Google knows better your language than you!



Google announced this summer an interesting update on Google Docs during the Google Cloud Next ’18 conference: from now on you will benefit from not only a spell-checker, but also a grammatical one! 1

The operation of the device is simple: when a possible grammatical mistake is detected, the portion of sentence is highlighted in blue. By right-clicking it, you can then accept one of the displayed proposals, or to refuse them all. You can also write your whole text and then correct all the faults using the “Tools” menu.2

We have to highlight there that if the result is as high as the expectations, it will be a huge progress in the field of natural language processing. Indeed, to build a spell-checker is not very complicated in itself: you just have to use a dictionary for that. If the program knows in which language you write (and computers manage to detect the dominant language of a document from its first written words), it loads the corresponding dictionary and compares all the words with this dictionary to know if what you write is correct or not. Well, OK, in practice, it is a little more complicated: it is necessary to be interested in the formal rules of agreement, but as these rules are formal, computers are rather good doing this job nowadays, as you can check by yourself using any office equipment.

But there, that goes farther: the detection of faults of structures in sentences requires that the machine has a good understanding of the language. Bye the bye, that’s the reason why Grammar Suggestions will use machine learning to progress. You can moreover volunteer to teach the program by making a request for Google and by using a beta version on Google Docs.3

At the end, when Grammar Suggestions will have grown up a little bit, it will write better than us…

Go further / usefull links


1■ Article taken from the Journal du geek, “Google Docs se dote d’un correcteur grammatical basé sur l’intelligence artificielle“, 18/07/25.

2■ Article from the G Suite Updates blog (Google), « New grammar suggestions in Google Docs launching to Early Adopter Program », 18/07/24.

3■ Take part in the tests



Expenditures and testosterone

The results of a study1 led by Gideon Nave, professor in Wharton School of the University of Pennsylvania, showed that the testosterone’s rate influenced men purchases2.

The initial hypothesis, which has been confirmed by the results of the study, is that the male slumbering in us tries on one hand to attract the eye of the females in the vicinity (like the bird who shows his most beautiful plumage to be spotted by potential sexual partners) and on the other hand to discourage his potential rivals, which are the other males circulating in the surroundings … You know, like dogs which raise the leg to urinate as high as possible, just to persuade the others that they are very big and thus very sturdy!

Well, the idea is that with man, the brands of the products which he raises can play this role.

The idea of the study is rather direct: products of the same quality but with different brands of which the prestige is different were proposed to a group of 243 men from 18 to 55 years old, half of them having beforehand received a dose of testosterone, the others a dose of placebo. And, guess what, the men who took a dose of testosterone were more sensitive to the brand than the others!

Well, then of course, the experiment is finer than that (the prestigious level of the brands was estimated before the experiment by a poll, the saliva of the participants has been checked to see if other hormones did not influence the tests, they were forbidden to have a contact with a woman just before the examination etc.).

It is a remarkable result, isn’t it? In any case, it helps us to understand the advertisings for fragrances or cars which, rather than to speak to us of the product, prefer to show us images which make raise our testosterone’s rate!


1■ Nature Communications, n°9, Article n°2433, “Single-dose testosterone administration increases men’s preference for status goods“, publié le 3 juillet 2018

2■”Testostérone rimerait avec achat premium“, article taken from Stratégies, 4th July 2018


DeepFake special FX are more and more frightening

It’s been a while since we know how to reproduce somebody’s voice with a speech synthesis software1, in order to make him say things he didn’t say… But now on we can also create or modify videos in that way (in real time), so we can see the guy telling things or doing things… he never told or did! And it’s impressive how the result is realistic for the human eye. In the domain of Deep Video Portraits, Michael Zollhöfer is very implicated in the current innovations, as he is, for example, a member of the HeadOn project.

We found a number of videos on this rather disturbing topic… Thus we share them with you:

Deepfake Videos Are Getting too Good :

‘HeadOn’, An AI That Transfers Torso, Head Motion, Face Expression And Eye Gaze :

Generate a new video of a completely different person’s body performing those actions :

FRENCH / Les « deepfakes », savant mélange de « deep learning » et de « fake news » :

Bonus / Age manipulation by video transformation (high level!) by Rousselos Aravantinos:

Notes (FR)

1■ Article du journal « Le Monde » du 02/04/2017 « L’appli qui imite les voix »

Make cannabis-based medicines legal, say UK drug advisers



Doctors in the UK should be able to prescribe cannabis-derived medicine, the government’s chief drug advisers have recommended, paving the way for a loosening of the laws governing access to the substance.

Cannabis is classed as a schedule 1 drug, meaning it is thought to have no therapeutic value and cannot be lawfully possessed or prescribed. It may be used for the purposes of research but a Home Office licence is required.

