieve a paradigm shift in informatics.
Getting ordinary PCs to solve huge scale problems UNO Takeaki
A computer program has to have a good blueprint for how to proceed huge scale operations. Such a program has to be coded with well-designed algorithms. My goal is to get ordinary PCs to solve huge scale problems that only high-performance computers can solve right now by developing new algorithms. There are two basic targets in algorithm research: theory and practical application. In my research, I design the most efficient ways for solving real problems while trying to fully understand the theoretical research results.
My main research interest is discrete mathematics—more precisely, graph theory and theoretical computer science. The abstract graphs we study consist of lines and points. Although this structure might sound like it has nothing to do with the real world, it's actually used all around us. For example, the idea in graph theory is successfully applied to the frequency band assignments for mobile phones and the car navigation systems. Unfortunately, not many people study discrete math for practical applications, but I'm sure that our research helps develop network technology and some related problems in the real world.
For equitable human intellect sharing in an information-oriented society SADAKANE Kunihiko
In the current information-oriented society, efficiently processing massive volumes of data that grow daily is a key topic. Specifically, we need to develop technologies that can compress large volumes of data to make data burdens as light and manageable as possible, while also enabling high-speed search. Achieving both goals isn't easy. Each day my research focuses on the issue of balancing high compression rates with search speed.
Using a mathematician's eye to read the structure of algorithms HAYAMI Ken
Numerically solving a problem via a computer requires algorithms to show the calculation procedure. My research activity is focused on mathematics and clear elucidation of the structure and characteristics of algorithms. More specifically, I am working on such questions as: "Does the algorithm derive correct answers?" "Will it result in failure halfway?" "How quick does it derive an answer?" and "Isn't a better algorithm available?"
Understanding Natural Language Through Formal Language Theory KANAZAWA Makoto
Formal language theory studies mathematical properties of sets of strings of symbols. As dry as it may sound, it sometimes brings surprising insights about the nature of human language. The recent solution to a longstanding open problem about a very simple formal language called "MIX" may help explain why nobody speaks a language with completely free word order.
Finding the truth about type theory and reflecting it in a program TATSUTA Makoto
In many programming languages, data and functions have many different "types." Abstracting the theory of these types and analyzing them with the use of "type theory" of mathematical logic enables automatic deducing of the types. This also helps prevent bugs from occurring due to the use of a wrong type (failed deductions cause errors). I study type theory and primarily aim to deepen the understanding of mathematical truths that lie in programming languages and calculations. Uncovering the truth of type theory will facilitate more practical forms of contribution such as design of programming languages based on type theory and the absolute prevention of errors in programming.
In the 1980s physicist Richard Feynman formulated quantum computation, which is based on quantum mechanics and designed to facilitate massive parallelism by innumerably combining the two different numerals conventionally used in computers: 0 and 1. Quantum computers, the dream computers of the next generation, are capable of drastically improving calculation ability. Their realization requires overcoming a major technical hurdle. Concerning the procedures for quantum-mechanically simulating the state of a substance's electrons, I therefore take an approach of combining theoretical discussions on algorithms with experiments on physicality.
Using new ideas to explore possibilities of quantum computers NEMOTO Kae
Throughout the 20th century, quantum mechanics reshaped physics from what it was in the days of Newton. The quantum world has been increasingly clarified in the current century and led to significant growth of quantum informatics to utilize electrons in blazing new trails in information processing, communications, etc. Research in quantum computers is more difficult and requires longer-term commitments than for other issues in quantum informatics. Quantum computers are thought to have the potential to bring together all the world's currently existing computers in a single chip as small as a fingertip. The question is how to realize them. From the standpoint of theoretical physics, my research activities are focused on the formulation of theories.
Fusion between quantum and information begins at a deeper level MATSUMOTO Keiji
I research many different aspects of quantum information. If I were to sum up my research in a word, it would be "entanglement." What I find interesting about the quantum information arena versus other fields is the fusion of different concepts at a deeper level. For example, introducing the theoretical structure of telecommunications to physics developed the theory of quantum entanglement. A closer looks at this finds a degree of adaptation of its basic concept to physics. Putting the adapted part of the concept to the original theory of telecommunications introduces new insight into problems of regular telecommunications. Through the fusion between quantum and information or between physics and informatics at a deeper level, the creation of a new concept is about to begin.
For many years I have been involved in quantum optics. This research area emerged in the 1950s, aiming at elucidating the quantum mechanical properties of light through theory and experimentation. The current central theme in quantum optics is optical quantum information processing. In layman's terms, this is the realization of quantum computers using light.
Applying Bose-Einstein condensation to quantum information BYRNES Tim
I have become interested in the phenomenon of Bose-Einstein condensation (BEC). Various different methods for building quantum computers have been proposed and are currently being developed, although it is still unclear at this stage what a quantum computer will look like. BEC offers unique advantages in being able to handle large numbers of particles simultaneously in a truly quantum mechanical manner. By constructing theory that is directly applicable to experiment, it is my hope that they can be more useful to wider society, such as quantum-based devices. Quantum information as typified by quantum computers is a vibrant new field. The significant impact it is likely to have on society makes it an exceedingly challenging area.
