T 2353/22 - Traversing a data structure

Key points

• The real surprise is at the end of this decision against a refusal.
• "The invention concerns a method, [...] for supporting the storage of, access to and display of lineage metadata about data stored in a storage system. The lineage metadata of a data object provides information about the sources from which the data object was derived. For instance, how the data object was generated, from which source it was imported, how it has been used by applications, how it relates to other datasets or how its modification will affect tables"
• The claim is lengthy.
• "The method according to the claimed invention includes a first step of receiving metadata from a data source, the metadata describing nodes and edges. Each node represents a metadata object, which can be a data element or a transformation. An edge represents a one-way effect of one node upon another node. According to the description, page 10, first paragraph, the data elements can represent, for instance, "datasets, tables within datasets, columns in tables, and fields in files, messages, and reports". An example of a transformation is "an element of an executable that describes how a single output of a data element is produced"."
• After reading the decision, I don't see any limitation to technical data. I understand the metadata could be metadata of pension contracts or insurance contracts, for example. The method is just generic.
• "After the step of receiving metadata from a data source, the claimed method further includes steps of generating a data structure representing the received metadata and receiving a query for lineage metadata. In response to receiving the query, the data structure is accessed and a response to the query is generated and sent to a computer system for display. The response includes the lineage metadata responsive to the query."
• The claim also includes: "in response to receiving the query (114, 126), traversing the data structure (134) in accordance with the walk plan (130) to collect lineage metadata stored in the data structure (134) that is responsive to the query (114)," 
• The walk plan is about following edges from node to node and is included in the query, i.e. preparing the walk plan is not a part of the claimed method.
• I understand graph theory, a mathematical theory, is used by the features of the edges and nodes.
• The Board: "It is true that since computer programming involves technical and non-technical aspects, it is difficult to distinguish between the "programmer as such" who, as long as they only develop abstract algorithms, are not a skilled person within the meaning of the case law, and the "technical programmer" (see also T 697/17, reasons 5.2.4). "
• "The subject-matter of claim 1 differs from the method of D2 in that it includes features (c1) to (f), (f2), (g) and (g1). These features specify details of the data structure (features (c1) to (c4)), the step of receiving a query for lineage metadata including the identification of a data element, a type of lineage and a walk plan (features (d) to (d2)), the steps of traversing the data structure and collecting data using the data structure and the walk plan (features (d3), (d4) and (e) to (e4)), the step of generating a response including lineage metadata responsive to the query ((f) and (f2)) and the steps of sending the response to a computer system and displaying the lineage data (features (g) and (g1))."
  
• "None of the cited prior-art documents disclose the combination of distinguishing features (c1) to (f), (f2), (g) and (g1). In the board's opinion, it would be within the ordinary skills of the computer expert to arrive at the data structure specified in the distinguishing features, which corresponds directly to the non-technical lineage structure. However, the board is not convinced that the skilled person would arrive at the combination of all the distinguishing features, including a walk plan to direct the way the data structure is traversed and the data is collected as claimed."
• "Therefore, the subject-matter of claim 1 of the main request involves an inventive step (Article 56 EPC)."
• Now, as a question to readers, this decision was issued by Board 3.5.07. Would Board 3.5.06 have decided differently (see yesterday's post about T1998/22 Wide and deep machine learning) ? 
• 
  
• 
  

EPO 

The link to the decision can be found after the jump.

T 1998/22 - Machine learning

Key points

• "It is noted that the use of a "cross-product feature transformation" in the "wide machine learning model" [as specified in the amended claims] may contribute to establishing that the advantages recited in paragraph [13] are actually achieved, as emphasised by the appellant (statement of grounds appeal, pages 10 and 11), but these advantages [see below] remain of a non-technical nature."
• "According to the description [paragraph 13], "in general, a wide machine learning model can memorize feature interactions through a wide set of cross-product feature transformations and a deep machine learning model can generalize unseen feature combinations by applying embedding functions to the input features", and "by including both [a] deep machine learning model and [a] wide machine learning model, the wide and deep machine learning model can obtain both benefits of memorization and generalization and thus can perform better on predicting an output from a set of input features" (paragraph [13])."
  
  
• "As to the argument that only a "technical expert" could have devised the features of the invention [i.e. the CardinalCommerce test], the board notes, as a general word of caution, that this kind of enquiry may be helpful in some cases to separate non-technical features from technical features - in particular to identify business-related features - but does not constitute a definite test as it only concerns which kind of considerations underlie some features of the invention and not which kind of effects are achieved by it. For instance, a claim to a computer-implemented simulation may involve features which are based on expertise in the technical field of the technical system that is being simulated. This alone would however not be sufficient to conclude that these features contribute to the technical character of the claim (G 1/19, reasons 122, 125, 141 and 142)."
• "In any case, the board tends to consider that claim 1 does not reflect any considerations beyond computer programming (which encompasses the design of algorithms) and mathematics."

EPO 

The link to the decision can be found after the jump.

T 1425/21 - Would improved machine learning be technical?

Key points

• This decision was published in April 2024, but I missed it then. Credits to Phyllis Luana Graf (linkedin) for flagging it on LinkedIn.
• The invention is about machine learning, applicant Google. The parent application was filed in 2015 and was granted.
• "The application relates to machine learning models such as deep neural networks. It proposes to approximate "cumbersome" machine learning models with "distilled" machine learning models which require less computation and/or memory when deployed. For instance the distilled model may be a neural network with fewer layers or fewer parameters. The cumbersome model may be an ensemble classifier, possibly combining full classifiers with specialist classifiers. The distilled model is trained on a "plurality of training inputs" and the associated outputs of the cumbersome model, so as to "generate outputs that are not significantly less accurate than outputs generated by the cumbersome machine learning model"
• "The Board notes that the features differentiating the invention from D1, or even the entire set of features defining the distilled model and its training, as a difference to a known cumbersome learning model, are mathematical methods which cannot, under the estab­lished case law of the boards of appeal (the "COMVIK" approach), be taken into account for inventive step unless they contribute in a causal manner to a tech­nical effect."
• "[The Board] accepts that the distilled model has reduced memory requirements when compared to the cumbersome model; after all this is expressly claimed. However, a reduction in storage or computational requirements of a machine learning model is insufficient, by itself, to establish a technical effect. One also has to consider the performance of the "reduced" learning model"
• The Board could have explained its reasoning in more detail. Is a technical effect the same as progress? Is it a generally accepted principle that a technical advantage constitutes no "technical effect" if there is also a technical disadvantage? E.g,. for a chemical reactor (my technical field), does a higher throughput constitute no technical effect if the reactor is more expensive? (or vice versa)? Or is this only a rule for software?
• "It is not credible in general that any model with fewer parameters can be as accurate as the more complex one it is meant to replace. For example, the complexity or architecture of the reduced model may be insufficient or inadequate for the given problem."
• "The Board also does not see that the temperature-based training process ensures that the smaller model has an equivalent accuracy. It is not clear how exactly the temperature must be first set (for both models), and then varied, and what accuracy may be expected. The application simply does not discuss this."
• "The soft outputs [of the distilled model] represent a class probability obtained according to a form of the softmax equation using a "temperature" parameter T, which is set higher during training than during subsequent use. [The softmax equation uses e^(x/T) instead of e^x.
• I can't evaluate the Board's technical assessment, of course. Apparently, there was not a lot of evidence on the file.
• " In principle, it appears possible to argue that the smaller model represents a "good" trade-off between resource requirements and accuracy, i.e that the smaller model may be less accurate but have (predic­tably) smaller resource requirements. However, the application lacks any information in that regard."
• Is the decision consistent with T 1952/21 ? 
• 
  

EPO 

The link to the decision and an extract of it can be found after the jump.

