UNYCOM | Knowledge & Legal

How much does a flash of inspiration cost?

Software provider UnyCom and Know-Center have jointly developed a method for an automated evaluation of expected property right costs. In the process, machine learning methods are applied, which use known evaluations to evaluate new similar property rights.

The registration of property rights, such as patents, is an essential measure to ensure the utilization potential of intellectual property. However, it is time-consuming and cost-intensive to generate and manage patents. Given the growing volume and complexity of inventions, it poses a major challenge for innovative companies. The economics of any course of action, such as the registration in certain regions or renewing a patent, must be well deliberated.

„Within this project, we intended to gain initial experience with machine learning methods and determine the best method for performing cost predictions. On both counts, Know-Center was the right partner and provided us with excellent support. Thanks to this cooperation, we were able to effectively implement cost prediction and increase prediction accuracy.“

Michael Kohlhuber, Development Manager at Unycom

Innovative algorithms forecast costs

UnyCom GmbH offers specialized software solutions for intellectual property management. Many of the most innovative and largest European companies are using UnyCom in more than 60 countries and manage more than one million IP rights by means of the software. The cost evaluation is based on a set of rules that is not very flexible and is costly to maintain.

Know-Center has been researching machine learning techniques since 2001 and has access to advanced algorithms and methods via its international research and enterprise networks. They were applied during the project to predict three key cost-driving factors in the IPR process:

  • Country decision: predicting expected registration costs in certain countries.
  • Maintenance decision: should an IP right be renewed or abandoned, especially in light of expected costs in the following year.
  • Budgeting: Prediction of the expected budget for an entire IP portfolio in a subsequent year.

„The approach proposed by Know-Center is based on machine learning, which has the advantage that factors that have the greatest impact on costs are automatically considered most. To this end, we did not have to apply any maintenance-intensive calculation rules. At the same time, this approach is transparent and easy to explain; we are certain to have chosen the right path.“

Christian Baar, Business Analyst of Product Development at Unycom

The project was funded within the framework of COMET – Competence Centers for Excellent Technologies by BMK, BMDW, Land Steiermark. The COMET program is managed by the FFG.