Google AI beats experienced human players at StarCraft II

An artificial intelligence (AI) known as AlphaStar — which was built by Google’s AI firm DeepMind — achieved a grandmaster rating after it was unleashed on the game’s European servers, placing within the top 0.15% of the region’s 90,000 players.

StarCraft II’s complexity poses immense challenges to AIs. Unlike chess, StarCraft II has hundreds of ‘pieces’ — soldiers in the factions’ armies — that move simultaneously in real time, not in an orderly, turn-based fashion. Whereas a chess piece has a limited number of legal moves, AlphaStar has 1026 actions to choose from at any moment. And StarCraft II, unlike chess, is a game of imperfect information — players often cannot see what their opponent is doing. This makes it unpredictable.

Read article in “Nature“.

AI – police wants it to stop violent crime before it happens


Police in the UK want to predict serious violent crime using artificial intelligence, New Scientist can reveal. The idea is that individuals flagged by the system will be offered interventions, such as counselling, to avert potential criminal behaviour.

However, one of the world’s leading data science institutes has expressed serious concerns about the project after seeing a redacted version of the proposals.

The system, called the National Data Analytics Solution (NDAS), uses a combination of AI and statistics to try to assess the risk of someone committing or becoming a victim of gun or knife crime, as well as the likelihood of someone falling victim to modern slavery.

West Midlands Police is leading the project and has until the end of March 2019 to produce a prototype.

Read article on NewScientist.




They said: ” The one who won a golden coin won a golden coin, but lost a hand. ” (Implied: the one hand which hangs on the coin and which is thus no more usable). We could also say: ” These objects that you bought, are you possessing them, or are they possessing you? ” Strangely enough, this philosophical interrogation seems nowaday to be a concrete one …


We all know that our invaluable personal data, those whom we agree to share with big companies which deal in it, are then used to categorize the human race, study it, analyze it, and finally plan, anticipate and influence the decision-making of any kind of person. This new transparency man has been highlighted by Olivier Ertzscheid in his article ” The man is a document as the others: from the World Wide Web to the world wide life.»1


What we have to understand is that this collateral damage is inseparable with technical progress. And this stands for several reasons. The first one is simply that a useful tool is a tool which meets a need, and that a tool which you’re going to use is a tool which is not going to need lot of time or energy from you to be used. In other words, an “intelligent” tool is a tool which has to know you better than yourself. To know your needs and to understand how you express them. Simply, if we try to meet your needs in a precise way and before you waste time and energy to formulate them in a high level language it is necessary to know you precisely. This condition is not negociable. Thus it is required there to amass an impressive quantity of information on what you are.

The other point is that in a general way intelligence is something collective. We notice it in all kind of domains: Isaac Newton said ” If I have seen further, it is by standing on the shoulders of giants. “2 speaking about his predecessors and, doing so, quoting Bernard of Chartres. Eric Raymond, co-creator of the “Open source” term, also leans on this idea to express the superiority of the Open source system on the Proprietary software one in his essay ” The Cathedral and the Bazaar “3. Finally, we also know from now on that a mass of chess players of varied quality manages to compete in a party with very strong players.4 In conclusion, if machines can surpass man, it is because man has only limited exchanges with his congeners, when machines can benefit from the experience of all almost without any limit. Thus, the more the machine is effective the more it has to have access to a wide source of data (which means of experiences).

Then we have to go with this issue: a tool helps you, but to improve itself, it has to steal information on you. Then: who’s possessing who? Could this be a symbiosis?


The issue here is that even if there is symbiosis, this knowledge of the man is problematic. As long as the technologies are there to help man, everything is ok. But what if they were used to destroy him? …

But nowadays, with the proliferation of connected objects, the issue is no more the simple trust, but the direct security of devices: what if the rat was hacked? It is moreover for that reason that the doll toy Cayla was removed from the market in Germany: this one recorded the voices of children and transmitted them to the servers of the manufacturer. The authorities worried that the information could be intercepted in hostile purposes.5

After all that, what do you think? Do you possess your objects or do they possess you? Or maybe other people possess you through your own objects?



