Technology News

That's enough. Who laughs last, laughs best. Should we consider that the melting of glaciers on the Antarctic continent is accelerating? Is there a risk of catastrophic sea level rise?

In its latest edition, Courrier International compiles a list of the craziest science projects… to reverse the trend. Although they are not realistic, they do give an idea of ​​the task to be accomplished…

In an article in the journal Science Advances, German scientists propose installing hundreds of snow cannons on glaciers. A massive undertaking that would require pumping billions of tons of seawater, not to mention building 12,000 wind turbines.

In total, an area the size of … Costa Rica would have to be covered with snow. Or 265 times that of Île d'Orléans if you prefer.

Another crazy project published in the journal Nature proposes building a 100-meter-high wall. The latter would prevent the ocean from eroding the glaciers' underwater base. Especially those in the west of the continent, which seem to be more at risk.

Without action, a catastrophe scenario could occur around the year 2150. The Pine Island and Thwaites glaciers would melt ten times faster than they do today. Enough to raise sea levels by four centimeters per year. One meter every 25 years.

At this point, many beach houses would prove as useful as a solar-powered flashlight. Or installing an ashtray in a motorcycle. Well, you know what I mean…

Source: Nautilus, quoted by Courrier International

(Taken from Calendario Romano website)

Quick question

For 20 years, the Catholic Church in Italy has welcomed the sale of a calendar featuring photos of priests considered “beautiful” and “attractive.” What revelation does the 2024 edition of Calendario Romano, also known as the “Pin-up Priests” calendar, bring?

has) Two of the “priests” photographed are actually women;

B) All photos were created by artificial intelligence. There are no priests;

vs) At least three priests pose bare-chested, which is considered inappropriate;

D) Several men photographed are not priests;

e) Ten of the twelve priests also appear in a calendar of handsome fascist activists.

Answer: d) Sources: The Telegraph and Calendarioromano.org

Rising

30 centimeters

Average increase in width of vehicles sold in the United States since 2013. Do the math. That means on average about three centimeters per year!

Source: thezebra.com

Quick question

In Britain, public opinion is abuzz over a dog abandoned in a supermarket car park. What detail of the case surprises many people?

has) The animal had six legs;

B) The animal was attached to a large bag containing 28 gold bars;

vs) The animal appeared to utter the phrase “Help me” as it barked. [Aidez-moi].

D) The animal wore a collar with a small handwritten note attached: “I am vegan”;

e) The animal was wearing an $18,000 leopard fur coat.

Answer: a) Since then, the dog has undergone surgery to “remove” two legs. Source: bbc.com

The duration

72 years old

Time has passed since the sale of alcohol was officially banned in Saudi Arabia. The country is preparing to approve the opening of a first outlet of the devilish liquid in Riyadh, but only for “non-Muslim” diplomats. There are whispers behind the scenes that this is a prerequisite for the country to host the 2034 World Cup.

Source: The Guardian

(AFP archives, Ludovic Marin)

Quick question

In France, President Emmanuel Macron is full of ideas for improving public schools. Can you identify TWO interventions he would like to test in a hundred schools?

has) The introduction of a uniform uniform for all students;

B) For fools, wear a fool's hat;

vs) Mandatory gymnastics sessions in the middle of the day;

D) Fines against parents of students who are absent without good reason;

e) Learning the national anthem “La Marseillaise” from primary school onwards.

Answers: a) and e). If the experience is conclusive, it will be generalized from 2026. Source: Le Monde

The forecast

2034

If this trend continues, this will be the year that a person's wealth will reach $1 trillion for the first time. More than Switzerland's current gross national product. Of course, it would be unfair to say that the fight against poverty is useless. Except that at the current rate, it will take another 229 years to eradicate poverty around the world. So calmed down?

Source: Inequality Inc., Oxfam International

The tendency

Good manners lessons

Are your office colleagues acting like a bunch of Ostrogoths? Do you leave the shared kitchen in an indescribable condition? Do you walk through the halls in your underwear? Don't know how to say hello anymore? Thanks again? In the USA, several companies believe they have found the solution: the etiquette course! The sector has actually exploded since the pandemic. Some training courses even last six three-hour sessions! Enough to go through the rules of etiquette during video conferences down to the smallest detail. Or the art of making telephone calls. Without forgetting the wise advice not to make too much noise while eating. A real eye-catcher. And we might as well warn those who don't say “thank you”! did not receive a passing grade when leaving…

Source: Los Angeles Times

(Wikimedia Commons, Xben911)

Old

192 years old

Approximate age of a Seychelles tortoise named “Jonathan” that lives on the island of St. Helena between Brazil and the African continent. It is now considered the oldest animal in the world. Jonathan was born in 1832 and brought to the island in 1882. At that time he had reached the respectable age of 50. At the time of Jonathan's birth, France was still ruled by a king, the United States had only 24 states, and Louis-Joseph Papineau was the most prominent politician in Quebec, which was still called Lower Canada.

