DFG project develops first knowledge graph for patent information
30.11.2022 12:38
Patent knowledge has so far been used primarily in an industrial and economic context. However, access to this knowledge remains denied to both researchers in science and society in general. This means that a large part of the opportunities from exploiting patent information to create new knowledge remains unused. The consequences are losses in innovation, quality and impetus for technology transfer. This means losses in international scientific and economic competitiveness.
To investigate the need for patent knowledge and its potential for use in research, FIZ Karlsruhe, in collaboration with the Leibniz Institute for Plasma Research and Technology eV (INP) in Greifswald, carried out a needs analysis in the form of online surveys at several Leibniz institutes, the results of which have reinforced the need to use patent knowledge in research. Building on this, the “Patents4Science” project, funded by the German Research Foundation (DFG), aims to create something new over the next three years: a patent-centric knowledge graph based on linked open data and a modern information infrastructure for networking patent knowledge with scientific literature and others specialist information – this is also a novelty. The Patent Knowledge Graph draws on domain-specific knowledge from the relevant specialist areas of plasma technology, additive manufacturing and battery materials in order to link relevant information in patents with important knowledge and information sources for the respective domains. This should enable researchers to easily access essential information in patents and use it in their research. This includes, among other things, the description of technical processes and equipment, the properties of materials and active pharmaceutical ingredients as well as details of special biomedical processes. As part of their work, researchers gain access to new solutions, experiments or technical specifications that are not or rarely published elsewhere via dedicated semantic information services.
An application example from the Leibniz Institute for Plasma Research and Technology (INP) impressively shows how important easy access to patent information is for targeted research and technological innovation: At the INP, research is being carried out to combat the corona pandemic as part of the extremely socially important research Solutions for plasma-assisted disinfection of indoor air have been worked on. This highly topical challenge was based on previous work on research into the effectiveness of plasma generators for cleaning indoor air and reducing viral loads, which has been published in numerous scientific publications and patents in recent years. This research and linking of information and data required considerable manual effort. The knowledge graph-based approach now being pursued in the Patents4Science project will make it possible to accelerate this process immensely in the future using artificial intelligence and Semantic Web technologies. For this purpose, search queries of the form “Which plasma generators can be used to decontaminate indoor air and produce a physical plasma that reduces SARS-CoV-2 viral load?” are automatically processed and can ultimately be answered in the form of prepared search results. This not only enables research and development processes to be accelerated, but also results in enormous savings in financial resources and a direct economic and intellectual competitive advantage.
In addition to the Leibniz Institute for Plasma Research and Technology eV, project partners of FIZ Karlsruhe are the Leibniz Institute for Materials-Oriented Technologies - IWT in Bremen and the INM - Leibniz Institute for New Materials in Saarbrücken. The IWT is about using patent knowledge to research completely new material combinations in additive manufacturing. In laser-based 3D printing processes, particles with special physical properties such as: B. Permanent magnets with high field strength are used together with alloy particles specially produced in the institute. Patent information is of enormous value for these processes. The rapidly developing field of novel battery materials, energy storage media and processes for recycling them presents both research institutions and industrial companies with the problem of having to orient themselves in a flood of information that is often rich in synonyms. Against this background, it is very important for the INM to be able to differentiate its own research from the patent situation as early as possible. This avoids duplicate research and the risk of affecting patent claims that are already protected. INM expects to achieve a simplification and acceleration of its own research in the defined research environment with the help of the automated information and research tool to be developed. In this context, new aspects should be discovered at an early stage, which will lead to a time advantage when researching new research fields.
The infrastructure developed in the project will be evaluated based on selected use cases from the project partners. This is primarily about using patent knowledge for efficient technology transfer into research on plasmas, 3D printing and new battery materials for next-generation energy storage. Further information can be found on the FIZ Karlsruhe website as well as under www.patents4science.org