1■ Olivier Ertzscheid, « L’homme est un document comme les autres : du world wide web au world wide life », Hermès, La Revue- Cognition, communication, politique, CNRS-Editions, 2009, pp.33-40

2■ Wikipedia

3■ Eric S. Raymond, “The Cathedral & the Bazaar”, 2010

4■ Par exemple : The world vs Arkadij Naiditsch

5■ 01net magazine, 17/03/08

Back on top of the article

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



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 »

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


Brain decoder for your inner voice


As you read this, your neurons are firing – that brain activity can now be decoded to reveal the silent words in your head…

The team found that certain neurons in the brain’s temporal lobe were only active in response to certain aspects of sound, such as a specific frequency. One set of neurons might only react to sound waves that had a frequency of 1000 hertz, for example, while another set only cares about those at 2000 hertz. Armed with this knowledge, the team built an algorithm…

It’s an already old article but still very good to read HERE.

What is Blockchain (Bitcoin etc…)

This is a  article from the blog CoinCentral (with their authorization).

What is blockchain?

Simply put, a blockchain is just a list of digital records (blocks) that are chained together using cryptography.

The financial industry today contains all types of middlemen – payments processors, banks, and credit card companies are just a few. These intermediaries help to establish trust between buyers and sellers and ensure the accuracy of data in the transactions. However, adding additional people and steps to the process oftentimes leads to cost increases and reductions in speed.

Enter blockchain. This technology eliminates the need for middlemen by providing a decentralized, trustless ledger system with little exposure to fraud. Bitcoin is the most famous example.

Although you primarily hear about the financial sector’s use of blockchain, the technology expands far beyond just simple transactions. Blockchain companies are disrupting tons of industries from data storage and supply chain to gambling and the Internet of Things.

How does blockchain work?


A blockchain is run by a large network of computers, called nodes. These computers validate and record transaction data on the network by solving complex mathematical algorithms.

Every node has a complete history of transactions, so if one were to try and maliciously change a record, the entire network would know and reject the change.

Transaction example

Bob wants to send Sally $5, so he submits his transaction to the blockchain. Every node in the network then receives his transaction request.

Each node checks for two things with the transaction data:

  1. That Bob is who he says he is
  2. That Bob has the $5 to send to Sally

First, the nodes check Bob’s identity using the private key that he provides. A private key is an ownership tool that identifies a source of funds.

Next, the nodes make sure Bob isn’t trying to spend money that he doesn’t have. Because the nodes all have a copy of the ledger of transactions, they can easily check whether or not Bob has the $5 that he’s trying to send.

If at least 51% of the nodes agree that Bob’s identity is truthful and he has enough money to send, then the transaction will go through. The nodes will also update the ledger on the network with the new transaction.

With each new transaction added to the chain, the previous transactions become harder and harder to manipulate.

This immutability is supported by hash pointers. A hash pointer is a cryptographic hash that refers to the previous data block in the chain. They allow you to confirm that no one has tampered with earlier transaction blocks.

A transaction becomes like a fly trapped in amber. You would need to remove each additional block (layer of amber) from the chain to access and manipulate the previous transaction data (the fly).

The benefits of blockchain


Cutting out the middleman shortens the process of transactions and data transfer. Validations are inherently built into a blockchain system, so there’s no need for lengthy approvals or complicated record checks.


As mentioned earlier, the more people and entities involved in a process, the more costly it becomes. The cost of running a blockchain network is far less than an intermediary doing the same job.

More accurate

Data on the blockchain is immutable and validated by mathematical computations. It’s nearly impossible to have any human error and/or fraud.

Blockchain TL;DR – Final Thoughts

A blockchain is a distributed ledger system that uses cryptography to link together bits of data. It removes the need for middlemen in transactions which leads to faster processes, reduced costs, and greater data accuracy.

Go further:

Back on top of the article

Privacy news : google and iOS 03-2018

1) Google logs

Google is also giving customers ways to verify what it’s up to in case they don’t trust Google enough to take the company at its word.

For instance, Google is expanding Access Transparency by offering customers a real-time log of any accesses to their data by Google’s own engineers and support staff, as well as justifications for that access. In doing so, Google is giving enterprise customers a way to make sure that its employees aren’t poking around when they shouldn’t be.

2) Siri on your locked phone

With iOS 11, Apple added a new setting that lets you choose whether you want previews of your notifications to appear on your lock screen. By default, iOS shows a preview of your notifications only when your phone is unlocked, via some form of authentication like Face ID. But Siri will read your notifications from third-party apps aloud even if your phone is locked. This means anyone with physical access to your phone could hear messages meant just for you. MacMagazine first reported the issue after one of its readers noticed the peculiar behavior.


Big brother is hearing you!


