HTML5 Webook

28/64

26Open InnovationAI Science Research and Development Promotion CenterAs a center for the promotion of eorts to create open innovation in the field of artificial intelligence (AI), as men-tioned in the NICT’s 4th Medium- to Long-Term Plan, and to implement the government’s plan for economic growth based on the R&D and commercialization of AI technology, we need a central organization that can promote external links. Based on these trends, the AI Science Research and Development Promotion Center (AIS) was founded in April 2017. Since then, we have also been working to promote the development of systems and raise awareness (including the creation of an AIS logo and PR activities via the NICT public relations magazine).At AIS, we aim to set up an R&D environment that is accessible to industry, academia, and government for the use of data collected by NICT, and to provide a one-stop service with a view to social implementation (Fig.1).Our priority R&D tasks for FY2017 are as follows:• Constructing an AI data testbedWe have started building an advanced AI testbed that allows various forms of AI data collected in NICT’s previous R&D efforts to be used on a nationwide scale via the Japan Gigabit Network (JGN) in order to acceler-ate the pace of AI-related R&D and demon-stration trials.• Promoting open innovation research projectsWe are promoting open innovation re-search projects by strengthening our AI-re-lated research ties with industry, academia, and government agencies.Here are the main topics of activities at AIS:Constructing and maintaining an AI data testbedOur AI data testbed is a platform aimed at the development of new AI technology and innovative creations based on this technology by applying and verifying the latest machine learning technology and the like to various forms of data collected and stored during previous R&D work at NICT. In FY2017, in collaboration with the ICT Testbed Research and Development Promotion Center, we set up systems in-cluding a large-scale storage system for AI data, and a GPGPU processing server for large-scale machine learning.Maintaining AI-related dataIn addition to organizing basic ideas for the sharing of data with third parties and rules for the use of data (Fig.2), we are also compiling lists of AI data and related applications that are available for sharing (language resources, • Prohibited uses(a) Uses that contravene any laws, regulations or ordinances, or are contrary to public order and morals(b) Uses that pose a threat to national security, or the security of individual citizens(c) Uses that place undue load on the Web server• Warning re infringement of third party rightsWhen data is covered by copyrights and other rights of third parties other than NICT, it is the user’s responsibility to ensure that these third parties have given their permission for use of the data unless it is clear that these rights have already been processed.Also, content acquired by, for example, cooperating with the APIs of external databases or the like should be handled according to the provider’s terms of use. • DisclaimersDisclaimers regarding all acts performed using the data (including the use of information such as editing or processing data), Public data not guaranteed to be complete, accurate, comprehensive or fit for any particular purpose, Data liable to be changed, moved, deleted etc. without prior notice.• Source descriptionState the source when using data, and include additional notes to show when the data has been edited, augmented, etc.• Individual terms of useParts of the data may be subject to additional individual constraints (fees, physi-cal/organizational access limitations, corporate status of user, usage methods, etc.), which must be complied with when specified.• Governing law and jurisdiction by agreementThese rules of use are to be interpreted based on Japanese law. The use of data according to these rules and any disputes concerning these rules shall be dealt with in the first instance by the exclusive jurisdiction of the Tokyo District Court.• Other considerationsFig.2 : Overview of the AI data sharing rulesFig.1 : Logo of the AI Science Research and Development Promotion Center (AIS)Research and Development