T 0323/21 - Mathematical device

Key points

• "The invention relates to an applied cipher technique and, in particular, to a method for determining whether a nondecreasing sequence exists, without revealing input data"
• "There is a method, known from the prior art, called secure computation, for computing encrypted results from encrypted data without decrypting any of the encrypted data. Encryption is performed that distributes pieces of a numerical value among a plurality of secure computers which cooperate to perform a computation in such a manner that the result is distributed among the secure computers without reconstructing the numerical value, that is, with the result and the original value being kept encrypted"
•  Claim 1 is directed to "A nondecreasing sequence determining device (2), "
•  The device of claim 1 according to the main request consists of a combination of functional units that cooperate to assess whether a nondecreasing sequence can be created from a sequence of sets of numbers. This is achieved by selecting elements, one by one from each set in the sequence. The claim does not contain any reference to any concrete use of the result. The various functional units of the device are defined by their mathematical roles in the determination of whether or not there is a nondecreasing sequence."
• "The technical nature of the claimed subject-matter is limited to the existence of the software code running on the claimed device."
• "the mathematical functionalities defined in claim 1, to the extent that they do not interact with technical features to produce a technical effect, cannot justify the existence of an inventive step. "
• "The algorithm underlying the claimed device is exclusively of a mathematical nature. It is without any interaction with the device on which it operates. The technicality of the claimed device lies only in its materiality. Concretely, its technicality is limited to the combination of software code with the associated processing unit. The claimed device with its program code implements a purely mathematical method deprived, in its generality, of any technical purpose."
• "Independent claim 1 does not contain any concrete reference to a use of the claimed device and method. Therefore, independently of the fact that text matching is not technical in itself, no technical contribution can be derived from any specific use of the claimed invention if it cannot be derived, explicitly or implicitly, from the claimed subject-matter. "
• " More fundamentally, the effect put forward by the applicant is also not sufficient to confer technical character to a method which is mathematical by nature. In the absence of interaction with the device on which the algorithm is to be run, the recited method does not add to its technical character."
• Note, the ED rejected the application for lack of clarity. The Board's preliminary opinion was that the claims were clear and sufficiently disclosed, with a single sentence about 'technical contribution' under the inventive step. 
• There was no auxiliary request specifically directed to secure computing, it seems.
• 
  
  

EPO 

The link to the decision can be found after the jump.

T 1132/20 - A cloud-computing provider's non-technical business decision

Key points

• As for difference (a), although document D3 does not disclose any other "middleware function" such as firewall or encryption/decryption functionality, these types of functionality are well known in the art. In the context of document D3, a cloud-computing provider's decision to provide such functionality on an optional basis and configured in accordance with the tenants' individual preferences (as negotiated in per-tenant policy agreements or "contracts") is a non-technical business decision.
• It's been a while since we saw the CardinalCommerce business person on this blog. Apparently, he/she/they moved on from insurance, gambling and tax advice to cloud computing services. 
• "the problem-and-solution approach as developed in the case law of the board of appeal is not a mechanistic tool for assessing inventive step but should rather be seen as a useful framework which can help answering in an objective manner the question whether, having regard to the state of the art, the skilled per son could and would have arrived at the invention. The problem-and-solution approach must be applied having in mind the purpose of Article 56 EPC, which is to prevent a patent from being granted on routine or otherwise obvious modifications of what, at the effective filing date, had been available to the public. "

EPO 

The link to the decision can be found after the jump.