Source: The Telegraph

The length

1500 kilometers

Prolonged traffic jams paralyzed Belgium on January 17 after 25 centimeters of snow fell. If we line them up, these traffic jams would stretch almost the entire route between Quebec and Chicago…

Source: La Libre Belgique

The punishment

12 years of forced labor

Sentence passed on two 16-year-old teenagers from North Korea who had the misfortune of watching videos of music or films from South Korea.

Source: The Guardian

To respond to this column, write to us [email protected]. Some answers may be published in our Opinions section.

The Windows 11 preview was released as part of the Windows Insider program

Windows 11 Build 23620 is rolling out through the Windows Insider Program Dev Channel. This is a troubleshooting access release. There is no novelty and change.

With this preview, Microsoft is simply addressing issues. It affects the start menu, the file explorer or even the widgets and the task manager.

In detail, Redmond fixes a display issue in the Start menu in the list of all applications when the text scaling is greater than 100%. Usually, fixing an undocumented problem makes File Explorer more reliable. On the Widgets page, hiding the Microsoft Start flow should now work. Likewise, the spacing and fonts used on certain settings pages should no longer be incorrect.

Windows 11 Build 23620, known issues present

Like File Explorer, Task Manager becomes more reliable and language access blocking when configuring new languages ​​is resolved. Microsoft says it has fixed crashes affecting the Narrator. One of these occurred when trying to add new natural voices.

Since this is the DEV channel, this preview is not recommended on a production PC. Two known issues without resolution are listed. The first concerns copilot. It is explained

The first time you start or after updating Copilot when using voice access, you must use the Show Grid commands for the first time to click in the Ask Me Anything box.

The second bug concerns the keyboard navigation between the settings subpages and the parent settings page, which no longer works.

The discovery of new materials is a lengthy and expensive process that often relies on complex empirical experiments and simulations. To accelerate the discovery of solid electrolytes for batteries, a very active area of ​​research due to the environmental and economic issues surrounding energy storage, researchers from Microsoft and Pacific Northwest National Laboratory have developed an innovative approach that combines AI and high technology. Performance computing (HPC) in the cloud. Their method allowed them to explore a vast chemical space and predict new stable and functional materials, which they then synthesized and characterized experimentally.

Electrolytes are materials that allow ions to be transported between the electrodes of a battery. In lithium-ion batteries, used in a variety of applications from smartphones to electric vehicles, the electrolytes that transport ions between the battery's two electrodes, the anode and cathode, are liquid, potentially flammable or toxic.

Developing solid-state batteries that offer safety, performance, and durability benefits is a key goal of materials scientists. This is also the aim of the ELIAS project, led by Saft and implemented by a consortium of academic and industrial stakeholders, launched in May 2023 and supported by France 2030.

However, it is difficult to find solid electrolytes that meet all required properties such as thermal stability, electrochemical stability, ionic conductivity and compatibility with other battery components.

To address this challenge, researchers from Microsoft and Pacific Northwest National Laboratory used AI models to filter more than 32 million candidates based on stability, band gap, electrochemical stability window and ion diffusivity criteria. Li or Na. These AI models are based on graphical neural networks capable of representing and learning the properties of crystal structures, as Google DeepMind researchers recently demonstrated with GNoME.

They trained them using data from quantum computing and public databases and deployed them on cloud computing resources, reducing the time and cost of materials discovery.

Among the candidates filtered through the AI ​​models, the authors selected the most promising ones to subject them to more precise calculations based on density functional theory (DFT) and ab initio molecular dynamics (AIMD). These calculations, performed on Microsoft's Azure Quantum Elements, which provides access to a cloud-based supercomputer suitable for research in chemistry and materials science, helped confirm the materials' stability and conductivity, as well as other properties such as hardness and Evaluate density and cost. The researchers also eliminated rare or toxic materials, identifying 23 final candidates, including five already known ones.

