Our partners and attorneys are highly qualified and highly experienced to advise you in all areas of Intellectual Property law. We advise start-ups, SMEs, and multinational corporations and ensure that your inventions, brands and designs are expertly protected.
This forms part of a wider report on the 2022 updates to EPO Guidelines for Examination.
We applaud the EPO’s decision to add a first example assessment of inventive step for a technical application of artificial intelligence, and another new example assessing inventive step for a claim including mathematical method steps.
See Part G, Chapter VII, sections 5.4.2.4 and 5.4.2.5. We look forward to seeing more of these examples in the 2023 Guidelines update.
This new example of G-VII 5.4.2.5 is a method that includes use of a mathematical algorithm to control an industrial process. The usefulness of the example is its confirmation that a computer-implemented mathematical method step can contribute to the technical character of a claimed invention when the claims are functionally limited to the control of an industrial process.
This is consistent with the EPO’s comments on patentability of “specific technical applications” of mathematical methods in Guidelines G-II, 3.3. The fact that the mathematical method steps can contribute to technical character is significant, because the EPO only considers features that contribute to technical character when it is assessing inventive step.
In the example, a neuro-fuzzy controller (a computer-implemented algorithm) combines a neural network and fuzzy logic rules to map statistical relationships between input variables and output variables for the purpose of automatic adjustment of process parameters of a thermal spray coating method. Without an appropriate control, the properties of the coatings would vary significantly.
The example of Guidelines G-VII, 5.4.2.5 includes this claim:
A method for coating a workpiece comprising:
a) applying a material to the workpiece by thermal spray coating,
b) monitoring the thermal spray coating process in real-time by detecting properties of particles in a spray jet and supplying the properties as actual values,
c) comparing the actual values with target values,
and, in the event that the actual values deviate from the target values,
d) adjusting process parameters for the thermal spray coating automatically by a controller on the basis of a neural network, said controller being a neuro-fuzzy controller which combines a neural-network and fuzzy logic rules and thereby maps statistical relationships between input variables and output variables of the neuro-fuzzy controller.
As noted in the EPO Guidelines, step a) clearly relates to a technical industrial process, and step b) is required for the control of the coating process and so also contributes to technical character (even though the specific monitoring apparatus is not claimed). Step c) determines whether a control action is necessary, and so this is also relevant to the technical purpose in the context of the claim as a whole. Step d) relates in part to the physical adaptation of the process parameters (which is clearly technical), and in part to the mapping operation of a neuro-fuzzy controller – the latter is technical in this context because it is used to adjust the control parameters and therefore controls the coating process.
Since the output of the neuro-fuzzy controller’s processing is explicitly stated to be used in the control of an industrial process, which is a technical effect, this mapping and adjusting in step d) also contributes to the technical character of the invention.
The example in the Guidelines continues by discussing a situation in which a citation D1 discloses most features of claim 1 including automatic control of a thermal spray coating process including automatic adjustment of process parameters based on a neural network analysis – i.e all features except the second part of step d) which is the use of a neuro-fuzzy controller combining a neural-network with fuzzy logic rules to map statistical relationships between input and output variables. Compared with D1, the claim recites a new mathematical method step that is used to produce a known technical effect in a new way.
The point to note here is that this difference would be enough for patentability, in the absence of additional prior art.
The EPO’s example ends with a conclusion that the example claim was not inventive, because the only point of novelty over D1 (the use of a fuzzy-controlled neural network) was well known at the filing date of the application. The Guidelines example refers to a second prior art document D2, which is a document disclosing a combination of a neural network and fuzzy logic rules providing a neuro-fuzzy controller for a highly non-linear processes. We should not be distracted by this negative conclusion, because the conclusion is based on specific prior art – i.e. the negative conclusion arises from the facts of the particular case on which the Guidelines example is based.
The important point is that all features of the claim were capable of contributing to an inventive step – they would not be disregarded by the EPO.
Software-implemented mathematical methods are often protectable in Europe, including technical applications of AI if the claims are functionally limited to a technical purpose.
Category: Latest Insights | Published: | Read more