T 1741/22 - Displaying medical data

Key points

• The case was given publication code [B] by the Board, i.e. of general interest.
"Claim 1 of auxiliary request 10 contains the following limiting features (board's labelling):
(a) A system for analysing glucose monitoring data indicative of a glucose level in a bodily fluid, comprising: an input device, a data processing device, an output device, a display device, and machine-readable instructions that are executed by the data processing device,
(b) [wherein the machine-readable instructions cause the data processing device to] receive continuous glucose monitoring data via the input device, the continuous glucose monitoring data indicating a glucose level sampled for a person in a bodily fluid at a plurality of sample times over a measurement time period in a continuous glucose level measurement, and comprising a plurality of continuous glucose profiles, each of the glucose profiles comprising a plurality of glucose values with a glucose value for each of the plurality of sample times over the measurement period, wherein the plurality of glucose profiles is determined on different days by sampling the glucose level on each day over the measurement period, wherein the measurement period is 24 hours;
(c) [wherein the machine-readable instructions cause the data processing device to] for the plurality of continuous glucose profiles, determine a plurality of minimum glucose values and/or a plurality of maximum glucose values for a selected group or each of the plurality of respective sample times
(d) [wherein the machine-readable instructions cause the data processing device to] provide first display signals representing the plurality of minimum glucose values and/or the plurality of maximum glucose values for the selected group or each of the plurality of respective sample times;
(e) [wherein the machine-readable instructions cause the data processing device to] output the first display signals via the output device to the display device; and display a first graphical representation according to the first display signals on the display device."
• Essentially, the system has a display to display minimum and maximum values from 24 hours of received "continuous glucose monitoring data" (from a wearable glucose measurement device,  I guess). The system itself is not a continuous glucose monitor.
• "The appellants [proprietors] argued that the distinguishing features of claim 1 of auxiliary request 10 over D1 were features (c) and (d), i.e. determining and displaying minimum/maximum glucose values. [D1 shows percentiles]. [The proprietors] stated that the technical effect of the distinguishing features was to provide an "improved analysis of glucose monitoring data". In particular, "the plurality of minimum/maximum glucose values may correspond to medically relevant outlier values", which "would otherwise be averaged out in the context of known methods employing percentiles as in D1". "
• The question is whether these features provide for an inventive step. The Board approaches it from the question of whether the features provide a technical contribution ("contribute to the technical character of the invention").
• As an aside, a test of the Board's reasoning is if it makes sense if the distinguishing feature was "calculating the third harmonic eigenvalue" (instead of minimum/maximum, just making up a fancy term here) and the clear and convincing evidence is that this value, hidden in what was considered random noise before the invention, turns out to be life-saving information. That helps to distinguish between 'this is obvious for any scientist in the field but I don't have the right prior art at hand' and 'it is very smart but not technical'.
• The Board: "if the mere generation of "new data" were sufficient to contribute to the technical character of the invention, Article 52(2) and (3) EPC would contain meaningless limitations of patentable subject-matter, as e.g. mathematical methods are supposed to constantly generate "new data"."
• The Board then sets out that steps C and D are not actual measurement steps and are not, for that reason, technical.
• Note that even including a technical measurement step may not necessarily mean that the further steps of processing the resulting data are automatically technical in the sense of G1/19 (T 0489/14, r.7.4). 
• "The appellants [proprietors] referred to T 2681/16 and to the Guidelines for Examination in the EPO in support of their view. "
• Dealing with the argument based on the GL first in this post: "the Guidelines for Examination in the EPO (in its applicable version of March 2022 and also in its current version of March 2024), section G-II, 3.3, which relates to the technical contribution of mathematical methods, lists "providing a medical diagnosis by an automated system processing physiological measurements" among "examples of technical contributions of a mathematical method". As providing a "medical diagnosis" - whether done by a physician or by an automated system - is devoid of any technical character (see e.g. G 1/04, Reasons 5.3 and 6.3), this example is clearly erroneous. As there is no further explanation, let alone a reference to any case law, the board sees no reason to speculate on how the Guidelines came up with this example (cf. Article 20(2) RPBA)."
• I will quote the relevant parts in G1/04 in full. It would have been helpful if the TBA had identified the relevant sentences - especially because G 1/04 was about Art. 52(4) EPC 1973, i.e. about industrial applicability, not about non-inventions of Art. 52(2) EPC 1973. 
• I will highlight below some parts of G1/04, cited paragraphs, that are at first sight a bit at odds with the TBA's reasoning based on the same paragraphs of G1/04, to the extent I can understand the TBA's reasoning in all its brevity. 
• G 1/04: "5.3 Since diagnostic methods referred to in Article 52(4) EPC are inventions within the meaning of Article 52(1) EPC (cf. point 4 above), it follows that, in a situation where the deductive medical or veterinary decision phase is a purely intellectual exercise, i.e. a step of a non-technical nature, such a method must necessarily further include preceding steps (cf. point 5 above) of a technical nature, in order to satisfy the requirements of Article 52(1) EPC. The subject-matter of a claim including technical and non-technical features may satisfy the requirements of Article 52(1) EPC if the non-technical features interact with the technical features in order to bring about a technical effect (cf. T 603/89 (OJ EPO 1992, 230), point 2.5 of the Reasons). 
• "6.3 In the judgment of the Enlarged Board of Appeal, the qualification of an activity as having a diagnostic character may not depend on who is involved. The wording of Article 52(4) EPC is unequivocal in that the exclusion relates only to the method, and not to the person carrying out the method. Furthermore, no indication can be found in the preparatory documents to the EPC which would restrict the exclusion of diagnostic methods from patentability to a certain group of persons such as medical or veterinary practitioners. Also, as already mentioned under point 6.1 above, defining the medical or veterinary practitioner on a European level within the framework of the EPC is difficult if not altogether impossible. To allow the grant of a European patent to depend on the involvement of such a person would therefore introduce legal uncertainty into the patent granting procedure. Thus, whether or not a method is a diagnostic method within the meaning of Article 52(4) EPC should neither depend on the participation of a medical or veterinary practitioner, by being present or by bearing the responsibility, nor on the fact that all method steps can also, or only, be practised by medicinal or non-medicinal support staff, the patient himself or herself or an automated system. This also reflects the well-known fact that technological advances penetrate human and veterinary medicine and the medical and veterinary profession. Today, and more than at any time before, technology is about to fundamentally alter how and by whom health care is administered, with the result that human and veterinary medicine is gradually being reshaped by technology. In a changing medical or veterinary environment brought about by technological progress, the need for reconsidering the relationship between medical or veterinary practitioners and non-medicinal support staff will become more pressing than ever before. This will have implications for the non-medicinal support staff in terms of profile and expansion in that a great variety of diagnostic and other information will have to be procured and gathered by these persons. Moreover, no distinction should be made in this context between essential method steps having diagnostic character and non-essential method steps lacking it. The reason for this judgment lies in the fact that, again contrary to the requirement of legal certainty, the assessment of the factual and legal situation in connection with these issues could change considerably in time. As has been mentioned under point 6.1 above, consideration might be given to exploring the possibility of protecting the activities of medical and veterinary practitioners by other means on the national level."
• The Board also " disagrees with the finding in T 2681/16 that providing an overall "measure" of the glucose variability and a prediction of glycemic events amounts to a technical effect.". "In the present case and in the case underlying T 2681/16, where the "physical reality" is typically the "patient's blood", the interaction with the physical reality ends once blood glucose measurements are carried out, directly on the relevant physical entity "blood", or indirectly e.g. on another bodily fluid [note; and these physical steps are not claimed in the patents] The [note: actually claimed steps of the ] provision of overall glucose variability and a prediction of glycemic events are mathematical steps or intellectual activities which take place in the absence of this interaction with the physical reality and are therefore not "measurements" in this sense."
• "generating (and displaying) further data by an evaluation or interpretation of these measurements (as done according to features (c) and (d) here) amounts to "measurements" generated merely by a cognitive or mathematical exercise that is inherently non-technical. "
• G 1/19 r.98: "Calculated status information or physical properties concerning a physical object are information which may reflect properties possibly occurring in the real world. However, first and foremost, they are mere data which can be used in many different ways. There may exist exceptional cases in which such information has an implied technical use that can be the basis for an implied technical effect. Still, in general, data about a calculated technical effect is just data, which may be used, for example, to gain scientific knowledge about a technical or natural system, to take informed decisions on protective measures or even to achieve a technical effect. The broad scope of a claim concerning the calculation of technical information with no limitation to specific technical uses would therefore routinely raise concerns with respect to the principle that the claimed subject-matter has to be a technical invention over substantially the whole scope of the claims"
• G 1/19, r.128 mutatis mutandis: " In the Enlarged Board's view, calculated numerical data reflecting the physical behaviour of a system [from measurements] usually cannot establish the technical character of an invention in accordance with the COMVIK approach, even if the calculated behaviour adequately reflects the behaviour of a real system underlying the [measurement]. Only in exceptional cases may such calculated effects be considered implied technical effects (for example, if the potential use of such data is limited to technical purposes"
• See also r.98: "calculated status information or physical properties concerning a physical object are information which may reflect properties possibly occurring in the real world. However, first and foremost, they are mere data which can be used in many different ways. There may exist exceptional cases in which such information has an implied technical use that can be the basis for an implied technical effect. Still, in general, data about a calculated technical effect [?] is just data, which may be used, for example, to gain scientific knowledge about a technical or natural system, to take informed decisions on protective measures or even to achieve a technical effect. The broad scope of a claim concerning the calculation of technical information with no limitation to specific technical uses would therefore routinely raise concerns with respect to the principle that the claimed subject-matter has to be a technical invention over substantially the whole scope of the claims (see point E.I.b above, referring to T 939/92)."
• See also the broken technical chain fallacy as reviewed in T 1670/07: "the Board sees something of a well known argument that could be termed the "broken technical chain fallacy" after decision T 1741/08 - GUI layout/SAP. This decision dealt with the fairly common situation that arises in connection with graphic user interfaces (GUIs) where a technical effect might result from the user's reaction to information. The decision essentially concluded (see point 2.1.6) that a chain of effects from providing information to its use in a technical process is broken by the intervention of a user. In other words, the possible final technical effect brought about by the action of a user cannot be used to establish an overall technical effect because it is conditional on the mental activities of the user. This applies to the present case because any possible technical effect depends on the user's reaction to the itinerary."
• These observations are without prejudice to the question of whether claims involving the display of medical diagnostic data should benefit from the exception of G1/19 r.98 and r.128 of an implied technical use and effect of the calculated or displayed data, the possible argument being the potential use of the data is (practically, realistically) limited to medical purposes (methods of treatment of patients) and that medical purposes are technical purposes.

EPO 

The link to the decision and an extract of it can be found after the jump.

T 1952/21 - Is AI / machine learning as such technical?