“At present, cannabis-derived products can vary greatly in their composition, effectiveness and level of impurity. It is important that clinicians, patients and their families are confident that any prescribed medication is both safe and effective.”

The ACMD has tasked the Department of Health and Social Care and the Medicines and Healthcare products Regulatory Agency with producing a definition for the products that could be prescribed.

The council also recommends that clinical trials urgently take place to further establish the safety and effectiveness of different products.

Read complete article on TheGuardian

Get more :


The addiction in the video games: a disease recognized by the WHO!



As reports the « Journal du geek » newspaper1, a new milestone was put in the debate concerning health and video games. Indeed, the WHO registered the addiction on the video games on its list of diseases.

If this piece of news can relaunch the debate, because seeming to agree with those who think that video games are a danger for society, Shekhar Saxena, who manages the department of mental health and drug addiction of the WHO, does not take a stand in this debate. Indeed, he declares to the AFP (french-language international news agency), standing neutral, that ” the WHO does not say that any habit to play video games is pathological “. According to the WHO, the addiction is defined by a ” loss of control over the game ” having ” harmful consequences “, as the relinquishment of ” certain activities “, ” like the sleep and the meal “.

Then the WHO doesn’t want to stigmatize video games. Moreover, numerous doctors think that the concerned players are urged by other reasons to be interested in the games; the dependence would be then a symptom of their disease, and not the disease itself.

But to understand the stakes in this announcement, it is already necessary to understand what it means. And the WHO calls back on this matter that its classification of diseases has for objective to notify countries and to help them to make decisions on ” the allocation of resources for the prevention and the treatment of the pathology “. Then what we need to understand, is that the addiction in the video games is becoming a social problem! Not only some individuals here and there, but a big fringe of the world population!

And if this problem emerges only now, it is probably because the domain of video game largely evolved these last years. The editors managed indeed to set up practices which urge the players to play more and more for a long time. As for example the random improve of the characters abilities during the time spent to play. This practice common to all the role player games (RPG) forces literally the player to stay in front of its screen to be able to move forward in the game, inducing a practice called “farming” (the name of which referred to the relation set between the player and his character). Or still the creation of big universes, games in which the scenario is not linear any more and encourage the player to explore the world of the game. Or still, on the contrary, to shorter the cycles of the games and to eliminate the plot aspect of the game (like that could be the case in arcade games, where it was necessary to pay at the beginning of each and every reload of the game).

This last practice, in particular, is in adequacy with the evolution of the materials used to play. From now on, a big part of the games (and thus players) are enabled on portable systems: smartphone or tablet. What it means, is that everyone can have access to a game in only a few seconds, as soon as a down-time appears during the day, because from now on, everybody lives with his smartphone.

But the question is probably wider than the video games addiction. Wouldn’t it be, more fundamentally, about an addiction to technology and to screens? For the record, in 2014, even the Prime Minister of the French Government was forced to prohibit mobile phones during the Council of Ministers2

Look further / Useful link(s)


1■ Journal of geek : WHO considers video games addiction as a disease

2■ Figaro magazine : Ministers leaved without their smartphones during the meetings


The love at first sight science



In 2017, scientists managed to observe and produce the loving mechanism to voles of prairies1. Before going further in this article, it is necessary to know that this species (Microtus ochrogaster) is curiously an excellent choice to model the human loving behavior. Indeed, voles look like us emotionally talking: they are monogamous but have sexual interactions with other partners than their other half, they become aggressive in the presence of other individuals of the same sex and depressed when they lose their partner. Furthermore, the prefrontal cortex of voles and humans can modify the behavior of the deep layers of the brain.

That’s what happens to the voles when they fall in love… and thus that’s probably what happens to us either in such a situation. The neuroscientists of the Emory University of Atlanta were able to observe that the loving feeling coincides, in the brain of voles, with the takeover of the accumbens core by the prefrontal cortex. In other words, the center of reflection and decision suddenly takes over the center of pleasure while the subject falls in love. It would thus be the intellect which would be at the origin of the Love feeling! And it would be it which would make us perceive the presence of our partner as a reward.

The same team was able to verify it: having introduced photoperceptible genes into certain neurons of a female, they put her in a cage with an unknown male (and they made sure to prevent any physical contact between them). Then they activated the neurons of the prefrontal cortex of the female with a frequency determined during preliminary phases of observation, before the experiment. Then they placed the female in an environment containing various males: in 10 cases out of 12, the female then preferred the male that was chosen by the scientists.

Even if this behavior are not directly transposable to a human (whose brain is more complex), this experiment lights us on the way we work. And, potentially, this knowledge can help us treating disorders as autism, which corresponds to a difficulty in creating social links.

