Some legal livetweets coming up as we watch the Enlarged Board of Appeal on G1/19 live-streamed from the @EPOorg. (Possibly 1600 watching.) #g119

We are looking at the basic question: can you patent (computer) simulations in Europe?

The questions are referred from T0489/14 - subject matter: simulating the movement of pedestrians through a building. https://www.epo.org/law-practice/case-law-appeals/recent/t140489ex1.html

One of the issues is T1227/05 - this found that a claim to simulating noise when designing chips was patentable. https://www.epo.org/law-practice/case-law-appeals/recent/t051227dp1.html The Board in T0489/14 said they would need to deviate from T1227/05.

23 amicus curiae briefs were filed - https://www.epo.org/law-practice/case-law-appeals/eba/pending/g1-19.html.

The referral is admissible - at least in part.

The first part of the second question may not be admissible but the rest appears admissible.

The ELBoA think this may not be admissible - "2. If the answer to the first question is yes, what are the relevant criteria for assessing whether a computer-implemented simulation claimed as such solves a technical problem?"

Some legal gymnastics (as is common) to explain why the supposed "deviation" between BoA cases might not be as much of a deviation as suggested.

Looking at "functional technical features" - ELBoA are saying that T1227/05 does support both finding that a claim to a simulation has technical features and lacks technical features.

The Appellant makes a good point that if you require the simulation to be materially produced, you would exclude much of cryptography and machine learning. Using the trick of stressing the "engineering" in "engineering software".

First referred question tracks the language of T1227/05 (Infineon).

The Appellant is discussing T1265/09 (IEX) on call centre simulation - https://www.epo.org/law-practice/case-law-appeals/recent/t091265eu1.html and T1630/11 - https://www.epo.org/law-practice/case-law-appeals/recent/t111630eu1.html on multi-processor simulations.

Good point by the Appellant that guidance is needed regardless of whether a claim has 1) a preceding "measurement" step or 2) a following "control" / "material production" step. (This would be more Anglo-centric formalistic cop-outs.)

Now hearing from the @EPOorg on behalf of the President.

EPO also feels the referral is admissible. All parties say that both requirements of uniform application of law and fundamental importance required by Article 112 EPC are met - https://www.epo.org/law-practice/legal-texts/html/epc/2016/e/ar112.html.

Guidance on "technical" is not found in the statute - it must be found from the (Boards of Appeal) case law. #G119

"Simulation" appears to have caught a nerve with the EPO and the filers of the amicus curiae briefs - it's a Trojan horse for wider questions raised by machine learning and modern numerical methods. We might get some guidance or we might get a cop-out. #G119

Civil engineering is an interesting field as there is often a bias against it (patent people much prefer chemistry or hard-core electronics). Hidden in the referral is the question: is civil engineering "technical" enough? #G119

If you refer to "physical reality" then why is "noise in circuits" okay but "people in a building" not okay?

Hidden is that some people feel the latter is too "fuzzy" and not "hard" engineering. The referral is good because it brings these prejudices to light. #G119

EPO making some good points. The irony is that the maths behind "noise" or "people" is often the same! (Often differential equations, finite methods and/or digital approximations to the same.) So the only difference is the use case (circuit vs building). #G119

Enlarged Board now stating that only the Enlarged Board decides whether the Enlarged Board decides. #G119

Tea break! I'm enjoying this. #G119

We're back.

We haven't decided on admissibility but are assuming the questions (apart from 2a) are admissible. #G119

Computer simulation as such = method not including steps preceding or following (e.g. without explicit measurement or control). EBoA is following case law for computer-implemented methods - i.e. computer simulations as such may be considered as computer-implemented methods. #G119

Without deciding anything, the preliminary view is that the answer to 1) is positive (Yes).

Cool - we might have a discussion on how simulations/CIIs may differ from schemes for business - it would be good to hear this from the EBoA. #G119

EBoA introducing the distinction between "natural processes" and "human activities".

Trying to sync up "human activities" with "non-technical", similar to the case law on the non-technical nature of "cognitive effects". #G119

Where does "physical activity", i.e. actions in the world, split into "natural processes" and "human activities"?

We're getting into a bit of naturalist dualism here. #G119

We'll maybe need to wait until we get a case in 5-10 years on MRI/TMS/EEG machines to face the problem of "human activities" being "natural". #G119

Appellant saying it is difficult to define "physical entities" within "physical realities". Stresses simulation is a tool and that we should look at teleology - what is the technical purpose of the claim?

(Don't think the EBoA would like that.) #G119

Appellant refers to pre-COMVIK "Sohei" - think it's this one - T 0769/92 https://www.epo.org/law-practice/case-law-appeals/recent/t920769ep1.html. #G119

Nice reference by the Appellant to a "technical person" & their "knowledge" - who studies & devises the system helps with defining "technical".