« Keep your friends close, but your enemies closer. » A wise sentence at the godfather’s time, but we shall note that Don Corleone didn’t know about smartphones… If smartphone became man best friend these few last years, it is also becoming his worst enemy … The cause? It dwells much too close to you! You use it as an intermediary in all that you make: when you communicate with other people, visit Web sites, take photos or videos … It never leaves you alone, and thus, a big part of your life passes through it. What makes it an informer of choice for whom wants to know what your life looks like.

Maybe you think that you’re not so important, that your common life couldn’t be interesting for whoever, but that’s very wrong! Advertising executives, sellers of all kinds, and even government intelligence agencies of numerous countries are interested in your private life …


That’s why we wrote this article. They noticed recently that several smartphone’s applications listened to us without knowing it. And when we say “listen to”, we do not talk about the history of your browser, but about your smartphone’s hot mic! Among them, unsurprisingly, the Google1 application… It naturally collects all your vocal requests and this is why you use it. But did you already wonder how Google knew that you had said ” OK Google “, informing the application that your next sentence would be a request? There is only one possible answer: well, Google has to permanently listen to you, to know when you are going to pronounce ” OK Google “! This is the way the firm records your phone calls, your private conversations and in a general way all the sounds that reach your smartphone’s mike. That’s hard news. But on the other hand, the application is very forced to listen your mike to function, and you know that. Then OK, recording all of this instead of erasing could be considered unfair, but not unexpected. Just you to know, Eric Schmidt, CEO of Google, had said in 2009: ” if you make something and if you want that nobody knows it, maybe you should begin by not making it. “!

Google allows you to see what was recorded on you, and even to erase these contents … At least in theory (because before being erased, the data could have been be stored in another server designed to Google’s internal use, or simply because they were hacked before being erased!). You can see your history and delete it via this link (if you are not connected to your Google account, the page will be empty):

Let’s remember here that every good IT specialist will confirm you that on the Internet, to fully erase something isn’t completely guaranteed!

In brief, all this to say: think before trusting technologies. In that case, you directly authorized Google to listen to you all day and all night long!!

However, things can be more pernicious: imagine that approximately 250 Android games2 does the same as Google! Well, then, officially, not exactly the same: the responsible code for the leak of data is thought ” to detect television programs as a background sound, and not your private conversations ” says the Alphonso company, which developed it. They say that the user is warned of the use of his smartphone’s mike: you know, during the installation of an app, the list of the demanded access rights which you never read … The New York Times specifies that if the general conditions mention an access to the microphone, they do not specify it’s use nor that this one will continue to be listened to when the application will be closed and even when you’re not using your smartphone! All of this only to play to Basketball 3D! But then again you can find the incriminated games by typing “Alphonso Automated” in the Google Play search bar. And then erase (but too late) the indiscreet applications from your smartphone.

That can be a little bit tiresome, but it becomes necessary to read the list of demanded access rights when you download an app. At least if you want to be able to chat quiet of intimate subjects. During your read, wonder about the real requirements for the smooth running of the application, and if strange authorizations are demanded, do not download the app! Once again, contrary to Google, Basketball 3D shouldn’t need to access your microphone to work on your smartphone! That makes this request suspicious …



Now is the time for a small word about Facebook, which is also interested in your smartphone. In an article from January 21st, 2018, Marianne3 gives you an example of what is nowadays possible to do and what explains sometimes certain coincidences of friends’ suggestions with people that you just met at the bar: ” it is possible to them to detect that two smartphones are in the same place at the same time. But it’s even more intrusive: by comparing the data of every available smartphones, the social network is able to determine if the people face each other or walk together. “.



And we shall end our article with the worst. There are things against whom the common sense and stubbornness (for the reading of the terms of sale) are insufficient. The Russian company of IT security Kaspersky Lab published on January 16th, 2018 a report on the malware named Skygofree4 which spread on smartphones using Android since 2014. This malware infects smartphones through Web sites imitating those of the main telephony operators. Very difficult to detect, it is able to locate the mobile and to cross its position with the listening. Skygofree could also pirate WhatsApp, listen to phone conversations, read SMS and more widely see contents of the memory of the smartphone. And also take photos and videos without acknowledging the user of that. It’s clearly spy material!

Thus in conclusion: as your smartphone watches you, watch your smartphone!

Cool stuff : Webcam Cover 0.7mm THIN – Magnet Slider – Protects your privacy


1■ Article from Sputniknews website Google vous écoute en permanence

2■ Article from New-York Times website That game on your phone may be tracking what you’re watching on TV

3■ Article from Marianne magazine : Facebook redouble de nouveautés pour vous espionner

4■ AFP article Un logiciel de surveillance cible les mobiles Android


Back on top of the article