Examining patent applications involving artificial neural networks
Published 25 July 2024
Summary
1. On 19 July 2024 the Court of Appeal handed down its judgment in Comptroller-General of Patents, Designs and Trade Marks v Emotional Perception AI Ltd [2024] EWCA Civ 825.
2. The judgment concerns an appeal from the earlier judgment of the High Court in Emotional Perception AI Ltd v Comptroller-General of Patents, Designs and Trade Marks [2023] EWHC 2948 (Ch) which had overturned the decision of the Comptroller in BL O/542/22. The appeal relates to an invention involving an artificial neural network (ANN) and the question whether it is excluded subject matter under the “program for a computer” exclusion of section 1(2)(c) of the Patents Act 1977.
3. By its judgment the Court of Appeal allows the appeal and upholds the decision of the Comptroller, holding that the invention is excluded from patentability under section 1(2)(c) as a program for a computer.
4. The judgment holds that an ANN is a computer and that, whether it is implemented as hardware or software, the weights and biases of the ANN are a program for a computer and are therefore within the purview of the “program for a computer” exclusion.
5. The judgment strongly emphasises that this does not mean ANN-implemented inventions are unpatentable. It simply means that ANN-implemented inventions are in no better and no worse position than other computer implemented inventions.
6. Following the judgment, the office is making an immediate change to practice for the examination of ANNs for excluded subject matter. Patent examiners should treat ANN-implemented inventions like any other computer implemented invention for the purposes of section 1(2).
7. Examiners should apply the Aerotel approach to assess whether an ANN-implemented invention makes a contribution which is technical in nature. The judgment expresses the four steps of the Aerotel approach as follows:
(1) Properly construe the claim
(2) Identify the actual contribution (although at the application stage this might have to be the alleged contribution)
(3) Ask whether it falls solely within the excluded matter
(4) If the third step has not covered it, check whether the actual or alleged contribution is actually technical
8. As with any other computer implemented invention, the AT&T signposts may be a helpful aid for examiners in considering whether an ANN-implemented invention makes a technical contribution. Examiners should apply the signposts carefully, particularly in so-called “better computer” (i.e. better ANN) cases where the invention is said to improve the functioning of the computer (i.e. the ANN) itself.
The invention
9. The invention provides media file recommendations, for example music track recommendations, to a user by passing music tracks through a trained ANN. In this way the invention makes suggestions of similar music in terms of human perception and emotion irrespective of the genre of music.
10. To achieve this, the ANN of the invention is trained in a particular way. It involves training the ANN by considering both natural language descriptions of a music file as it might be perceived by a human (for example the metadata relating to the music file) and the physical properties of the music file (for example the tone, timbre, speed, and loudness).
11. Once the ANN is trained, music tracks are passed through it to produce outputs which are compared to a database from which recommendations of similar tracks are produced. The recommendations are provided to an end user by sending a file and message.
The judgment
12. Lord Justice Birss deals with the appeal by answering two questions (see [55]). The first question involves asking what a computer program is and whether there is a computer program in an ANN. If there is a computer program in an ANN, then the second question asks whether the claimed invention makes a technical contribution.
13. In answering the first question, Birss LJ starts by defining a computer as a machine which processes information. Birss LJ then defines a computer program (noting this is the same thing as a “program for a computer”) as a set of instructions for a computer to do something. One can therefore say that a computer is a machine which processes information in a particular way, and the program is the set of instructions which cause the machine to process the information in that particular way (see [61]).
14. Turning to consider an ANN, Birss LJ states that however it is implemented, such a machine is clearly a computer – it is a machine for processing information. Irrespective of the manner in which an ANN is implemented (hardware or software), the weights and biases of the ANN are a computer program. They are a set of instructions for a computer to do something. The set of weights and biases as a whole instruct the machine to process information it is presented with in a particular way (see [68]).
15. The judgment therefore holds that the exclusion from patentability of a program for a computer as such in s1(2) of the 1977 Act is engaged in this case. Whether implemented as a hardware ANN or a software ANN, the weights and biases of the ANN are a program for a computer and are therefore within the purview of the exclusion ([70]).
16. In answering the second question, Birss LJ emphasises that the fact that s1(2) of the Act is engaged in a case of an ANN-implemented invention does not mean it is unpatentable. The fact the exclusion is engaged simply means that ANN-implemented inventions are in no better and no worse position than other computer implemented inventions ([71]).
17. In order to assess whether there is a technical contribution, the judgment goes through the steps of Aerotel, which it expresses (see [31]) as follows:
(1) Properly construe the claim
(2) Identify the actual contribution (although at the application stage this might have to be the alleged contribution)
(3) Ask whether it falls solely within the excluded matter
(4) If the third step has not covered it, check whether the actual or alleged contribution is actually technical
18. It is noted that there are no issues of construction at step 1 and no dispute as to the contribution at step 2 (see [73]). The analysis therefore focusses on steps 3 and 4.
19. The judgment notes that, put simply, the program here provides file recommendations. That is what it does. It is therefore not concerned with improvements inside the running of a computer ([76]). The fact that a file is actually sent or there is an external transfer of data does not help. What distinguishes it from standard file transmission or data transfer is that the file or data represents a better recommendation (see [77] and [81]).
20. The judgment states that what makes the file worth recommending are its semantic qualities. This is a matter of aesthetics, or as the Hearing Officer put it in his decision, they are subjective and cognitive in nature. They are not technical (see [79]).
21. The judgment reaffirms that the AT&T signposts can be very helpful when considering whether a computer program makes a technical contribution, particularly in “better computer” cases where an invention is said to improve the functioning of the computer system itself (see [34]). In this case, the judgment found that the signposts did not assist the applicant in identifying any technical effect ([82]).
22. The judgment therefore concludes that the application is excluded from patentability ([83]).
23. The judgment decides there is no need to consider the mathematical method exclusion. However, the judgment observes that it is hard to see why even if the weights and biases of an ANN are not to be regarded as a computer program for some reason, they are not in any case a mathematical method, and so the very same analysis based on the Aerotel approach would apply with the same result (see [84]).
IPO practice regarding ANNs and excluded subject matter
24. Following this judgment, patent examiners should treat ANN-implemented inventions like any other computer implemented invention. Examiners should use the Aerotel approach to assess whether the claimed invention makes a contribution which is technical in nature.
25. As with any other computer implemented invention, the AT&T signposts may be a helpful aid for examiners in considering whether an ANN-implemented invention makes a technical contribution. Examiners should apply the signposts carefully, particularly in so-called “better computer” (i.e. better ANN) cases where the invention is said to improve the functioning of the computer (i.e. the ANN) itself.
26. For the avoidance of any doubt, this notice supersedes patent practice notice PPN 03/23 which stated that examiners should not object to inventions involving an ANN under the “program for a computer” exclusion.
27. The Manual of Patent Practice (MoPP) and the office’s guidelines for examining patent applications relating to artificial intelligence (AI) inventions will be updated to reflect the Court of Appeal’s judgment in due course.
28. Any comments or questions arising from these guidelines should be addressed to either:
Ben Buchanan,
Intellectual Property Office
Concept House
Cardiff Road
Newport
South Wales
NP10 8QQ
Telephone: 01633 814742
Email: Ben Buchanan
Nigel Hanley
Intellectual Property Office
Concept House
Cardiff Road
Newport
South Wales
NP10 8QQ
Telephone: 01633 814746
Email: Nigel Hanley