They then synthesized and characterized the structures and conductivities of their best candidates. Instead of using either lithium (Li) or sodium (Na) ions as conductors, they took an unconventional approach: combining these ions. The introduction of Li instead of Na significantly improved the ionic conductivity in a promising new solid electrolyte material, Na2LiYCl6, compared to the starting material Na3YCl6. Furthermore, the introduction of lithium resulted in a significant reduction in activation energy compared to the starting material, suggesting a more efficient ion diffusion process. This improvement in conductivity and reduction in activation energy is likely due to the presence of Li+ ions in ion transport as well as possible changes in the crystal structure.

This new material opens up perspectives for the design of versatile solid-state batteries. The authors emphasize that integrating AI and HPC in the cloud not only accelerates materials discovery, but also democratizes the discovery process by making computing resources easily accessible and reproducible.

Their work illustrates the potential of AI and HPC in the cloud to transform materials discovery and advance scientific and technological innovation. The authors plan to continue their research by exploring other chemical areas, refining their AI models, and validating their candidates in solid-state battery systems. They hope their approach will inspire other researchers to discover new materials with targeted properties.

Article references:

“Accelerating computational materials discovery.” with artificial intelligence and cloud high-performance computing: on a large scale from screening to experimental validation” arXiv:2401.04070v1 08/01/2024

Authors :
Chi Chen1, Dan Thien Nguyen2, Shannon J. Lee2, Nathan A. Baker1, Ajay S. Karakoti2, Linda Lauw1, Craig Owen3, Karl T. Mueller2, Brian A. Bilodeau1, Vijayakumar Murugesan2, Matthias Troyer1

1 Azure Quantum, Microsoft,
2 Directorate of Physical and Computer Sciences, Pacific Northwest National Laboratory,
3 Microsoft Surface, Microsoft.

ChatGPT was launched in late November 2022 and quickly surprised the world with its amazing performance. The text generation application was able to deceive many readers, even the most attentive ones, because they were unable to distinguish texts created by artificial intelligence (AI) from those written by a human. But how could what many thought was impossible yesterday become reality so quickly?

“The explanation for this rapid rise of artificial intelligence and ChatGPT can be viewed as a triangle, the three vertices of which are equally important. First, the computing power of computers has increased dramatically. Second, the amount of quality data for training neural networks has exploded. Thirdly, there have been several innovations in the architecture of neural networks,” explains Professor Nicolas Doyon.

Nicolas Doyon, professor in the Department of Mathematics and Statistics and researcher at the CERVO Research Center, explained the rise of AI and Chat GPT in his talk “The Mathematical Secrets of ChatGPT”.

At the invitation of the Continuing Education Department of the Faculty of Science and Technology to hold a general public conference on this topic, this professor from the Department of Mathematics and Statistics and researcher at the CERVO Research Center spoke about some milestones in the history of AI and certain scientific and mathematical principles popularized, on which the success of the famous computer application rests.

A champion chess machine

One of the greatest achievements in artificial intelligence dates back to 1996, when the computer Deep Blue managed to beat world chess champion Garry Kasparov. Deep Blue was programmed to create a tree of possibilities, assign a value to the final positions of the tree's various branches, and then determine the best possible move.

However, this approach, which worked well in chess, was less suitable for the game of Go, whose board forms a 19 x 19 grid – providing many more possibilities for moves than chess' 8 x 8 format. The tree of possibilities became too large even for a computer. “That’s why,” says Nicolas Doyon, “the researchers then said to themselves: ‘This doesn’t correspond to our way of thinking at all.’ How could we take inspiration from the way the human brain and neurons work to improve artificial intelligence?”

Mimic neurons

By studying how human neurons work, we found that they do not respond to all messages they receive. A message must reach a minimum threshold for the neuron to emit a so-called action potential, which always has the same strength and shape regardless of the intensity of the original message. This action potential is passed on to the next neuron via a synapse. It's an all-or-nothing law.

However, synapses are not just used to transfer information from one neuron to another; Their plasticity would play a central role in learning. In fact, researchers have found that the connection strength of synapses changes over time. “Simply put: the more frequently a synapse is used, i.e. the more it transmits an action potential to the next neuron, the stronger it becomes.” Under the microscope we can clearly see that the dendritic spine, an area of ​​the neuron, becomes larger, when a person learns. In short, by getting bigger and stronger, the synapse gradually changes the way we think,” explains the professor.

