To print this text, all you want is to be registered or login on Mondaq.com.
In this choice, the European Patent Office refused to
grant a software program patent on an environment friendly spam filter. Here are the
sensible takeaways of the choice T 2147/16 of September 7, 2021 of
Technical Board of Appeal 3.4.03:
Key takeaways
An extra technical impact should be confirmed and particularly and
sufficiently documented within the disclosure of the invention.
The invention
The Board in cost summarized the invention of the applying
underlying the current choice as follows:
2.1 The current invention has as object to offer an environment friendly
spam filter. Spam messages current the inconvenience of
“cluttering” the consumer’s electronic mail field, such that the consumer
could simply overlook an essential message within the circulate of quite a few
spam messages. Spam messages take up a considerable quantity of the
electronic mail visitors. The share of spam consists of greater than 70% of world
mail visitors. The spam messages accumulate in consumer mail bins and
take reminiscence house and time to wash out (web page 1, line 10ff of the
description as filed).
2.2 Among the spam messages there are messages with laptop
viruses resulting in hurt of laptop {hardware}/software program and messages
which result in phishing sources that may be a reason behind theft of
passwords and private knowledge. Prior strategies of spam detection, which
sometimes contain grouping of messages by varied standards, usually
lead to false positives, i.e., when the spam message results in
the group of reliable messages or when the reliable message is
grouped as spam.
To overcome these downside, the applying suggests a particular
algorithm, inter alia together with steps of
hashing, clustering, weighting and so on.
Fig. 1 of EP 2 811 441 A1
Here is how the invention is outlined in declare 1:
Claim 1 (important request)
(A) A pc-applied methodology for classification of
digital messages being e-mail messages as spam or reliable on
a consumer mail server (200), the strategy comprising:
(B) receiving (510), by a pc processor (15) of the consumer
mail server (200), all digital messages (180) directed to
completely different customers who’ve electronic mail addresses registered on the consumer
mail server (200);
(C) classifying (515) the obtained digital messages (180) as
reliable or spam messages;
(D) figuring out (535) unknown messages that might not be
categorised as reliable or spam;
(E) acquiring (550) metadata (860) of the unknown messages,
(F) the metadata (860) together with at the least a set of hash sums
(830) for an unknown message and an IP handle of a sender of the
unknown message;
(G) inserting (553) the metadata (860) of the unknown messages
into one cluster of a plurality of clusters forming a cluster index
tree primarily based on a level of similarity between a newly-arrived set of
hash-sums {h1; h3 … hk} from the metadata (860) and present units
within the cluster index tree
(H) in accordance with: D = Sigma wj /okay ,
(I) whereby the sum is fashioned over all weighting elements wj
similar to hash sums within the cluster index tree which are
matched by hash sums within the newly-arrived set {h1; h3 … hk};
(J) ranking (590) every unknown message in accordance with a
ranking of the cluster the place the metadata (860) of the unknown
messages was assigned to,
(Okay) whereby the ranking of the cluster relies, at the least, on a
variety of related hash sums (830) of unknown messages obtained from
completely different IP addresses of message senders contained in mentioned
cluster; and
(L) classifying (560, 565) unknown messages as reliable or
spam primarily based on message rankings.
Is it technical?
The first occasion Examining Division refused the applying
because of the following causes:
(a) a spam filter didn’t serve a technical goal;
(b) it didn’t produce a related technical impact;
(c) it didn’t suggest technical concerns.
Besides some minor inadmissible modification and readability points,
the Board in cost was primarily involved with the evaluation of
technicality and creative step. In this respect, the Board
indicated that it agrees to the arguments offered by the Examining
Division that the claimed methodology, significantly the distinguishing
characteristic, lack technical character.
To defend its software, the appellant argued that the claimed
methodology facilitated an improved efficiency of a pc and
laptop community and decreased on the tip consumer laptop the reminiscence
quantity of the onerous disk required for the reliable digital
messages as a consequence of reducing an quantity of spam. Furthermore, the
appellant said that the strategy decreased a quantity of the e-mail
visitors to the tip consumer laptop and that it decreased a load for
the processor on the finish consumer laptop:
5.3.6 The Appellant argued that within the proposed algorithm reasonably
than processing all obtained digital messages on every particular person
consumer computing gadget, the invention used a consumer mail server.