Key points

• Inventive step is at issue for a claim directed to a "machine learning system" comprising generic hardware and some sotware.
• The Board: "The application relates to reinforcement learning. In reinforcement learning, an agent explores the environment according to a policy, determining which action the agent takes (e.g. move right) at every juncture as a function of its current state (e.g. its position in the environment). The agent receives rewards, positive or negative. In this way it can "learn" the value of the various actions and states. The goal of training is to maximize a value function which reflects the expected sum of rewards given a certain action. The application builds upon the method of D1, called A3C (asynchronous advantage actor-critic). That method separately approximates the policy and value models as neural networks. The raw input (describing the environment) is preprocessed in sequence by a feed forward network. The result is fed to the value and the policy networks. Developments of that method [i.e. the distinguishing feature of claim 1, as I understand it], termed NoisyNet A3C in the current application, inject randomness into the training by using stochastic weights (e.g. by adding random noise or using stochastic models) in the policy and value networks. This allows for further exploration of the parameter space )." [initernal citations omitted]
• The decision does not seem to explicitly identify the distinguishing features of claim 1 over D1. Claim 1 recites: "wherein said feed-forward neural network comprises stochastic units". This may be the relevant feature.
• The Board extensively summarizes the applicant's arguments, but I'm more interested in the Board's reasoning.
• The Board: "The Board remarks that pseudorandom number generators were known to the person skilled in the art. Their use, in general or in the more specific context of "stochastic units" [] does not change in substance the computer, which remains as "deterministic" as any conventional computer. So the Board cannot see a contribution on this level."
• I'm not sure if "well known" and "non-technical" are the same under the EPO approach to inventive step (cf. the US approach in Alice, 573 U.S. 208)
  
• "The system for reinforcement learning as claimed is a neural network, comprising various sub-networks, implemented on a computer. The network, as a whole, defines a mathematical function mapping inputs into outputs. Effectively, the claim is to a mathematical method implemented on a computer."
• I guess this observation is true for any type of software?
• "the Board holds that the Enlarged Board decision G 1/19, addressing the patentability of computer-implemented mathematical models for simulation, should be the starting point when assessing the technical character of reinforcement learning. It is commonly accepted that a large part of the findings in G 1/19 apply to any computer implemented inventions.
• " In G 1/19, the Enlarged Board of Appeal stated (reasons 137) that (simulation) models by themselves are not technical but that "they may contribute to technicality if, for example, they are a reason for adapting the computer or its functioning, or if they form the basis for a further technical use of the outcomes of the simulation". However, "such further use has to be at least implicitly specified in the claim"."
• The Board finds there is no implicit further technical use in the case at hand.
• " The Board has already explained above that the functioning of the computer, or the computer itself, are not adapted. "
• Regrettably, I can't find the 'above' quickly.
• "This conclusion is consistent with that in the case T 702/20, which is in many ways similar to the present one, where this Board (in a different composition) decided, also following G 1/19, that a trained machine learning model, namely a neural network, can "only be considered for the assessment of inventive step when used to solve a technical problem, e.g. when trained with specific data for a specific technical task" []."
• "The Appellant also argued that reinforcement learning was technical based on an analogy with the case law regarding cryptography, in particular RSA [T 1326/06]
• From that case: "Nach Meinung der Kammer handelt es sich beim sicheren Austausch von elektronischen Nachrichten um eine technische Wirkung, die zu erzielen als eine technische Aufgabe angesehen werden muss. RSA löst diese Aufgabe mit mathematischen Mitteln. Mit RSA gelang ein Durchbruch in der Entwicklung der Kryptografie: RSA wird als das erste praktikable, konkret implementierbare asymmetrische Kryptosystem angesehen und ist heute in zahlreichen kryptografischen Sicherheitssystemen eine zentrale Komponente. Die RSA zugrundeliegende Mathematik dient somit unmittelbar der Lösung eines konkreten technischen Problems."
•  "The Board notes that [...] RSA and reinforcement learning are different and serve different purposes. In particular, RSA and other cryptographic methods have a specific, and at least implied, purpose, namely data security. This is not the case for reinforcement learning. So the findings regarding RSA cannot directly be transferred to reinforcement learning."
• This raises the question: why is 'data security' technical and 'machine learning' not?
• To me, as a simple chemist, ChatGPT feels just like database management technology, turbo code interleavers for wireless communication, and copy/paste (clipboard) on the computer: I have (basically) no idea how it works, but it definitely makes my computer more useful, even though they are perhaps not all prize-winning.
• Whether introducing randomness for "further exploration of the parameter space " in the given context is obvious or not, I don't know. 
• "The Board accepts that the use of the term technical in the case law of the Boards of Appeal may differ from its use elsewhere in society, especially from its colloquial use. However, this does not mean that the Boards of Appeal interpret the law incorrectly: it is common place that the legal interpretation of a term may differ from its colloquial meaning. In particular, the Boards use the term "non-technical" to denote matter excluded under Article 52(2) and (3) EPC. Any alternative interpretation of the terms "technical" and "non-technical" can only be used to justify the patentability of subject-matter to the extent that it does not contradict the law, in particular the exclusion of mathematical methods."

The link to the decision and an extract of it can be found after the jump.

T 0183/21 - Inventive training method for machine learning

Key points

• The application relates to a system for recommending content in a video-on-demand platform. The system is trained and retrained.
• The application states that "providing data from a huge number of users to retrain a recommender system presents challenges in that it takes up system resources." 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 claim specifies, essentially, measuring the performance, comparing it to a desired performance level, and using the comparison result to set the amount of training data provided to the recommender system. Moreover, the recommender system is trained for each user separately.
• "[t]he recommender input controller 16 seeks to adapt the type and amount of usage data provided as training data to the recommender system 18 for each individual client device 22 to provide the minimum amount of data to drive the recommender system 18 towards the predetermined level of recommendation performance yref for each client device 22"
• " there is a positive correlation between the amount of training data specified by the control parameter and the measured performance metric received in the subsequent iteration"
•  "In the subsequent iteration, training data are derived from "usage data" from the client device associated with a user to which the recommendations are provided. The amount of training data derived from the usage data is based on the generated value or values of the control parameter"
• " The technical effect of the distinguishing features listed under point 8.1 is that the use of network bandwidth required to provide the training data to the recommender system is minimised, as is the amount of storage necessary for storing said training data in the communications system including the client device and the recommender system"
• "The board has come to the conclusion that this technical effect is achieved, on average, over substantially the whole scope of the claim (see, for example, G 0001/19, point 82)."
• " Since achieving a maximum performance metric of the recommender system is of paramount importance in the method of document D1, the skilled person would not use a "reference performance metric" which might be different from [i.e. lower than] a "maximum achievable performance metric", and would have no motivation to consider using a closed-loop control algorithm as claimed."
•   Starting from the disclosure of document D1, if the reference performance metric is exceeded, the skilled person would stop changing the amount of training data but would not decrease it, so that the measured performance metric oscillates towards the reference performance metric" [i.e. the possibility of decreasing the amount of training data results in oscillation of the performance].
• The board therefore considers the subject-matter of claim 10 and that of the corresponding claims 1 and 14 of the main request to be inventive (Article 56 EPC). It follows that the decision of the examining division is to be set aside.

EPO 

The link to the decision is provided after the jump, as well as (an extract of) the decision text.