How can these biological facts be represented mathematically? “One way to transfer the all-or-nothing law into mathematics,” answers Nicolas Doyon, “is to use the Heaviside function.” In mathematics, functions often go continuously from 0 to 1. “The Heaviside function, on the other hand,” he explains, “is a function that has the value 0 until the input to the function reaches a certain threshold. Then suddenly it goes to 1.”

All or nothing can be represented mathematically by the Heaviside function.

“To illustrate the role of the synapses,” he adds, “we assign weights to the different inputs of the neuron.” From the graph, we can see that after determining the numerical values ​​of the inputs, we assign these values ​​to the weight of the Synapse, add the results of these multiplications to obtain a weighted sum, and finally we check whether this output value meets the required threshold, resulting in 0 or 1.

“To illustrate the role of synapses, we assign weights to the different inputs of the neuron,” explains Professor Nicolas Doyon.

Train the network

In recent years, artificial intelligence has made major breakthroughs thanks to the development of deep learning. “We now work with neural networks with several layers: an input layer, intermediate layers and an output layer. Between a neuron in one layer and a neuron in another layer there is a connection strength, also called synaptic weight, and as the network learns, each of these weights is adjusted,” notes Nicolas Doyon.

And how does the network learn? Through training, the researcher states. Consider the case of a neural network tasked with confirming whether the photo is that of a cat or a dog. We will assign a value of 0 to the cat and a value of 1 to the dog. To train the network, we will use thousands or even millions of images of these little creatures and examine the percentage of well-classified images. If the network does not give the correct answer, it did not get the correct output value because the synaptic weights were not well matched. We will therefore continue to adjust these weights until we achieve a very high success rate.

But how do I adjust the weights? “One of the things we use is the gradient descent method. To illustrate this, we can imagine a person trying to descend to the base of a mountain as quickly as possible. This is easy to imagine if there are only two inputs. On the x-axis we represent the success rate associated with different weights by which we multiplied the first entry, and on the y-axis we represent the success rate associated with different weights by which we multiplied the second entry. The error is displayed on the Z axis. Then it is possible to visualize the point where the error is smallest and try to adjust the weights so that they move in that direction,” explains Professor Doyon, who adds in the same breath that the principle, Although always the same, it is more difficult to visualize in reality when the number of parameters to be adjusted is in the millions or even billions.

We adjust the synaptic weights using the gradient descent method.

Math and reading at the heart of ChatGPT

The exact numbers are of course not disclosed publicly, but we can estimate that ChatGPT needs to adapt a network of 60 to 80 billion neurons, 96 layers and 175 billion weights. For comparison: there are around 85 billion neurons in the human brain. “The comparison remains a bit lame,” agrees Nicolas Doyon, “because our neurons are not quite similar to artificial neurons, but we are roughly in the same order of magnitude.”

When the computer application is asked to define itself, it responds: “ChatGPT uses a deep neural network structure. It is important to note that ChatGPT does not possess deep understanding or self-awareness. The answers are based solely on the statistical probabilities of the words or phrases.” To generate a text, ChatGPT calculates the probabilities that another word sequence will follow from a word sequence and then suggests the most likely sequence.

To achieve this, ChatGPT had to train billions of data points. The content of this reading is of course subject to confidentiality. However, it can be assumed that the network was trained on over 300 billion words. “If you read 300 words per page and one page per minute 24 hours a day, you would have to read for 1900 years to absorb that much information,” explains the mathematician, to get an idea of ​​​​the scale of the problem, using the library as a basis for learning ChatGPT.

“If you read 300 words per page and one page per minute 24 hours a day, you would have to read for 1,900 years to absorb that much information.”

– Nicolas Doyon on the supposed 300 billion words that make up the ChatGPT training database

Between amazement and fear

ChatGPT's sometimes breathtaking performance captures the imagination of some who see the future as a science fiction movie where artificial intelligences dominate the world. However, it is not this scenario that worries those among scientists who would like to see greater regulation of AI development. Rather, their intention is to prevent certain slips associated with human use. They also want us to take the time to better understand and analyze the negative impacts of this technology.

“What could possibly go wrong? Apparently students can use ChatGPT to cheat. Plus, people can lose their jobs. Recently, striking writers in Hollywood called for limiting the use of AI in screenwriting,” recalls Nicolas Doyon.

