DOLORES MODIC, Kyushu University, Japan
Understanding Intellectual Property Data – the How and Why?
A single piece of intellectual property data is usually devoid of context, thus its usability is limited. E.g. for SMEs to be able to extract value from the information about the patent applicant, they need to know whether the applicant is still the owner. Is the patent in use? Is it embedded in their products, their processes? Similarly, university technology transfer office needs to understand how to match their seeds with industry needs. Presently, the field of intellectual property rights and patent informatics is a lively one. IPR data is big data, as its characteristics are high volume, high variety, and high velocity of changes. Consequently, connecting different types of IPR data presents a challenge. Yet, when large amounts of IPR data are connected, a new ecosystem for (open) innovation appears. The first part of this paper is concentrated on new advances in patent (IPR) informatics, focusing on the usability of new data-merging techniques (including disambiguation and allocation techniques), on new IPR data types (such as linked open data, LOD) and finally, promising new knowledge extraction techniques. The second part asks why bother? Discussing how advances can be harnessed for improvement of IPR management as well as better evidence-based policy-making.
DALIBOR FIALA, University of West Bohemia, Czech Republic
Online Sources of Patent and Bibliographic Data: Accessing and Processing Them
The established bibliographic databases have a number of functionalities and can, in principle, serve as search engines of academic papers, citation indices, or calculators of bibliometric indicators. Web of Science, Scopus, ACM Digital Library, DBLP, CiteSeerX, and Google Scholar belong to the best known ones. Similarly, Espacenet, PatentScope, and Google Patents are examples of search engines designed for patent databases. Larger amounts of data from these two types of databases can be successfully used for bibliometric measurements, citation and collaboration networks analysis, for the visualization of the production and quality of scientific research, or for the evaluation of innovation in various fields and countries. These data can be acquired only manually in some cases but also automatically in some others. In this talk, I will give an overview of bibliographic and patent databases and the possibilities of data acquisition from them and will focus in more detail on Web of Science, Scopus, Espacenet, and Google Patents.
MARKUS ABEL, University of Potsdam, Germany
Knowledge Engineering in Software Development at Two Examples in Machine Learning
We present two showcases of knowledge engineering with positive and negative outcome. Both examples reside in the domain of data analytics and/or machine learning. The first example concerns sound reconstruction from measurements and is relevant for music industry. The second example concerns the use of a specific machine learning strategy to control complex systems. We explain the difficulties related to IP rights on software, how they are tackled and what is relevant in a european and american context. The whole presentation will be given from the perspective of a developer and where his main difficulties were found.