Used a lot in practice but the Boards generally avoid making any statement on this - again I'll think the EBoA will sidestep. #G119

Some good points within some old classic computer-implemented invention (CII) arguments - what if the simulation says *not* to build a particular chip or building, surely negative decisions are as important as active positive decisions? #G119

Appellant citing case law to ask: Is it okay just to have "technical" "parameters" to have a technical effect? #G119

Appellant: no legal difference between virtual and actual methods - it all comes down to "technical", which can be applied the same in both cases, and is defined by the case law. #G119

Appellant: a simulation running on a computer running on data with a technical meaning in mind, with a technical purpose, is patentable. #G119

Appellant is stressing that graph theory applied to logic circuits is "technical" not just mathematical, and stresses that the some applies to certain simulations, and that a technical person from a field of engineering was needed for the solution. #G119

Appellant citing case law saying that modelling the weather is "not technical" as you can't change the weather. But Appellant makes a good point - what if the weather model is used to control window blinds and the claim is limited to this? #G119

Appellant stressing that the claim relates to improvement of the building structure, not just people milling around in a wood. #G119

Problem is how do you reconcile this "technical purpose" argument with later infringement? It would surely require the "technical purpose" to be within (& adequately limit) the main claim, e.g. if the claim is to modelling a building design then only this purpose infringes. #G119

The @EPOorg is stating that the COMVIK approach as set out within the Guidelines is stable and liked by all as a fair manageable way to evaluate computer-implemented inventions. #G119

Good point by the EPO - for a simulation to be "technical" , it needs to be useful - this is only the case if it accurately & realistically models a technical system and provides useful design information to an engineer. #G119

Such a simulation would include "technical considerations" if it reflected (at least) in part technical attributes of the simulated system. Evaluating the "technical" effect is independent of the form of the claim. #G119

This pings it back to the question: "is a building technical enough?".

EBoA could side-step by saying "pedestrians" = "human activities" = no patent.

#G119

EPO saying that it is better to define "technical" dynamically via the case law as what is an "invention" changes with time.

E.g. deep learning models were mainly archaic academic models until around 2011. #G119

EPO asking: if we omit the measurement step, do we need the simulation model to be obtained by measurement?

(I.e. this then provides the link to "physical reality".)

Assumption of EPO is that claim *does not* need to assume a model from measurement.

#G119

EPO also assumes we are *not* talking about any "technical effect" in the better functioning of the computer running the simulation. #G119

[Small tribute now being left at the shrine of COMVIK - https://www.epo.org/law-practice/case-law-appeals/recent/t000641ep1.html] #G119

EPO distinguishing between "potential" & "virtual" technical effect:
"potential" - occurs via use of the invention;
"virtual" - effects that are achieved in a computer that correspond to effects that would be seen if the modelled process was happening in the real world. #G119

EPO is saying "potential" technical effects are not generally to be relied upon unless we are talking about computer programs, where we are talking about the effects directly obtained when the program is executed. #G119

These "potential" technical effects seem similar to the Appellant's teleological effects (technical purpose). Likely that these will not be seen as provided a concrete technical effect. #G119

EPO saying that generally a virtual technical effect does not solve a technical problem. #G119

Now talking about direct and indirect measurements - latter where we take raw data and compute a quantity as opposed to measure it, the indirect technical measurements may then be used to solve a concrete technical problems. #G119

"virtual sensors"/"digital twins" provide these indirect technical measurements and can provide a technical effect - provides a real/concrete measurement but via different paths. #G119

EPO now discussing aircraft wing shapes - and how simulations of wind tunnels may be used instead of real wind tunnels to estimate and test aeronautical properties. Design of experiment is patentable, applies to both actual/virtual cases. #G119

In the above case - both virtual and concrete experiment designs are alternative solutions to the same technical problem - both should be patentable. #G119

Here the EPO is talking about simulations that directly provide technical data rather than just "data" that needs to be further analysed by a user to derive technical data. #G119

This is a nice position suggested by @EPOorg - in the general pedestrian modelling case you could say that you are not providing any concrete parameters of the building design as an output but if you were providing these it could be patentable. #G119

EPO - link to direct measurement / physical reality is not needed - this syncs with the development of "inventions" and increasing digitalisation. Key issue is whether a technical problem is solved and whether there is a technical purpose. #G119

COMVIK - mere fact that a feature relates to a specifically excluded field - https://www.epo.org/law-practice/legal-texts/html/epc/2016/e/ar52.html - does not mean it cannot provide a technical effect - need to consider the invention as a whole & the technical problem solved. #G119

EPO: But some exclusions are more exclusive than others. #G119

Examples of simulation "features": algorithms used, models used, computer-implementation, what is being modelled, purpose. #G119

EPO: models are closest to excluded "mental acts" - in the simulation method, the "model" solves a purpose, this purpose needs to be technical. #G119

EPO: logic to dismiss just adding a computer works for business methods but more problematic for mental acts - "paper & pencil" test not conclusive - as mental acts may involve technical considerations and may be replicating physical experiments & measurements. #G119

EPO: Sufficient condition: limit the object of the simulation to a technical system and that the output of the simulation provides technical information relating to that technical system. #G119

EPO: problem is more generic claims to simulations of "technical systems" - this requires more consideration of "technical purpose" - we need to see it in the claim.

