春雨直播app

Business and governmental organizations need to make strategic decisions swiftly but judiciously. To this end, they rely on findings derived from various data sources around the world to guide their decision-making processes. For example, if an organization is considering how to use an innovative form of technology, or what science/technology strategy it should adopt, it will need to gather information about the fields of research that are attracting the scientific community’s attention, including who the key actors (such as research institutes) are. It will also need to verify what stage its rivals are at in relation to the technology it is developing, and it needs to ascertain any complementary technologies. However, much of this information ultimately goes unused, even if it is useful and accessible. One reason is that decision makers become overwhelmed by the sheer volume of information. A case in point is nanocarbon (a type of material that has attracted attention for its potential to solve various social problems); while there were fewer than ten thousand articles on nanocarbon in around 2000, there are now over forty thousand. Business and governmental organizations have so far followed an expert-based approach, in which they rely on experts to analyze the data and derive the key findings. The Delphi method is a typical example of this approach. However, the problem with this approach is that the volume of data becomes too vast for humans to process fully.
 
This book proposes a solution to the problem: knowledge structuring. Knowledge structuring uses a computer-based approach to expedite the decision-making process. The method exploits the citations found in articles and patents. As citations represent the purpose of citing the work, they provide in-depth information about the connections between the content. Knowledge structuring links these bottom-up indicators to create a citation network in which the articles and patents are nodes. This network is then analyzed in two steps. First, the clustering method is used to form clusters of articles and patents. Experts can then use the automatically extracted data to define the key information in each cluster. As a result of this step, what once consisted of hundreds of thousands of papers is reduced to only thousands. The second step involves analyzing the content of the papers in each cluster using Natural Language Processing and similar methods. One can then consult the automatically generated information to pinpoint the key researchers or research organizations in each cluster (each representing a field of knowledge), including which fields of knowledge are growing and which are on the decline. Finally, the heat map method is used to determine the strength of the connections between the fields of knowledge (as voluntarily indicated by the authors concerned). To take an academically topical subject as an example, suppose an organization has developed a robotic technology and it wants to find out the potential applications of this technology in Japan’s super-aging society; this organization could use a heat map to ascertain the relevance of its robotic technology for each of the particular issues related to the super-aging society phenomenon.
 
The authors have now established the Academic Landscape System. The Academic Landscape System is an online system that combines knowledge structuring with a model that uses machine learning to predict which papers will win acclaim in the future. The system is currently in use at the University of Tokyo. With broader use, it should become an exemplary example of how to combine a computer-based approach with expert-based approaches.
 

(Written by Ichiro Sakata, Professor, School of Engineering / 2018)