6. Reproducibility

You are viewing the 9th edition (2019) of this publication; for the 10th edition (2022) see here

6.1. Repeatability

6.2. Hypothetical embodiments

6.3. Variants

6.4. Use to which invention is put

6.5. Reach-through claims

6.6. Reproducibility without undue burden

6.7. Trial and error

6.8. Post-published documents

New decisions

T 161/18

1. Die vorliegende, auf maschinellem Lernen insbesondere im Zusammenhang mit einem künstlichen neuronalen Netz beruhende Erfindung ist nicht ausreichend offenbart, da das erfindungsgemäße Training des künstlichen neuronalen Netzes mangels Offenbarung nicht ausführbar ist.

2. Da sich im vorliegenden Fall das beanspruchte Verfahren vom Stand der Technik nur durch ein künstliches neuronales Netz unterscheidet, dessen Training nicht im Detail offenbart ist, führt die Verwendung des künstlichen neuronalen Netzes nicht zu einem speziellen technischen Effekt, der erfinderische Tätigkeit begründen könnte.

OJ Supplementary Publications

Case law 2020: In T 161/18, the application used an artificial neural network to transform the blood pressure curve measured on the periphery into the equivalent aortic pressure. As regards how the neural network according to the invention was trained, the application disclosed only that the input data should cover a wide range of patients differing in age, sex, constitution type, state of health, etc. to prevent the network from becoming specialised. However, it did not disclose what input data were suitable for training the network or even a suitable set of data for solving the technical problem in question. As a result, the skilled person could not reproduce the network's training and so could not carry out the invention. The invention, which was based on automated learning, in particular in connection with an artificial neural network, was thus insufficiently disclosed, because the training it involved could not be reproduced owing to a lack of disclosure in this regard. See also chapter I.C.4. "Effect not made credible within the whole scope of claim – neural network".
Case law 2019: In T 383/14 (sorting table for grape harvest), the board held that a claim was an attempt to define a device in terms of ideal conditions, i.e. those required for its theoretically optimal or nominal operation. However, when considering a claim, the skilled person would readily understand that the conditions of actual operation would not be the ideal ones defined there. Thus, on reading the claim at issue, they would immediately grasp how the table would operate in practice after a harvest and so understand the contested term "only" not in an exclusive sense, but one compatible with the actual operation of all mechanical devices, whose reliability or success rate was always less than 100% and even lower in the specific case of sorting or grading. Thus the board was unconvinced by the opponents' argument that not "only" the grapes passed through the openings because it was unrealistic to expect as much when reproducing sorting; the fact that not only the grapes passed through resulted from an occasional failure entirely to be expected in the case of a sorting device and allowed for by the case law.