The European Patent Office (EPO) has issued official guidelines on the patenting of artificial intelligence and machine learning technologies. The guidelines became valid on November 1st, 2018.
The new guidelines aim to help in the assessment of whether a particular AI-related invention has the necessary technical characteristics to receive a European patent. According to the new guidelines, regardless of whether or not they can be “trained”, artificial intelligence and machine learning models and algorithms are considered abstract mathematics in nature. In order to be patentable, a proposed AI/ML entity must have “a technical character as a whole.”
When determining whether the claimed subject-matter satisfies this condition, the guidelines note that expressions such as “support vector machine,” “reasoning engine” or “neural network” may not qualify, as these are regarded as terms for mathematical methods which do not have a unique technical character of their own.
The EPO provided examples of mathematical methods (ie artificial intelligence and machine learning related applications) which can pass the technical character threshold:
- Digital audio, image or video enhancement or analysis, e.g. de-noising, detecting persons in a digital image, estimating the quality of a transmitted digital audio signal;
- Separation of sources in speech signals; speech recognition, e.g. mapping a speech input to a text output;
- Encoding data for reliable and/or efficient transmission or storage (and corresponding decoding), e.g. error-correction coding of data for transmission over a noisy channel, compression of audio, image, video or sensor data;
- Encrypting/decrypting or signing electronic communications; generating keys in an RSA cryptographic system;
- Optimising load distribution in a computer network;
- Providing a medical diagnosis by an automated system processing physiological measurements
- (A generic purpose such as “controlling a technical system” is not sufficient to confer a technical character to the mathematical method. The technical purpose must be a specific one.)
One filed example is a classification of text documents that was classified as a linguistic purpose rather than a technical purpose. Meanwhile, two patentable examples use neural networks in heart-monitoring apparatuses to help identify irregular heartbeats. The EPO distinction again is that the neural networks in these examples are not purely abstract models but are used to achieve technical purposes.
The new guidelines set the European patent standard for artificial intelligence and machine learning technologies. It would seem the EPO is more likely to grant patents to integrated AI and ML solutions, rather than the core technologies themselves.
Author: Victor Lu | Editor: Michael Sarazen