Science meets Industry

Traditionally, collaboration with researchers from the academic environment has been close in nearly all areas of industrial development. Ever since the global economy has been preparing for the transition to a knowledge economy, there has been a need not only for academic cooperation in product development but also for cooperation in information and knowledge management, involving all industrial sectors and product categories. This situation is accentuated by the fact that successful application of information science requires relevant knowledge on the respective domain.

For this reason, one of the first measures of the IRF is to establish a lasting dialogue between industry and information science in the form of cross-domain knowledge transfer.

The needs of industry

Product development, in particular, produces enormous amounts of so-called unstructured data (mainly text but also images, speech, video etc.). In addition to documentation, research results, DNA sequences, structural designs or patent results, enormous volumes of new data are created on a daily basis. Most of these data are not available beyond their production date as they are literally lost in the company-wide network in no time.

This phenomenon is particularly problematic if an organisation depends not only on its productivity but if in addition, its net worth is largely dependent on the intangible asset of its in-house knowledge base.

Patent literature constitutes a key element of these documents, for the following reasons:

• Patents are an intangible asset per se. Patents are registered, published, traded and licensed. Therefore it is extremely important for industrial enterprises of whatever size to master all disciplines related to patent documentation.
• Patent specifications are highly sophisticated and complex documents that feature and combine a wealth of scientific, economic and legal knowledge.
• Patents protect products whose development has been immensely cost-intensive. To give an example, developing a new drug until its approval may cost up to USD 1.7 billion.
• The current patent pool comprises approximately 60 million documents with a gross volume of approx. 25 terabyte worldwide. This corresponds to five million times the collected works of Shakespeare, or two and a half times the entire printed collection of the US Library of Congress.
• Faced with these data volumes, human processing is becoming increasingly complex, costly and, last but not least, time consuming. Due to the constant increase in new patent submissions, the industry is coming under growing pressure and looks for ways of easing that pressure by relying on information technologies.

The needs of information science

Like other research fields, information retrieval has special requirements that facilitate its scientific work. For example, to review results, it is necessary to characterise the user and the user’s information needs. A query that triggers a given result may be excellent for one user yet useless for another.
Interestingly, opinions largely concur on whether a patent search triggers right or wrong results. For that reason, an evaluation of IR techniques by patent information professionals is one of the most precise known methods of measuring quality, efficiency and effectiveness. Moreover, considering the fact that patents are the most sophisticated documents altogether given their average size, number and text quality (approx. one billion words per three million documents), they are evidently best suited as a basis for building a reference system.
This reference collection will not only serve ongoing research but will also help to better measure existing IR algorithms and methods with greater precision.

IRF facilitates dialogue and cooperation

To build a scientific reference system at a quality unprecedented within this domain, it is necessary to engage in inter-disciplinary dialogue and consistent, bi-directional knowledge transfer. The IRF considers one of its key tasks the facilitation and sustainable preservation of such a dialogue. For this reason, the IRF organises an annual symposium inviting approximately one hundred experts from both domains (IP and IR) for an exchange of ideas on specific problems and related solutions. This event triggers a number of inter-disciplinary working groups focusing on a variety of issues. The scientific results of these working groups and all other projects/experiments submitted during the year are presented on the occasion of the scientific IRF Conference, which is the second annual event.