For query/comment, please send a message to [office  (a　t)  compview.　titech.ac.jp]

Our objectives

• Establish new scientific methodologies focused on computation
• Train scientists versed in the methods based on a computation-centric point of view (=> Educational Program)

What is Computationism?

If we consider everything that can be represented in a computer as "computation", then it is possible to view everything that can be analyzed by humans as computation. "Computationism" is a framework which seeks to rethink various scientific phenomena from the point of view of computation.

Each field of science embraces a particular "point of view" which can be considered the defining philosophy of that scientific discipline.

For example, physicists view phenomena from a "physical point of view”, and biological scientists view phenomena from a "biological point of view". A point of view is a new point of view focused on computation, which differs from these traditional points of view. The goal of computationism is to develop new methods of scientific inquiry based on this new perspective.
 

What are the advantages of focusing on computation?

• Discovery of new concepts and relationships (see Example 1).

• New approaches to modeling in particular, the formulation of computationally efficient models (see Example 2).
• Previously unforeseen applications of existing algorithms, or to the development of revolutionary new algorithms (see Example 3).

Towards the establishment of new scientific methods based on a computational point of view

Our GCOE plans to develop and implement fundamental theories and technologies in order to establish computation as a new, viable framework for scientific modeling and analysis.

• Computational analysis, advanced algorithms:
  research on randomness, theory of optimization algorithms.
• New approaches efficient modeling, data analysis, and visualization techniques.
• Foundations for the Next-Generation of HPC Technologies: Post-Peta -HPC technology using accelerator boards, analysis of grid computing.

While our GCOE will pursue basic research to establish computation as a foundation for science, it is actually future generations of scientists who will apply these new methods and continue to develop the field. Thus, we will focus much of our effort on an educational program whose goal is to train computational scientists with the background necessary to pursue research in this new field. We seek to train doctoral students with skills in both high-performance computing and mathematical modeling, and will establish a special educational program for doctoral students which consists of the following:

• Coursework: Lectures and seminars on theoretical and practical techniques which will serve as a foundation for computational science.
1. Fundamentals of algorithms and modeling
2. Techniques for efficient use of Grid-based supercomputing resources (e.g., TSUBAME); Fundamentals of high-performance computing
3. Problem-solving techniques in selected, specific application areas
• Research Rotation System: In addition to performing research in their own advisor's lab, students will be assigned for a period of 2 months to another lab participating in the GCOE. The student will perform collaborative research directed by a faculty member other than the student's advisor. This will also be an opportunity to get acquainted with researchers from another field. We will make a particular effort to send students to the labs of our international collaborators.
• Mathematical Science Forum: An opportunity for doctoral students affiliated with our GCOE to present their results. External researchers (in particular, technologists and researchers from industry) will be invited as commentators to evaluate the presentations.
• Internship: We will strongly support efforts by our doctral students (including pos doc.s) to present their work to industry. For this we are planning some internship, i.e., temporary assignment of doctoral students to industry or other research institutions.

It will be challenging for doctoral students to participate in these programs while also making progress on their own research. However, we believe that the opportunity to apply their research to real, scientific problems and explore research in related fields will broaden their research interests, and will prove to be valuable in their future careers.

In order to introduce the Computationism to a wide audience and recruit students with the talent to enter this field, we will hold a two-day seminar (Introduction to Computationism). This seminar is intended for students considering advancement to doctoral studies, and it will be held several times each year.