Summary of RWC Project Achievements During FY1997

Junichi Shimada Managing Director, Real World Computing Partnership Director, RWCP Research Institute

I will talk about how the RWC Project has been carried out in terms of numbers and from a macroscopic perspective. First, Table 1 shows the members of this Partnership who conducted research and development in 1997.

Among them, the System Science Laboratory of the National University of Singapore (abbreviated as "ISS"), was reorganized and renamed as the Kent Ridge Digital Laboratory (abbreviated as "KRDL"), starting in 1998. And the Japan Iron and Steel Federation withdrew from the Partnership at the end of fiscal 1997, but instead, Sumitomo Metal Industries joined us in 1998.

The overall scope has been summarized by Professor Tanaka, so I will explain what themes map to what fields (Table 2).

There are two fields: Real World Intelligence and Parallel and Distributed Computing. Real World Intelligence includes such domains as Multimodal Function, Autonomous Learning, Self-Organizing Information Base Function, Theory and Algorithm Bases, and Real World Adaptive Devices. Parallel and Distributed Computing includes such domains as Multiprocessor Computing Technology, Seamless Parallel and Distributed Computing Technology, and Parallel Application Technology. This table shows what themes are assigned to each of them. Currently, 31 laboratories belong to Real World Intelligence field and 23 laboratories to Parallel and Distributed Computing field, making a total of 54 in the RWCP.

This Partnership subcontracts researches to mainly domestic and overseas universities to complement the entire research. The number of subcontractors are: 20 for Real World Intelligence and 10 for Parallel and Distributed Computing as shown in Table 2. In summary, Real World Intelligence has 51 themes and Parallel and Distributed Computing has 33 themes, giving a total of 84 themes.

Table 3 shows the research achievements of each of these themes. The top row indicates years since 1992 and the left-most column indicates technology fields and domains. The upper table shows the number of patent applications during those years. The number of patent applications during 1997, when the second phase of this project started, is 36, indicating a modest start after the first phase. We are quite proud of these numbers, which are relatively high in this field where the development of programs is the main task. The lower table shows the number of publications since 1992, which are classified in the left-most column as presentations at academic societies and other forms of publications. The first phase had 1,300 in total, while during 1997, the first year in the second phase, 338 papers were published.

Next, I'd like to present some of the academic awards that we have received (Table 4). Not all the awards are covered but most are listed here. As you can see, we are proud of not only the quantity but the quality of our achievements.

The final goal of the Project is the promotion of industry. So far I have introduced the achievements in terms of researches. Now, let's look at how those achievements are penetrating society.

Some of them were already mentioned by Mr. Okada from MITI and listed at the top is MMPC (MultiModal Personal Computers), PCs that understand combinations of speeches and gestures and which have been installed in some companies already. There have been user groups set up at Distributed Laboratories. In addition, we've heard that one domestic company is considering installing them in their mobile devices. Also we plan to give a demonstration of MMPC at AAAI, one of the biggest international conferences on artificial intelligence that will be held in the USA this summer. The second is a self-organizing technique called Galaxy. This is taking shape and currently is available for trial runs at the RWC's website.

Next comes PC Clusters, which are PC-based parallel computers. These are a milestone on the road to seamless computing systems, and its intermediate results were demonstrated at SC97 (International conference on SuperComputing & Network '97) last year, and attracted quite a lot of attention. If you like, you can check this by looking at the video presentation in front of the reception desk. As Mr. Okada mentioned, this system is now being used at the Los Alamos Research Center. The table also shows that the PArallel Protein Information Analysis called PAPIA may enter practical use. This has been developed as a type of "killer application" for parallel and distributed systems, and is attracting great interest from the biochemical industry and is now under a technology transfer. You can use the PAPIA on RWCP's website. Please try it out if you're interested.

I have briefly introduced an overview of our researches to complement the presentations of Mr. Okada and Professor Tanaka.