Go further

1■ Science et vie, august 2017, “Les secrets de l’amour enfin révélés” [“Science and life”, “Secrets of loved finally uncovered”]

Data sciences chair opened at the Collège de France


In January, 2018, a chair of data sciences was opened at the Collège de France by Stéphane Mallat, French researcher who had worked out in 1987 an algorithm at the origin of the future JPEG2000 format before creating a start-up producing chips for TV which allowed to improve the resolution of the image (production of a high-resolution image from a standard signal), after which he interested himself in deep-learning algorithm for problems connected to the automatic recognition of images.

In an interview for the newspaper La recherche1, he explains the reasons of the creation of his chair at the Collège de France and gives indications onto the contents of the classes he gives.

For him, it was important that the name of chair was ” Data Sciences ” with the plural form because it is a multidisciplinary field of research. Indeed, although the used tools are always the same (applied mathematics, IT mathematics and IA, information theory etc.), the handled sets of data concern any sorts of sciences (physics, biology, cognitive sciences, economy, social sciences, etc.). Yet each of these sciences have their own approach to the problem of big data, what makes it a massively multidisciplinary domain. Moreover, Stéphane Mallat supports that the emergence of this discipline is not due to a scientific necessity, but rather to a social and university pressure, because these methods are deeply revolutionizing our societies (like chairs of computer science were imperative in universities, fifty years earlier). In fact, the actual pressure is such as Stéphane Mallat works at present on the opening of another chair of data sciences at the Ecole Normale Supérieure (a French school to form searchers). As a result, and as the domain just begins to crystallize, the main objective of his chair at the Collège de France will be to create a common vocabulary for all the concerned scientific disciplines, to describe problems connected to large-dimension data. In other words, to put the bases of this new science by creating a new vocabulary.

Historically, if we want to understand where from comes this emergent domain, it is because of the accumulation of the data and the increase of the computing power that the applied math and the computer science met to give birth to machine learning. Because historically, we were first capable of storing a large amount of data, before knowing what we could do with it. Globally, the data sciences are used to reach two types of objectives: the modelling of a set of data (to generate new data, compress data, reconstruct or improve the quality of an image etc.) and the prediction (which consists in giving meaning to a set of data). At present, deep learning techniques work well for these uses, but we fail to understand why. Thus it is a whole domain of research to understand this, in order to make them more reliable for critical applications such as medical diagnosis or autonomous cars. Other areas of research concern the reduction of the number of dimensions of the problems by discovering and using multi-scales hierarchies (observation of the data with various scales) and symmetries (invariances) in the handled set of data.

And because we have on one hand data warehouses and on the other hand quite an arsenal of applied math to use, one of the distinctive characteristics of this domain is that it is at the same time theoretical and experimental. It is indeed, according to Stéphane Mallat, from empirical approaches and remarkable intuitions of several researchers and engineers that were born the recent and sudden progress we know in the techniques of visual and vocal recognition, machine translation or still in Go and chess games. And it is for him the experimental search in this domain which brings to the foreground new mathematical problems; and that is why this correspondence between math and application is at the heart of his classes.

Go further


1■ La recherche, february 2018


Elon Musk wants to create a website to note the journalists and the media


Some billionaires want to buy media. Elon Musk wants to rate journalists’ credibility.


“Going to create a site where the public can rate the core truth of any article & track the credibility score over time of each journalist, editor & publication,” Musk tweeted.


“The last thing we need is another rich and powerful dude threatening to silence any journalist who doesn’t see things his way,” Timothy Karr, senior director of strategy at advocacy group Free Press, told CNN. “The implication in Musk’s actions are that all news media are untrustworthy. That’s a shameful message to be spreading.”

Could be true. But we’d like this Timothy Karr to say the same about Soros, who is much more Dangerous. Curiously, we think that he wouldn’t.

Musk’s idea is rather a good one. More and more people realize that Press writes less and less Truth, and behaves like a political censorship. We have the same problem in Europe.


Some colors boost your brain

Existing research reports inconsistent findings with regard to the effect of color on cognitive task performances. Some research suggests that blue or green leads to better performances than red; other studies record the opposite. Current work reconciles this discrepancy. We demonstrate that red (versus blue) color induces primarily an avoidance (versus approach) motivation (study 1, n = 69) and that red enhances performance on a detail-oriented task, whereas blue enhances performance on a creative task (studies 2 and 3, n = 208 and 118). Further, we replicate these results in the domains of product design (study 4, n = 42) and persuasive message evaluation (study 5, n = 161) and show that these effects occur outside of individuals’ consciousness (study 6, n = 68). We also provide process evidence suggesting that the activation of alternative motivations mediates the effect of color on cognitive task performances.

Read it in science mag.