The application underlying this decision relates to controlling a recommender system configured to provide up-to-date predictions of user preferences for products within a large set, for example within a Video on Demand (VOD) catalogue. In view of, the European Patent Office controlling the type and amount of data used for training the recommender system in the claimed manner achieves a technical effect. Thus, a European Patent was granted. Here are the practical takeaways of the decision T 0183/21 (Controlling the performance of a recommender system/BRITISH TELECOMMUNICATIONS) of September 29, 2023 of Technical Board of Appeal 3.5.07:
Key takeaways
Recommending products is generally recognised as having non-technical character.
Optimizing resource usage (e.g., with respect to bandwidth consumption or computational effort) for training a recommender system is considered technical.
The invention
The application underlying the present decision mainly concerns the controlling of a recommender system configured to provide up-to-date predictions of user preferences for products within a large set, for example within a Video on Demand (VOD) catalogue. According to the description, providing data from a huge number of users to retrain a recommender system presents challenges in that it takes up system resources. For example, transferring user preference information from a large number of client devices takes up bandwidth within the communications network connecting the client devices to the recommender system. Moreover, training activities have a high computational cost for the recommender system.
Thus, the need to retrain must be balanced against the quality of recommendations being provided. The type and amount of training data provided to the recommender system is thus controlled in order to drive the performance of the recommender system towards a desired performance value (or to be maintained within a desired performance range). The desired performance value may not be the optimum level achievable, but may be a level available to the majority (if not all) of the users of the services being offered.
Fig. 1 of WO 2013128154 A1
Here is how the invention is defined in claim 1 of the main request:
-
Claim 1 (Main Request)
Is it technical?
Both the Board in charge and the Appellant agreed that the closest prior art document D1 failed to disclose features 3-8. D1 discloses a recommender system in a communication system including a client device associated with a user to which the recommendations are provided.
At first, the Board stated that recommending products is not generally recognised as having technical character (see T 1869/08, Reasons 2.6 to 2.10, and T 0306/10, Reasons 5.2). The Applicant agreed and responded that the purpose of the present invention was not about recommending products but rather to limit the amount of resources used.
The applicant argued that the above-identified distinguishing features provide the technical effect of minimizing the use of network bandwidth required to provide the training data to the recommender system. Same applies to the amount of storage necessary for storing said training data. The amount of training data is indirectly limited via the tendency/convergence of the measured performance metric towards (or oscillation around) the predetermined level of recommendation performance yref which is not necessarily the maximum achievable level of recommendation performance.
The Board agreed and thus defined the objective technical problem as reducing the use of network bandwidth and the amount of sotrage in a communication system including a client device and a recommender system in communication with the client device.
According to the Board, D1 discloses that the data processing algorithms of the invention are re-trained to reduce the differences between actual feedback and earlier predictions with the sole purpose of achieving a maximum performance metric of the recommender system.
Therefore, the skilled person would not use a “reference performance metric” which might be different from a “maximum achievable performance metric”. Moreover, even if the skilled person were to use such a “reference performance metric”, they would not be able, without exercising inventive skills, to derive the amount of training data and also the specific training data per se (i.e. implicitly the nature of the training data or the chosen training data), from a “positive correlation” with the measured performance metric (yielding either an increasing or decreasing amount of training data) and usage data respectively.
Instead, the Board stated that the skilled person would at best incrementally increase the amount of training data until the reference performance metric is at least almost achieved and then stop changing the amount of training data. Thus, the skilled person would not consider decreasing the amount of training data so that the measured performance metric oscillates towards the reference performance metric.
As a result, the Board concluded that claim 1 of the Main Request involves an inventive step and thus set aside the decision under appeal.
More information
You can read the whole decision here: T 0183/21 (Controlling the performance of a recommender system/BRITISH TELECOMMUNICATIONS) of September 29, 2023.