The application underlying the present decision relates to a server search service . However, the European Patent Office refused to grant a patent since the distinguishing features would merely relate to linking of cognitive data. Here are the practical takeaways of the decision T 1615/17 (Linking place information based on position/RAKUTEN) dated October 21, 2022 of Technical Board of Appeal 3.5.01:
The present invention relates to a service that allows a user to search for information relating to a facility (e.g., a store or a restaurant) based on geographical location. A server stores information about the facilities in a corresponding database and provides the search functionality to users of mobile terminals.
The users may populate the database (i.e., a user may give a name to a facility at a geographical position). Accordingly, the problem may arise that different users input different information describing the same facility. For example, position information of the facility may differ because the devices of the respective users utilize different positioning systems. As a result, the server may have issues recognizing that the user actually describe the same facility resulting in poor search results (cf. application as filed, para. ).
In order to overcome this problem, the present invention proposes to group information records relating to positions included in a predetermined geographical range and having names that are associated with each other according to a dictionary. Finally, the resulting records are assigned a common identifier in the database allowing to appropriately link the information to the facility/place (cf., application as filed, paras.  and ).
Fig. 5 of EP2650839 A1
Here is how the invention is defined in claim 1 of the main request:
Claim 1 (Main request)
A server (10) comprising:
an input information storage means (14) for storing input information where position information indicating a geographic position, a word given to the position, and a user ID identifying a user having given the word to the position are associated with one another;
a dictionary storage means (15) for storing dictionary data indicating associations between words;
an association means (17) for extracting a plurality of input information where the geographic positions are included in one geographic range and the words are associated with each other by referring to the input information storage means and the dictionary storage means, and associating the extracted plurality of input information with each other by assigning a common identifier to the plurality of input information; and
a registration means (13) for entering the plurality of input information associated with each other by the association means into a result storage means (14b), wherein
the dictionary storage means (15) stores dictionary data generated by associating different words given by different users when the different words are given at a predetermined number or more of common positions, and
the association means (17) determines whether the words indicated by the input information are associated with each other by using the dictionary data.
Is it technical?
The examining division argued that claim 1 merely distinguishes itself from the prior art in the handling of cognitive data. The appellant argued that the subject-matter of claim 1 achieves the technical effect of improving a database being in an inaccurate state (i.e., having multiple uncorrelated records relating to the same facility) resulting in improved search functionality of the server.
Furthermore, the appellant argued that:
The generation of dictionary data and the grouping of information was performed automatically by analyzing position data having a latitude and a longitude. Such data was technical by nature as it had been measured by a position measurement system. According to G 1/19 -Pedestrian simulation, measurements had technical character.
The invention was motivated by technical considerations of the position measurement system. It would have required technical expertise to recognize that a position acquired by means of a GPS function was prone to measurement errors, since the GPS receiver was subject to different noise and/or reception conditions varying among different devices.
However, the Board did not follow the appellant’s argumentation. Relating the linking of the records, the Board stated:
The Board’s view is in line with decision T 309/10 -Archival and retrieval/MULTEX which held that linking data records by means of a common identifier (parent identifier) was an administrative matter that did not contribute to inventive step under the “Comvik approach”. The Board in that case compared the indexing of database records with the work of a librarian (see point 9). The librarian would choose one representative form for all the variants of each entry (e.g. “IBM”, “iBM”, “IbM”, and so on). There would be nothing technical in what the librarian would be doing. He would simply be a good administrator, solving the non-technical problem of storing and locating data (books).
The Board further did not follow the appellant’s argument regarding the technical position data:
In the Board’s view, the position data is cognitive information just like the names given to the positions in claim 1 and data entries in T 309/10. A position is just a point in a geographic coordinate system. Claim 1 does not contain any features relating to how the position information was measured using e.g. GPS, but even if it did, the position information is simply given as an input to the algorithm. The position information does not retain any technical character related to how it was measured in the subsequent data processing.
Moreover, the Board does not consider that the association of information in claim 1 involves any technical considerations of how the position information was measured. Indeed, the same algorithm would apply if the position information was inputted manually by the user or a database administrator. Such manually inputted data would also be “inaccurate” due to human error or the lack of an agreed format. In previous cases of the Boards of Appeal (see for example T 1234/17 – Customization based on physiological data/ADIDAS AG, and T 1798/13 – Forecasting the value of a structured financial product/SWISS REINSURANCE COMPANY LTD), it has been held that it is not enough that an algorithm makes use of a technical quantity in the form of a measured physical parameter for it to be technical. What matters is whether the algorithm reflects any additional technical considerations about the parameter, such as its measurement. As set out above, the Board does not see any such additional technical considerations in the present case.
Finally, the Board stated:
That the improvement of the search results, which potentially leads to fewer search queries, is a mere consequence of the non-technical organisation of data. The only difference between this effect and the non-technical effect of a library indexing scheme is that the claimed invention is implemented on a computer system comprising a server and a number of mobile devices. However, this is not enough to confer technical character to the method of organising data. There has to be a further technical effect going beyond the normal effect of implementing something on a computer system. The Board does not see any such further technical effect in this case, because the invention does not improve the functioning of the server in a technical sense.
Accordingly, the Board did not see a further technical effect or motivation based on technical considerations
As a result, the Board stated that claim 1 of the main request does not provide any technical effect which goes beyond the straightforward implementation of an algorithm.
Therefore, the Board dismissed the appeal due to lack of inventive step.
You can read the whole decision here: T 1615/17 (Linking place information based on position/RAKUTEN)