T 1768/20 - Obtaining a design model by simulations

Key points

•  " Claim 1 of the main request reads as follows: "Method for generating a design model with an improved characterization of a standard cell with adaptive body biasing in a circuit design process of said cells in digital circuits fabricated in a foundry, whereby the cell is defined by subject to variations due to the fabrication process P of said foundry, a supply voltage VDDnom and an operating temperature T, the method comprising the following steps: considering static deviations in the supply voltage VDDnom ..." 
• " The board considers that, in the current case, the result of the calculations effected by the method steps of claim 1 is a design model with an "improved" characterisation that is obtained by simulation (see characterising part of claim 1, for example) of a standard cell which [design model (?)] is provided in a library file. Such a library file is, at least according to the description, used by customers in later stages of the overall design process [] but the method of claim 1 does not contain steps specifying the use of the produced library file in later design or manufacturing steps. A use of the design model in a manufacturing step would have been particularly relevant since the Enlarged Board of Appeal considered the inclusion of a manufacturing step to be an argument in favour of patentability (see point 134 of decision G 1/19)." 
• " The board is aware that claim 1 ... specifies the ultimate aim of using the improved design model of a standard cell for fabricating integrated chips (digital circuits) in a foundry. However, the manufacturing of a designed digital circuit using the improved design model is neither specified as being part of the method of claim 1 nor implied by specifying the ultimate aim of the design process. Since the result of the method of claim 1 is calculated data for an improved design that is obtained using simulation, it follows from points 97, 98 and 138 of decision G 1/19 that the improved design does not contribute to inventive step since no "further" technical effect, such as controlling a machine in the foundry during a manufacturing process, is derivable." 

The link to the decision is provided after the jump, as well as (an extract of) the text of the decision.

T 1225/19 - Proximity-based reminders

Key points

• The invention is essentially the idea of setting a reminder that is triggered if another person is nearby.
• "The entire scheme can be summarized by the following, purely administrative instructions: "please, remind the user of an action (item), when a contact which is involved in the action (item) is or will be close to the user, in order to realise the action (item)"".
• The Board considers the claims to lack an inventive step based on a straightforward application of the Comvik approach.
• "in addition to the non-technical features, the subject-matter of claim 1 includes technical features, in particular a "processor ... on an electronic device". Thus, the subject-matter of claim 1 relates to an invention within the meaning of Article 52(1) EPC."
• "The features defined in claim 1 represent a mixture of technical and non-technical features. An invention consisting of a mixture of technical and non-technical features and having technical character as a whole is to be assessed with respect to inventive step by taking account of all those features which contribute to the technical character whereas features making no such technical contribution cannot support the presence of inventive step (see T 641/00"
• "All [the technical] features as such are individually known and are also known as being, if necessary, integrated in a notorious computer system. This represents the technically skilled person's notorious knowledge."
• " When starting from a notorious computer system, like a PDA [personal digital assistant] , the business person provides the following purely administrative/business scheme as business constraints to the technically skilled person for implementation on the notorious computer system ... the following, purely administrative instructions: "please, remind the user of an action (item), when a contact which is involved in the action (item) is or will be close to the user, in order to realise the action (item)", whereby the board considers "action" and "action item" being equivalents."
• "Having this business scheme in hand, the skilled person, a software programmer, has to solve the objective technical problem of providing the computer implementation and automation of the above formulated business method in the notorious electronic device. The provided technical solution neither calls for a particular technical implementation nor provides a further technical effect beyond the straightforward computer implementation. In particular, the implementation of this business constraint ("please, remind the user, ... the action item") relating to the determination of the proximity of a contact and the user, is held by the board, contrary to the appellant's view, not inventive. "
• "features a) and b) derive immediately from the straightforward implementation of the non-technical business requirement mentioned above. All other features relating to the remaining steps of the computer implementation of the business scheme are also straightforward implementations since the technical means required for all these various procedural steps are well-known standard features and the definition according to claim 1 is technically unspecified. This is not disputed by the appellant. 
• Since the result of the straightforward implementation of the above cited business scheme is the method defined in claim 1, the board concludes that the subject-matter of claim 1 of the main request does not involve an inventive step "

EPO 

The link to the decision is provided after the jump. 

T 2852/19 - Reselling tickets

Key points

• "The invention relates to optimised occupancy of an event. The aim is to avoid seats remaining empty if visitors cannot reach the event location in time (see introduction of the description).
• The invention proposes to automatically determine whether visitors to an event are in a defined area at a certain distance from the venue. GPS data from visitors' cell phones is used for this purpose. If ticket buyers are not in said area shortly before the start of the event, an alert is sent to them and their ticket is resold in case of cancellation."
• " D1 discloses GPS tracking of event visitors within a specific radius, but D1 does not mention reselling a ticket. D1 is considered closest prior art."
• "D1 therefore does not explicitly disclose: (a) dynamically adapting the distance/area with time; (b) requesting a visitor to cancel a seat reservation; (c) offer the seat for resale."
• The Board identifies two effects: effect (i) of features (a) and (b) is optimising the seat occupancy; effect (ii) of features (b) and (c) is to increase the profit by reselling seats shortly prior to the event in case a visitor has cancelled its venue.
• As a comment, possibly cancelling reservations leads the ticket being offered to people on the waiting list thereby increasing seating capacity.
• " The effect (i) is technical. The effect (ii) is non-technical."
• The Board does not comment on the relation between the feature "(b) requesting a visitor to cancel a seat reservation" and the broken technical chain fallacy (T1670/07).
• The Board, after a review of the Comvik approach: "Therefore, the problem may be formulated as [how to] "optimising the seat occupancy and implementing features (J) and (K), i. e. prompting the user to indicate whether they plan to attend the event and offer the seat for sale"."
• Features J and K are: "(J) sending an alert to the user prompting the user to indicate whether they plan to attend the event; (K) offer the seat for sale to another user based on the user indicating that they will not attend the event.."
• "it is obvious in view of the disclosure of D1 that the size of the area is reduced (e.g. from a 5-mile radius at 7:30 PM to a 2-mile radius at 7:45 PM) as an event start time approaches"
• "Feature (b) relates to an economical model, i.e. reselling a ticket by encouraging a visitor to cancel its reservation if it cannot meet the appointment time. The system of D1 reveals that a visitor is reminded of its reservation ("remind a user of a reservation"). D1 further discloses the option to "cancel a reservation if the user cannot meet the appointment time" by mobile phone. It would be obvious that the reminder contains a link or a request to cancel the reservation if the appointment time cannot be met."
• As a comment, it is interesting to see the Board analysing the obviousness of a feature relating to an economic model.
• "Feature (c) is per se obvious, because it relates to a business method, i.e. reselling a ticket (reservation) in the case a visitor cannot arrive in time."
• The Board also finds feature (c) to be suggested by D1.
• "Features (F), (I), (J) and (K) are therefore obvious in view of the disclosure and teachings of D1."
• Note that features J and K were also included in the objective technical problem. 

 

EPO 

The link to the decision is provided after the jump, as well as (an extract of) the text of the decision.