Biological morphology exists in many different forms. As my interest grew in how these forms have been shaped, I came to research the correlations between genes and morphology by using the informatics method. The low-profile job of comparing genetic arrangements can lead to the availability of the genes through identification of roles merely from arrangement data.
Chemoinformatics - Towards Making a Guide to The Chemical Reactions' Complex World SATOH Hiroko
Chemical reactions are fundamental phenomena in nature. Much variety compounds have been generated by natural resources and synthesized artificially. Prediction of the reaction products from reactants and reaction condition is one of my research themes. Solving the chemical reaction prediction problem entails finding a solution from a huge number of possibilities of reactions that occur as a result of complicated interactions between several factors concerning structural and electronic properties of reactants, reagents, catalysts, and solvents, and conditions such as temperature, density, pressure, and reaction time. To find a solution, the space with broad diversity to be searched ought to be reduced. Hence, chemists must seek ways to reach the solution within an acceptable time by reducing the space in a rational way. To have success in predicting them, a rational transdisciplinary approach will be necessary. Chemoinformatics is a discipline that can act as guide to the complex world of chemical reactions.
Focusing on genomes and exploring new academic areas FUJIYAMA Asao
Genome science no longer focuses on structure determination. It has shifted to identification of biological significance written in genomes. Apart from the continuation of rapid increase of data on genome structures, databases for describing interactions and relationships between genomes are also increasing rapidly. Considering these facts, informatics is already imperative for the study of life science, and the fusion between these topics is rapidly progressing. However, it is not easy to facilitate communication among different fields of study. My current commitment aims at serving as a bridge between life science and informatics and between life science/informatics and other related fields or even the general public.
Humans learn quite naturally, and their learning activities can be categorized as follows: acquiring knowledge from a phenomenon resulting from an action; summarizing various types of knowledge; and using knowledge to improve the outcome of an activity. Computers have a hard time emulating such behavior. I'm working to develop an artificial intelligence technology that lets computers learn. Providing computers with the ability to learn is the goal of my research and its most interesting aspect.
Robots that "grow" through the experiences of touching, seeing and feeling INAMURA Tetsunari
"Can you hand me the thick book I always use?" In response, the robot says, "No problem. Is this the one you want?" and brings the book. That's the goal of my research. How can a robot gain the flexibility needed for self-assessment of circumstances? I think "embodiment" and "interaction" are critical issues. I'm trying to administer various experiences to robots—for example, teaching robots to move around inside an unfamiliar environment by avoiding obstacles or separating waste by type. It will be a while before you have a robot working next to you in ordinary life, but attempts in this direction are underway.
The study of artificial intelligence has achieved many different human abilities on computers. Human beings themselves have some abilities that remain insufficiently realized. These concern discovery and invention; creative acts of making new theories that no one else has known. I believe computers can do this by leveraging our accumulated information and knowledge and being embedded with parts equivalent to human beings' ability to make discoveries.
In a real environment, various sounds are present and we usually here mixture of them. For example, even if you try to use the speech recognition function of PC, not only your voice but sound from TV close to you may be inputed togather. Even if you try to record the piano performance of your daughter at a concert, noisy sneezing of a man next to your seat may be recorded together. Aiming to recognize only target signal from mixutre sound, and edit or modify it as you like, we have developed a technique to fast separate mixutre sound into each of source with multiple microphones.
Suppose a serious public health crisis occurs in Japan, or somewhere else in the world. What's the most effective way to minimize the risk (risk management) of problems caused by the crisis? The first course of action is to identify the crisis and take all possible response measures. To do this successfully, it's important to obtain accurate information on the crisis as quickly as possible. Several current systems scan public health information on the Web, but they all have drawbacks—limited language compatibility or lack of ease of use. The basic technologies for solving these problems are being developed in BioCaster.
Robots have been considered a dream technology. Today, however, we see the potential for robots all around us. For example, robots may take the place of humans in rescuing lives in disaster-hit areas and other hazardous places. Robot may also support us in many everyday life situations or serve as skilled secretaries to support our business with remarkable information processing ability. Realization of these functions requires appropriate action based on proper judgment of situations. One technology necessary for that purpose is artificial intelligence, which I have worked on for many years.
The study of artificial intelligence is divided into two different objectives: using computers to produce intelligence and making computers for assisting human intelligence. My research is based on the latter objective. My goal is to offer a form of web usage that provides tips for creative activities of humans. As I proceed with the study, I grow more interested in the web as an opportunity for dynamic communication, rather than as an aggregator of knowledge. I am now focused on how to make it work in human activity. I still have a long way to go to achieving my goal of realizing a form of web usage that provides tips for humans' creative activities. I believe the semantic web will be the first step toward that goal.
Specializing in network technologies, I aim to realize services that predict and satisfy people's wishes by taking an engineering approach for combining sensors, computers, networks, radio and other technologies. This is ubiquitous computing in the true sense of the term. Besides the engineering-related aspects, I will continue exploring service-related aspects of how computer networks will be used, in pursuit of realization of services available to all people regardless of age or gender.