Machine Learning Technology Probabilistic Generative Model. Time Series Data Analysis Technology Stream Data Control Technology Artificial Intelligence Technology IOT&Sensor Technology Semantic Web Technology Digital Transformation Technology
Most of the current IOT services are for specific domains, and the volume of service recipients is small compared to the volume of initial investment, making it difficult to establish a business. In order to increase the volume of service recipients, a platform that can build flexible services that meet individual needs is necessary.
Semantic Web (SW) technology considers the entire Web as one huge information DB, and is a technology to efficiently process the huge amount of information that exists there through automatic software processing. SW service technology is a technology to automatically build services that meet individual needs by modularizing web services.
By expanding the scope of SW technology to include IOT information, flexible services based on various information sources can be built, and one of the aforementioned barriers to the commercialization of IOT can be broken down.Data stream technology with semantic interpretation has been developed for the extension of SW to IOT technology.
A concrete example of the fusion of IOT and SW technology is the Smart city project. The pioneer of the smart city project is the project in Dublin, Ireland. The pioneer of the smart city project is Dublin, Ireland, the birthplace of SW technology, where various SW projects such as LOD have been implemented, and the Dublin smart city project is one of them. GAFA (Google, Apple, Facebook, Amazon), companies such as IBM and softbank, and public research institutes such as CNRS and INRIA are also participating in this project, making it one of the best of its kind in the world.
As for the application of SW technology to smart cities, in the tutorial “AI for Smarter Cities. Hype or reality? An application to diagnose and reason about traffic conditions in a city is reported in the tutorial “AI for Smarter Cities. In this application, information from social media such as twitter, road information managed by public agencies, bus operation information provided by bus companies, and other information in various data formats are collected as a data network using SW technology, and then processed using predicate logic and machine learning such as Markov Logic Network. The aforementioned service is realized by using a method imitating a combination of predicate logic and machine learning, such as Markov Logic Network, to perform inference (described as subsumption-based reasoning in the report).
The smart building project is a smaller scale project that targets buildings rather than cities. This project is also being implemented in various places in Japan and abroad, but the most famous one is probably the ADREAM project. In this project, more than 6500 sensors are placed in the building, and they are connected by SW technology to realize autonomic control (mainly energy saving) by building knowledge data called ontology.
As an extension of this project, a human support agent system with the addition of robots is being considered. Such a combination with robots would be one of the ways out for the combination of SW technology and IOT. One of the most famous research projects on the integration of robots and humans is the Cybernetics Cluster project at Aachen University of Technology. Here, they are trying to create a world where humans and machines coexist by combining various AI and robot technologies.
In addition, ontology technology, which is one of the SW technologies, has various applications in the domain of knowledge utilization, such as improving the efficiency of information use within a enterprise, failure analysis, and risk management technology, etc. By combining them, the above technologies can be expected to have even more significant effects.
Machine Learning Techniques applied to these include submodular optimization, which is the optimization of discrete data, compressed sensing using sparse modeling, and various inference techniques.
コメント