T 0702/20 - A neural network is (as such) a non-technical feature

Key points

• Board 3.5.06 in the headnote: "A neural network defines a class of mathematical functions which, as such, is excluded matter. As for other "non-technical" matter, it can therefore only be considered for the assessment of inventive step when used to solve a technical problem, e.g. when trained with specific data for a specific technical task."
• Obviously, this headnote concerns computer-implemented neural networks. Claim 1 is reproduced below. It defines a software product. 
• Inventive step is the critical issue. 
• The Board, on the Comvik approach: "following the so-called Comvik approach (T 641/00), it is made sure that only features contributing to the technical character of the invention are considered for the assessment of, in particular, inventive step (cf. G 1/19, reasons 30 (F), cited from T 154/04). In particular, "non-technical" features, understood in this context as features which, on their own, would fall within a field excluded from patentability Article 52(2) EPC (see, e.g., T 1294/16, reasons 35), can only be considered for this assessment if they contribute to solving a technical problem (see also T 1924/17, reasons 15 to 19)."
• As an aside, the Board recalls that "Even technical features may be ignored with regard to inventive step if they do not contribute towards solving a technical problem (see G 1/19, reasons 33)."
• The claimed neural network has a new structure because the claim recites that "the hierarchical neural network [is] formed by loose couplings between the nodes in accordance with a sparse parity-check matrix of a [low-density parity-check code]".
• The Board: "A neural network is composed of nodes, called "neu­rons", linked to each other by edges transmitting the output of one neuron to the input of another. Each neu­ron implements a parameterized mathematical function, typically a weighted addition of its inputs followed by a nonlinear operator (e.g. a threshold, [...]); the parameters are called weights."
• "In principle, it is possible, if cumbersome, to replace the inputs to each neuron by the mathematical functions implemented by the nodes of the previous layer and write down the mathematical function that the network implements as a whole, i.e. the output as a function of the input. "
• The Board: "The proposed network structure [...] defines a class of mathematical functions (see above points 7 and 8), which, as such, is excluded matter. As for other "non-technical" matter, it can therefore only be considered for the assessment of inventive step when used to solve a technical problem".
  
  
• "The Appellant argued that the proposed modification in the neural network structure, in comparison with stan­dard fully-connected networks, would reduce the amount of resources required, in particular storage, and that this should be recognized as a technical effect, following G 1/19, reasons 85."
• "The Board notes that, while the storage and computatio­nal requirements are indeed reduced in comparison with the fully-connected network, this does not in and by itself translate to a technical effect, for the simple reason that the modified network is different and will not learn in the same way. So it requires less storage, but it does not do the same thing. For instance, a one-neuron neural network requires the least storage, but it will not be able to learn any complex data relation­ship. The proposed comparison is therefore incomplete, as it only focuses on the computational requirements, and insufficient to establish a technical effect."
• The Board here seems to imply that reduced storage and computational requirements without reduced performance would be valid technical effects and would cause the claim feature of "loose couplings between the nodes in accordance with a sparse parity-check matrix" to be considered for inventive step.
• The Board's reasoning that a technical effect can only be acknowledged if an advantage is obtained without simultaneous disadvantages seems unusual to me. Cf., e.g, GL G-VII, 10.1: "if [a] worsening is accompanied by an unexpected technical advantage, an inventive step might be present".
• The Board concludes that "The claim as a whole specifies abstract computer-implemented mathematical operations on unspecified data, namely that of defining a class of approximating functions (the network with its structure), solving a (complex) system of (non-linear) equations to obtain the parameters of the functions (the learning of the weights), and using it to compute outputs for new inputs. Its subject matter cannot be said to solve any technical problem, and thus it does not go beyond a mathematical method, in the sense of Article 52(2) EPC, implemented on a computer."
• "For the sake of completeness, the Board also notes the following: even if, as the Appellant argued, general methods for machine learning, and neural networks in particular, were to be considered as matter not excluded under Article 52(2) EPC, it would remain questionable whether the proposed loose connectivity scheme actually provides a benefit beyond the mere reduction of storage requirements, for instance a "good" trade-off between computational requirements and learning capability."
• This issue ( 'remains questionable') seems highly fact-specific and will likely turn on the evidence. There may also be an aspect of 'plausibility'.
• 
  
• The Board: "whether a claimed invention is patentable [i.e., inventive] or not can often be decided by focusing on the tech­ni­cal problems it solves, and by means of which combina­tion of features, be they technical or not, and by answering the question of whether this combination of features is obvious"
• This observation is very interesting but sheds a different light on the headnote: all (distinguishing) features can only be considered for the assessment of inventive step when used to solve a technical problem. This observation also seems to shed a different light on the Board's reasoning, cited above, starting with "even if, as the Appellant argued ...". Whether or not general methods for machine learning, and neural networks in particular, are to be considered as matter excluded under Article 52(2) EPC, is irrelevant, the only question is whether the problem of "reduction of storage requirements" (possibly: while maintaining good learning capability)  is a technical problem and whether the claimed solution solves that problem and does so in an obvious way. 
  
  
• As a comment, I don't know if there is a clear hint or teaching in the prior art to reduce the memory and computational requirements of a neural network by "loose couplings between the nodes in accordance with a sparse parity-check matrix of a [low-density parity-check code]". According to Wikipedia, "in information theory, a low-density parity-check (LDPC) code is a linear error correcting code, a method of transmitting a message over a noisy transmission channel", which does not immediately seems to pertain to neural networks (duly noting that IT is not my technical field). 

Claim 1 of the main request defines:

A hierarchical neural network apparatus (1) implemented on a computer comprising

a weight learning unit (20) to learn weights between a plurality of nodes in a hierarchical neural network, the hierarchical neural network being formed by loose couplings between the nodes in accordance with a sparse parity-check matrix of an error correcting code, wherein the error correcting code is a LDPC code, spatially-coupled code or pseudo-cyclic code, and comprising an input layer, intermediate layer and output layer, each of the layers comprising nodes; and

a discriminating processor (21) to solve a classification problem or a regression problem using the hierarchical neural network whose weights between the nodes coupled are updated by weight values learned by the weight learning unit (20)

or comprising

a weight pre-learning unit (22) to learn weights between a plurality of nodes in a deep neural network, the deep neural network being formed by loose couplings between the nodes in accordance with a sparse parity-check matrix of an error correcting code, wherein the error correcting code is a LDPC code, spatially-coupled code or pseudo-cyclic code, and comprising an input layer, a plurality of intermediate layers and an output layer, each of the layers comprising nodes; and

a discriminating processor (21) to solve a classification problem or a regression problem using the deep neural network whose weights between the nodes coupled are updated by weight values learned by the weight pre-learning unit (22)

and

a weight adjuster (23) to perform supervised learning to adjust the weights learned by the weight pre-learning unit (22) by supervised learning; and wherein

the weights are learned by the weight pre-learning unit (22) by performing unsupervised learning; and

the weights between the nodes coupled are updated by weight values adjusted by the weight adjuster (23).

EPO 

The link to the decision is provided after the jump, as well as (an extract of) the text of the decision.

T 0467/20 - Moved through the vessel of the patient

Key points

•  Claim 1 is directed to a system comprising "a first instrument sized [e.g. a catheter] and shaped for introduction into the [blood] vessel of the patient", a similar second instrument, and a computing device configured to "obtain pressure measurements from the first and second instruments while the second instrument is moved longitudinally through the vessel of the patient"
• The Board: "The contested patent relates to a system for assessing the severity of a stenosis in a blood vessel."
• The Board finds the claim to be sufficiently disclosed and inventive.
• Of note, Art. 53(c) is not discussed. Hence, a computer configured to perform calculations on output from a device that by way of surgery is introduced into the body of a patient is not excluded. 
• This decision appears in line with established case law, see e.g. T 2136/19. For a different approach, see T 0944/15.
• The present decision also illustrates the difference in the effect of Article 52(2) and Article 53 on inventive step. Features defining subject-matter excluded under Art. 53(c) can very well contribute to inventive step.
• " it is common ground that D1 does not disclose feature 4.3 of claim 1, whereby the computing unit is further configured to "modify the visual depiction of the vessel to simulate one or more treatment options based on expected results of the one or more treatment options on the pressure differential". " 
• This feature pertains to the presentation of information. 
• " The appellant [opponent] formulated the objective technical problem to be solved starting from D1 as being to facilitate the selection of a treatment method." 
• The Board seems to follow this formulation.
• " The Board first notes that D1 itself does not suggest that the "integrated graphical image output" may be used to simulate various treatment options. Without hindsight, the person skilled in the art proceeding from D1 alone would not implement feature 4.3 without an inventive step." 
• The distinguishing feature is the presentation of information. Interestingly, the Board does not comment on the case law that " A feature defining a presentation of information produces a technical effect if it credibly assists the user in performing a technical task by means of a continued and/or guided human-machine interaction process (T 336/14, T 1802/13 and T 1185/13). " (CLBA 10t hed. I.D.9.12.10.b).
• Implicitly, the method of treatment is a technical method and selecting the right treatment method is (apparently) seen as a technical problem, such that "facilitate the selection of a treatment method" is considered to be a technical problem.

