On March 1, 2021, the European Patent Office (EPO) updated its Guidelines and the new update includes a substantial new section on how to patent database management systems. Given the importance of data and database protection for, e.g., artificial intelligence, this update could be significant in the next few years. The Guidelines are clear in this and other sections that there is a distinction to be made between improvements in accessing the data, such as improving the internal functioning of a computer and optimizing the search using the computer’s technical means, and improvements in search results that merely look at the information content of the data to produce a “better search”.
The EPO considers that database management systems are “technical systems implemented on computers to perform the technical tasks of storing and retrieving data using various data structures for efficient management of data”. There is nothing objectionable about this statement and is certainly in agreement with the common understanding of a technical method. The Guidelines emphasize that a method performed in a database management system is potentially patentable. The method uses
technical means (the computer) and is therefore not excluded from patentability under the provisions of Art. 52(2) and Art. 52(3).
However, this clarification does not make potentially patentable all types of database management systems and methods in a database management system (assuming that they are novel and inventive over the prior art). Granted patents will require technical features compared to the prior art which are considered to be inventive. These novel features need to be clearly explained in the patent description and included in the claims.
For example, features specifying the internal functioning of a database management system are normally based on technical considerations and they contribute to the technical character of the invention. Such features – if described in the description and set out in the claims – are taken into
account for the assessment of the inventive step.
The improvement of system throughput and query response times by automatically managing data using various data stores with different technical properties such as different levels of consistency or performance have been considered to justify the grant of a patent (see decisions, T 1924/17, T 697/17).
Another example would be the optimization of the structure of queries with respect to the computer resources needed (such as CPU, main memory, or hard disk). If the optimization improves performance, then it will contribute to the technical performance of the database management system and thus be patentable – assuming again that this optimization method is explained in the description and set out in the claims.
The Guidelines emphasize that not all features implemented in a database management system necessarily make a technical contribution by virtue of their implementation alone. A distinction needs to be made between technical implementation and providing “cognitive information”. A novel data structure, such as an index, hash table or a query tree, used in
database management systems to facilitate (technical) access to data or for improving the execution of queries is technical. Such novel data structures are functional since they purposively control the operation of the database management system to perform said technical tasks. Conversely, data structures defined solely by the cognitive information that is stored in the data structure would not be considered to contribute to the technical character of the invention beyond the mere storage of data.
In other words, a distinction must be made between executing structured queries and information retrieval. Information retrieval includes searching for information in a document, searching for documents themselves, and also searching for metadata that describes data such as texts, images, or sounds in the document.
The translation of linguistic considerations for information retrieval into a mathematical model with the aim of enabling the linguistic analysis to be done automatically by of a document by a computer can be seen as involving, at least implicitly, technical considerations. However, this is not enough to guarantee the technical character of the mathematical model. Further technical considerations such as those relating to the internal functioning of the computer system are needed.
A mathematical model for calculating the probability that a given term is similar in meaning to another term by analyzing the co-occurrence frequency of the two terms in a collection of documents does
not make a technical contribution per se since it is based on considerations
of a purely linguistic nature (i.e. based on the assumption that terms which
are related are more likely than unrelated terms to occur in the same
documents). The search results produced using this method of similarity
calculation would differ from the prior art that adopts another mathematical model only in that information with different cognitive content would be retrieved. This is a non-technical distinction and does not qualify as a technical effect, as was decided in a decision T 598/14. The concept of a “better search” is subjective. However, if the mathematical model was able to additionally optimize the execution time of structured queries in the retrieval system then it would be potentially patentable, as discussed above.
These Guidelines are not intended to be comprehensive and leave open many questions that will need to be answered in the next few years. However, the EPO is attempting to provide support to patent applicants and their attorneys on how to draft patents on database management systems that will lead to the grant of the patents.
If you have any further queries, then feel free to get in touch.