(Good point that we need to consider withdrawn Infineon claims.) #G119

EPO: These could be closer to pure "scientific discoveries" that are excluded - they lack a "technical purpose" or fail to provide a "technical solution". #G119

EPO: in preparation documents the "as such" exclusion was deemed more to apply to "scientific discoveries" rather than "scientific theories" and "mathematical methods" - interesting that "mental acts", "mathematical methods", "scientific theories" can be treated separately. #G119

EPO: there is more leeway for claims based on (but not exclusively to) "mental acts", "mathematical methods", "scientific theories" that concretely solve a technical problem, e.g. as compared to "business methods". #G119

EPO: good reference to ray-tracing and rendering - this is patentable but doesn't necessarily directly relate to a real-world object that exists externally. Simulation here is the light effects - on the laws of optics - creates perceptual reality in the human brain. #G119

EPO is making some good points - clarifying that the technical effect can involve perceptive effects - need to call this out as it is at odds a little with the "cognitive effect" argument - the difference is sub-conscious/conscious awareness. #G119

EPO: in a wind tunnel case - what is simulated could have a technical effect as it may be new and contribute to the technical purpose; in other cases, what is being simulated may not have a technical effect but how the simulation is used may do. #G119

EPO: machine learning models avoid the need for models that incorporate (directly) technical principles - they are just trained - hence it needs to be careful not to exclude these new inventions from patentability - they are valid alternative solutions. #G119

A round of applause for that marathon last point - lots of useful and valid points there made by the representative of the President of the @EPOorg. A good eye on the nuances (for the office, applicants and practitioners alike). #G119

Back from lunch and on time. How many of the 1600 are left? #G119

Now looking at questions 2 and 3 and questions within a communication from the Enlarged Board. #G119

EB: Q 2b - In particular, is it a sufficient condition that the simulation is based, at least in part, on technical principles underlying the simulated system or process? #G119

EB: wishes to distinguish between the claimed simulation method and the simulated object in the real world. The former and its implementation is the focus. #G119

The communication the EB refers to can be found here - https://register.epo.org/application?number=EP03793825&lng=en&tab=doclist - document from 22 June 2020.

EB: initial answer to question 2b) is "no", with an emphasis on the fact that there may be cases where it is not sufficient and more is needed. #G119

EB: question 3 refers to a 4th auxiliary request where the claim covers an iterative design process. Initial answer of the EB is that the answer is the same as 1) ("yes")- it doesn't matter whether the simulation is part of the design process. #G119

Everyone loves aircraft wing modelling - Appellant independently presents an example similar to the EPO's. #G119

Everyone loves differential equations. Only engineers (and some weird economists) use these surely? Appellant submitting that a simulation of a technical system will be technical as it involves technical considerations [in an ML case these could be implicit]. #G119

EPO: agrees with general sentiment that Q3 is not really that important - the answers to the other questions apply whatever the form of the claim. [This is the most sensible answer.] #G119

EPO: if is it *not* a sufficient condition that the simulation is based, at least in part, on technical principles underlying the simulated system or process, then another condition could be that the output relates to technical properties of the system that is simulated. #G119

EPO: an example of this could be parameters relating to the performance of the simulated system. #G119

EPO: wrt Q3 - if the simulation claim as such does not meet the sufficient condition(s), if that claim was embedded in a wider claim (e.g. to visualisation to assist a user), would that then fulfill the sufficient condition(s)? #G119

Debate is now closed for a short deliberation - then the Enlarged Board will come back (maybe re-opening debate with questions or the Oral Proceedings will be closed). #G119

It looks like the aim of the Enlarged Board is to hear these oral arguments and then go away and produce a decision off-line - but we know that their inclination is:
Q1) Yes.
Q2a) Not admissible.
Q2b) No - you also need a "technical" solution.
Q3) GOTO Q1.

#G119

This basically says: claims to computer simulations are allowed and they are to be treated just like any other computer-implemented invention. #G119

So far - TLDR: can you patent (computer) simulations in Europe?

Yes - they are no different from other computer inventions.

[We'll see if the EB add anything when they come back.] #G119

EB: All issues have been suitably discussed. We'll let you know. Decision will be in the post. #G119

Well done to @EPOorg - technically (no pun intended) went off without any hiccups - clear video, no buffering and good sound. Great for access to the legal process and for wider engagement. #G119