EPO 

The link to the decision is provided after the jump, as well as (an extract of) the text of the decision.

T 0698/19 - Technical effect must be derivable from the application as a whole

Key points

• The applicant of the application at issue is Swiss Reinsurance Company Ltd.
  
•  Board 3.4.03 in the headnote: " If non-technical features have both a technical and a non-technical effect, the technical effect must be taken into account when assessing inventive step, but the technical effect must be clearly derivable from the application as a whole".
• This statement appears to combine two distinct rules: 1) "If non-technical features have both a technical and a non-technical effect, the technical effect must be taken into account when assessing inventive step" and 2) a  "technical effect must be clearly derivable from the application as a whole".
• The arguments of the applicant: " Prior to automation, the handling of insured loss cases was handled by an insurance agent who determined the amount of loss to be reimbursed based on similar claims and the insurance policy. In the 1960s, attempts were made to automate the handling of claims by a computer. For this purpose, the procedures were parameterized, which led to an efficient (but less accurate) claims handling. However, there were parts that could not be parameterized. The present procedure solved this problem by automating these non-parameterized parts through pattern matching with historical data by use of seamless integration. The invention represented the complete automation of an insurance premium payment. Automation was technical as such and had a technical effect." 
• The Board: " Automation is admittedly technical in itself. This is inherent in the very fact that the method runs on a computer. The board agrees with the Appellant that in the present case technical and non-technical features are interwoven and that generally non-technical features can have a technical effect. If non-technical features have both a technical and a non-technical effect, the technical effect must be taken into account when assessing inventive step, but the technical effect must be clearly derivable from the application as a whole." 
• " The Board is not persuaded either that steps (b) to (d) in combination  have a technical effect, i.e. improving the stability of the automated system of the invention, for the following reasons."
• "Disclosure as to how the algorithm is implemented in practice must be provided in order to give evidence that the algorithm has any proved further technical effect with respect to known algorithms and that it provides an improvement over the prior art. The present invention has the object to provide a stable system. However, as discussed above, no details are given why the proposed system should be considered to provide better stability than any other system, for example, how and at which level improved stability or accuracy of the algorithm is achieved, which kind of data is used for pattern matching, how historical data is selected, prepared and compared, and which parameters are matched. The only data provided in the application is pure economical data (see e.g. Fig. 3)."
• Note, " The system should be stable and "give basis to better investment grounds for partners and clients supporting the system" (description, page 3, lines 28-33)." The term 'stability' of the automated insurance system may refer, possibly, to financial stability, but the Board's decision does not comment on that point.
• "To summarise, any further technical considerations or effect must be derivable from the application as a whole and the claims must comprise the specific features which contribute to the further technical effect of the invention, ... "
• The Board adds: "... since it must be made clear to third parties what the technical part of the invention is, and the technically skilled person needs these details to carry out the invention (Article 83 EPC). This is not the case for present claim 1." 
• As a comment, it may also be said that an inventive step is to be based on features specified in the claim and the technical effect provided by those features (if any).

EPO 

The link to the decision is provided after the jump, as well as (an extract of) the text of the decision.

T 1108/20 - Closed-circuit e-commerce entity

Key points

• The title of the application in this appeal is "METHOD AND STRUCTURE FOR SIMPLIFIED CODING OF DISPLAY PAGES FOR OPERATING A CLOSED CIRCUIT E-COMMERCE"
  
• Claim 1 begins with:  "A method for connecting and operating a closed circuit e-commerce entity comprising at least one shopping terminal, at least one provider and at least one entity center" 
• The Board, however, finds claim 1 to be inventive. The decision may illustrate that non-technical features do not contribute to inventive step under the Comvik approach, but there is, strictly speaking, no penalty for specifying non-technical features under the Comvik approach either if the claim additionally recites non-obvious technical features.
• Claim 1 specifies that "said shopping terminal including a touch screen, a CPU and a second memory is connected with said building server via an interface including a CPU and a first memory"
• The Board: "the board considers that the "CPU" mentioned in distinguishing feature (ii) can only be interpreted as relating to an additional CPU, i.e. in addition to the regular CPU that each server - including the server disclosed in document D2 - possesses. The same applies analogously to the "memory" mentioned in distinguishing feature (iii)."
• "The objective technical problem is how to support a large number of kiosks without overwhelming the server. This problem is solved by distinguishing features (ii) and (iii), since adding a CPU and a memory for processing kiosk requests at the interface reduces the load on the server CPU. However, neither this problem nor its solution are disclosed or hinted at in the prior art at hand."
• "Therefore the board concludes that the subject-matter of claim 1 involves an inventive step."
  
  
• As to the description: "The board notes that the [allowed set of claims] basically corresponds to the claims originally published and that the description as published supports these claims. Hence there is no need to adapt the description to the present claims in order to meet the requirements of Article 84 EPC in the present case." (italics added)
  
  

EPO 

The link to the decision is provided after the jump, as well as (an extract of) the text of the decision.