Design and implementation of progressive algorithms and knowledge
buildings went past a specific mathematical formulation of the
process of classifying digital messages. The invention utilised a
cluster ranking system to acquire and analyse metadata of the unknown
messages and classify these unknown messages utilizing a cluster index
tree knowledge construction. The cluster ranking was dynamically altering
through the course of the filling of all clusters with varied
metadata of incoming digital messages. The invention analysed
mass messages despatched by varied sources over time and up to date cluster
rankings and due to this fact allowed to classify unknown
messages extra precisely. The methodology due to this fact facilitated
an improved efficiency of a pc
and laptop community. Therefore, the strategy achieved the technical
results talked about in part VIII above.
In response, the Board agreed to the appellants arguments in so
far that the claimed algorithm could also be optimized for the pc
{hardware} and will thus have technical character. However, this might
solely be of speculative nature:
5.3.7 The Board agrees in to date with the arguments of the
Appellant as the mix of evaluating digital textual content content material by
similarity preserving hashing and dynamic cluster ranking could also be
thought of an algorithm optimised for the pc
{hardware} and will have a technical contribution.
However, this mere assumption shouldn’t be
ample. …
According to the Board, to be patentable, an algorithm should
fulfill the next three standards:
(a) the implementation of an algorithm in a way for filtering
spam messages will need to have a proved additional technical
impact or particular technical concerns;
(b) such additional technical impact should
be particularly and sufficiently documented within the
disclosure of the invention and be mirrored within the
declare wording;
(c) the algorithm should serve a technical
goal.
Concerning criterion (a), the Board said that particular particulars
as to how an algorithm is applied in apply and the way the load
is lowered should be offered with a view to give proof that the
algorithm has any additional technical impact. However, the current
software didn’t achieve this:
5.3.11 The current invention, web page 19, traces 1 to eight, discloses
that the cluster measurement needs to be optimised with a view to cut back the
load on the pc. However, additional particulars usually are not
offered, for instance, the vary of the optimum cluster measurement,
related parameters, the quantity of reminiscence saved or the ratio of
elevated pace.
With respect to criterion (b), the Board arrived on the
conclusion that the applying failed to offer any particulars how
the alleged additional technical impact is achieved:
5.3.14 However, the current invention doesn’t
present ample and particular disclosure, akin to
parameters, how the algorithm is optimised for the pc, neither is
this mirrored within the declare wording.
5.3.15 The Board is of the opinion that any additional
technical impact needs to be specified and sufficiently disclosed in
the invention and that the claims should comprise the
particular options which contribute to the additional technical impact
of the invention.
Finally, additionally the third criterion (c) was thought of to not be
fulfilled for the reason that claimed methodology fails to serve a technical
goal:
5.3.18 The Board is of the opinion that, if the algorithm
claimed within the current invention will depend on the
preferences of the consumer, the aim of the claimed methodology
could also be thought of non-technical. Spam filters
on the whole want a human conditioning for coaching an algorithm
which form of emails ought to to be categorised as spam messages and
which form of emails needs to be saved within the mail field. This relies upon
on the person consumer preferences, …
Against this background, the Board concluded as follows:
5.4.2 According to G 1/19 (causes 121) algorithms
to begin with outline (non-technical) constraints to be
thought of within the context of the
COMVIK method (T 0641/00). Depending on
whether or not they contribute to any technical impact achieved by the
claimed invention, they could or could not in reality be taken into
account within the creative step evaluation.
5.5 Obviousness
5.5.1 As mentioned above a particular technical
impact associated to the algorithm as outlined within the
impartial declare shouldn’t be sufficiently documented in
the current software and isn’t mirrored within the declare
wording. The declare due to this fact defines a mere implementation
of an algorithm with none further particular technical
impact. This mere implementation of an algorithm
can’t be thought of involving an creative step.
At the tip, the Board dismissed the attraction as a consequence of lack of
creative step.
The content material of this text is meant to offer a basic
information to the subject material. Specialist recommendation needs to be sought
about your particular circumstances.
