We've been waiting for this moment since Tensorflow was open-sourced.



I have been struggling with 5 different versions of Bazel and lots of code patches in Tensorflow to make it work on my humble machines (because Cuda capability 3.0 is not official supported, I will write about the final distilled steps in a future post), but I was terribly disappointed when I discovered that distributed computation was not supported in the initial release.
Fortunatelly, some Spark presentations and github repositories pointed to this feature to be supported soon.
And finally, today, an initial version of the distributed runtime, based on gRPC, has just been pushed into the official TensorFlow repository.

It's very promising...

Actualization (13th April'16): Official distributed version has just been released (release candidate).

More good news !

We have a new paper accepted in Eurographics 2016 Conference !!
Modeling and Estimation of Energy-Based Hyperelastic Objects.

It has been listed in the unofficial Ke-Sen Huang's Home Page already :-D

Inventor	Granted patents (OEPM)
RODRIGUEZ FERRE,JOSE MANUEL	243
GUERET, JEAN-LOUIS	230
AMMERMANN, EBERHARD	222
LORENZ, GISELA	187
THEURER, JOSEF	166
MUR GIMENO,EMILIO	159
CORMA CANOS,AVELINO	138
WOBBEN, ALOYS	136
SAUTER, HUBERT	132
BIEDERMANN, LUTZ	127

Jose Manuel Rodriguez Ferré, son of the founder of Feber and former manager of his company, stands remarkably with 243 patents and utility models. Unlike the inventors listed in the world top list in the Wikipedia, that appear to sign every single patent their engineers do for their companies (and, therefore, they share the applications with a large number of people), Jose Manuel Rodriguez Ferré is the only person in (most of) their applications. So, he is a particularly meritorious inventor because of the high number of patents and utility models as independent inventor.

In my experience, I know that individual patents require much more effort than the ones that are devised and drafted together.

Another aspect that draws my attention (and we discussed a little later) is the low proportion of Spanish inventors at this "top" part of the ranking. It is true that the world is much bigger than our country and that many multinationals are interested in protecting their inventions here, so it is expected to find inventors of all nationalities in this ranking, but I find our contribution particularly insignificant.

Moreover, it is especially disturbing when you consider that the process of creating patents is slowing down in Europe, while it speeds up in the rest of the world (every time I consult the databases of patents I get flooded with Chinese patents).

The structure of patentados.com webpage is regular and easy access (and faster than OEPM/SPTO website), so I webscrapped it to generate some graphics to help me understand the phenomenon of the invention in Spain.

The following chart can be a bit counterintuitive. First, you must notice both axis are in logarithmic scale, that is, every time a portion advances we increase an order of magnitude. Thus, the initial portion of the figure (the "top" ranking) is actually populated by a very small portion of total 28266 inventors.



As shown in the graph, only 25 people have 100 or more patents registered in Spain.

Only inventors with 6 or more patents are listed. Again, I stress the difficulty of seeing the "long tail" because of the way I draw the graph. For example, the number of inventors with a maximum amount of 8 to 6 patents is 15498, it is represented by three small steps of the green "stair". An uniform sampling would be less readable, especially in area of the most prolific inventors, they would have been relegated to an insignificant region at the corner of the figure.

As discussed above, it shock me to see only a low percentage of Spanish inventors at the "top". The figure below shows the percentages for the first 10, 25, 50 and 100 inventors in the ranking of the website. I'm afraid it is a constant for every section of it, employees of multinationals monopolize our rankings.



Could this possibly be the reason why these data is not directly made public? I do not know.

If someone is embarrassed, just invent!, there are lots of challenges to face. Crisis time is an opportunity time. Although, it seems most of people prefer to sit comfortably in front of the TV while watching football and to follow Unamuno blindly.

This last graph shows the number of inventors in function of the number of patents (and utility models) they have been granted in the SPTO / OEPM, without distinguishing whether they were obtained by passing the IET and the exam.



As can be seen in the overlapping notes, the number of inventors with more than 25 patents does not reach 1200. If we consider that about three quarters are foreigners working in companies that have extended their patents into national territory, I think there would be around 400 Spanish inventors with 25 or more granted patents. Not even  200 inventors reach the 50 patent "plateau".

In the Wikipedia entry about prolific inventors, it is indicated that the average number of patents by inventor in US is 3, so the threshold for considering a inventor as "prolific" is situated at least 5 times at that amount: 15 patents. Tens of thousands of US inventors could fit that category. In Spain, only 4085 people reach the 15 patents granted or utility models mark.

So, where should be located the threshold to be considered as a "prolific" inventor in our country? What is your opinion?