T 0929/18 - Refusal without prior art search

Key points

• This is an appeal against a refusal decision.
• The invention is about "mobile location data sharing". The filing date is in 2009; the search report was established in 2009; the first Communication was issued in 2012; next action: summons oral proceedings in 2017, refusal decision in 2017. The Board's preliminary opinion was negative. 
• The search report did not cite any documents but rather stated that "the technical aspects identified in the present application are considered part of common general knowledge" (link). 
• The Board: "The appellant's request that a patent be granted on the basis of the main request, or one of the first to third auxiliary requests, cannot be allowed because no patent can be granted without a prior art search (Article 52(1) in combination with Article 92 and Rule 61(1)EPC)."
• The independent claims as filed indeed seem to broadly specify location sharing in a group by a mobile phone [user device] for a limited time, i.e. with instructions to share the information for a limited time and to stop sharing the information thereafter. 
• I note that in 2009, the previous version of Rule 63 applied, and Rule 137(5), second sentence, accordingly does not apply to the present application. However, as the searched claims recited technical means, such as a user device and an output interface, no declaration under R.63 is issued; instead, a search report not citing any document is established (OJ 2007, p.592; T1411/08; T1242/04 r.9.2)
• "The examining division refused the application for lack of inventive step without referring to any document. Essentially, they considered that the technical character of independent [method] claim 5 resided solely in the implementation of an administrative scheme on a notoriously known electronic data processing system"
• I note that the claims as pending at the time of the refusal decision were significantly amended compared to the claims pending at the time of the search.
• "The Board does not see a problem with the practice of referring to notorious knowledge in such cases; it has the obvious benefit of avoiding the need to prove facts that are self-evident or indisputably well known."
• "According to the Boards' jurisprudence (see e.g. T 1242/04 [...]), it is possible to raise an objection of lack of inventive step without documentary prior art if the technical features of the invention are notorious, i.e. so well known that it cannot reasonably be contested that they are prior art."
• "Claim 5 refers to mobile user devices in a "peer-to-peer" data sharing group."
• " In its communication accompanying the summons to oral proceedings, the Board had misgivings about the peer-to-peer aspect of the invention. Firstly it did not appear to be part of the inventive concept in the sense that it contributed to the problem that the application set out to solve. It also rather appeared from the way in which the application presented the invention that "peer-to-peer" was part of the background framework or at least technically irrelevant." 
• "At the oral proceedings, the appellant argued that the application did not describe the claimed aspects of peer-to-peer networking as known, and moreover that limiting the time period during which the devices shared data was particularly advantageous in mobile peer-to-peer networks."
• "The Board considers that, taken together, the features of the mobile device, peer-to-peer, group data and address arrangement and the relay or router [...] define details of an electronic communication architecture that cannot be dissected into non-technical and technical parts. "
• "Although the Board still has doubts about the significance of the peer-to-peer feature it would be unfair to judge such issues without proper prior art. This would be to put the cart before the horse by putting the substantive examination before the search (see e.g. T 1515/07 []). Indeed, a search may reveal prior art which puts the invention in a perspective that is different from what was originally set out in the application."
• " Also, there has been a lot of development in mobile communication networks since the filing date of the application. Thus, what is self-evident now might not have been so in 2009. Therefore, the Board judges that a search is necessary."

EPO 

The link to the decision is provided after the jump, as well as (an extract of) the text of the decision.

T 0909/14 - Upgrading user accounts

Key points

• This is an appeal against a refusal decision. "The invention concerns a web service enabling the creation, storage and sharing of documents (see published application page 1, lines 6 to 19). Such services are known from the prior art. The key idea is to allow new (prospective) users of the service to access only part of the available functionality in order to reduce resources consumed (page 1, line 30 to page 2, line 13).  Claim 1 of the main request recites a method for creating and upgrading user accounts. The service's full users, having an account of the first type, may create their own documents and edit documents shared by others. By contrast, a prospective user, joining the service in response to an invitation from an existing user, gets an account of a second type not allowing him to create his own documents, but only to edit shared documents (page 12, lines 8 to 31)."
• "The main point of dispute in this appeal is whether discriminating between full users and prospective users with limited functionality is the solution to the technical problem of saving resources, as argued by the appellant (see section X, above), or a non-technical requirement that could be envisaged by the business person and thus part of the problem given to the skilled person under the COMVIK approach (see T 641/00 - Two identities/COMVIK) as argued by the division."
• "Decision T 1463/11 - Universal merchant platform/CardinalCommerce, further refining the COMVIK approach, set out that in the assessment of inventive step, circumstances under which inventions are developed in the real world have to be ignored to a certain extent. It introduced the concept of the notional business person who may formulate business requirements to be implemented but will not include in them any technical matter. The notional business person is a legal fiction representing a shorthand for a separation of business considerations from technical ones. Using this legal fiction is the price paid for an objective assessment; a real inventor does not hold such considerations separately from one another "
• "Under the CardinalCommerce approach, the test used in deciding on technicality of a feature is whether the notional business person could have come up with it. This boils down to determining whether there would have been at least one way to devise it without technical considerations. If so, it does not have technical character and may form part of the requirement specification regardless of what alternative ways of arriving at it were disclosed in the application or were conceivable."
• "In the present case, the Board judges that providing limited functionality for some users, such as not enabling the creation of documents, is a business decision which need not be based on technical considerations. There are apparent business reasons for doing it, for example charging the full user a higher subscription fee than the prospective user."
• The present decision is interesting because G 1/19 does not refer to T 1463/11 though it did refer to the follow-up decision T 0817/16. G 1/19 stated that "It was argued in the course of these referral proceedings that a simulation is of a technical nature and has technical effects if the relevant skilled person is a technically skilled person in the field of the simulated system or process (see e.g. the comments of the President of the EPO, points 23 to 25). This argument is partly based on T 817/16 (see Reasons, point 3.12), which relies on the (technically) skilled person in order to distinguish between technical and non-technical features. This approach may be suitable in some cases but may prove difficult in others where the skilled person for the simulation is different from that for the system represented by the model underlying the simulation. " 
• " Concerning "simulating" the full user interface [as specified in the claims], the Board judges that including in the user interface a feature which is useless without an account upgrade, for example an empty document container, relates to the presentation of information and is not based on any technical considerations. The appellant's argument attempting to prove that this part of the solution derives technical character from effects produced when the user operates the upgraded user interface is not convincing. In addition to being speculative, it is a typical example of the "broken technical chain fallacy" in the sense of T 1670/07 - Shopping with mobile device/NOKIA, reason 11."
• The appeal was filed in 2014; the Board's first action was a preliminary opinion in 2020. The appeal procedure took 8 years. I don't know what the average is for Board 3.5.01. The prospect of having the patent application pending for 8 more years may of course precisely be the reason for filing the appeal, especially for this kind of software invention. In a way, the backlog at Board 3.5.01 may sustain itself. 

EPO T 0909/14

The link to the decision is provided after the jump, as well as (an extract of) the text of the decision.

T 0955/20 - (II) Better search results are not technical

Key points

• The appeal is against the refusal of a patent application with applicant Google.
• The Board, on inventive step: " Furthermore, even if distinguishing features (b) and (c) could be seen to plausibly improve the "quality" of the translated query in the sense that it leads to search results that are "better" in the sense of closer in meaning to the desired search results, this would not be a technical effect (see decisions T 1569/05, Reasons 3.4 to 3.7; T 598/14, Reasons 2.3 and 2.4)." 
• T 1569/05: " The appellants regard the invention as belonging to the technical field of utilizing a natural language as a search input. In their view, the invention makes a technical contribution in a field not excluded from patentability," " In the Board's view, neither the mathematical model of meaning according to D3 nor the modified model according to the invention are within the technical area, since only the meaning of the words determines how they are represented, stored and selected, and since mathematical algorithms completely define the processing. In this respect the present invention is similar to the case T 52/85 - Listing of semantically related expressions/IBM (not published in OJ EPO), where the deciding board held that automatically generating a list of expressions semantically related to an input linguistic expression is basically not of a technical nature but a matter of the meaning of those expressions, ie of their abstract linguistic information content." 
•  T 52/85: "Such a semantical relationship is basically not of a technical nature but a matter of the meaning of those expressions, i.e. of their abstract linguistic information content; it does not relate to any physical entity. A semantical relationship can be found by performing mental acts only, with no technical means involved. This does not necessarily mean that any system automatically concatenating, in place of a human being, semantically related expressions to a list is excluded from patentability. Rather, this will depend on whether the manner in which it is automated, involves features which make a contribution in a field outside the range of matters excluded from patentability under Article 52(2) and (3) EPC."
• "the appellant's observation on document D1 does not call into question the board's finding that the distinguishing features provide no technical contribution over document D1. Hence, the subject-matter of claim 1 lacks inventive step (Article 56 EPC)."

EPO T 0955/20 

The link to the decision is provided after the jump, as well as (an extract of) the text of the decision.