In addition, the professor reveals, other problems are less obvious and more insidious. “For example,” he says, “AI in the area of ​​facial recognition would more easily recognize white men than women or people who are visible minorities.” This fact is a little surprising since we imagine a neutral artificial intelligence. It can't be sexist or racist. But because the AI ​​was likely trained on a database that contained more male and white faces, it inherited our mistakes.”

Another example the professor gives comes from DeepL, a translation application that uses the same principles as ChatGPT. “If we ask DeepL to translate “she reads” into Hungarian, he says, we get “ὄ olvassa.” If we ask him to translate the same Hungarian words into French, he will say “il lit”. For what? Since the database has a statistical bias, the male subject is more often found before the verb “read”.

The often hidden environmental problem should not be taken lightly. “People think that AI is virtual and has no impact on the environment. However, according to an article, ChatGPT drinks 500 ml of water every time you talk to him. This image was used to remind us that massive amounts of water are required to cool supercomputers. In addition to this resource, ChatGPT also requires a lot of energy. Some say that AI will soon use as much electricity as an entire country,” says Professor Doyon.

So what does the future of AI and ChatGPT look like? “I don’t know,” Professor Doyon answers humbly. “Are there things ChatGPT can never do? I have no answer. Every month we hear that Chat GPT has done something new. It is impossible to know where this will all end,” concludes the mathematician.

• For an overview of Nicolas Doyon's work

• Learn more about past and upcoming general public conferences organized by Continuing Education at the Faculty of Science and Engineering

• Watch the conference “The Mathematical Secrets of ChatGPT”:

At what age did you start writing independently? Today we often start earlier and more frequently, for more reasons and in more forms. Young people quickly understand the power and necessity that writing can have.

New forms of expression through writing are emerging: emails, tweets, emojis, mutations of rap, slam, graffiti, playing on both calligraphy and intelligibility, emotion, rhythm or polysemy. Usage and technical considerations also influence semantics, syntax, and grammar, which are expected to evolve over time. Even writing machines; If they can already adopt the style we want, they will soon be able to recognize ours among thousands.

Who says writing also means reading? Some bloggers have a significant number of loyal readers, others have virtually none and are relegated to the dungeons of the rankings. A writer's challenge has always been to be read, but today it's about competing with hundreds of millions of accessible short stories and works that have been effectively indexed, annotated, quoted and translated, which include “machine messages and rich AI-generated prose “to be added

So the idea of ​​cyborgizing our writing with that of AI is gaining ground… to be better disseminated using strategic words and key phrases compiled by algorithms that amplify the reach of our message, which is transformed by these considerations became… . The balance between popularity, originality and ethics becomes a real problem. For an AI, originality and consistency in both content and style are the highest values ​​of the texts that serve as its inspiration, they are its food that makes it appear “intelligent”. Future authors will inevitably have to further develop their relevance and the quality of their texts, while “communication agents” must, above all, increase their productivity.

The Scripture is not yet finished distributing its parts.

Denys Lamontagne – [email protected]

Illustration: Peshkova – DepositPhotos

Cygames, Inc.a Tokyo-based video game publisher and developer, announced that the demo of Granblue Fantasy: Relink has been available to download on the PlayStation 5 and PlayStation 4 platforms since January 12, 2024 (as mentioned in our article here). Remember that the full version will be available for purchase from February 1, 2024.

Granblue Fantasy Demo: Relink

The demo offers three modes:

1. Story mode Dive into a small part of the main plot! At the end of this mode, players will receive a reward that can be used in the full version of Granblue Fantasy: Relink.

2. Quest mode Take part in a selection of three quests that you can play alone or with other players online! At the end of this mode, rewards can be claimed in the full version.

Eleven characters are available in quest mode, each with their own unique fighting style. Favorites like Gran, Katalina, Rackam and more await you.

3. Tutorial mode Go through this mode to learn basic commands.

*The content of the demo may differ from the full version. *Game data cannot be saved or loaded in the demo. *Progress made in the demo will not carry over to the full version. *Demo rewards can only be claimed in the full version if the demo data is retained.

*The game is available on AMAZON.FR, see this link

Google Art Selfie 2 is here to revolutionize your selfie gallery! This innovation, available in the Google Arts & Culture application, turns your selfies into artistic portraits thanks to the magic of artificial intelligence.

Imagine turning your face into a museum piece of art! Using advanced algorithms, this feature allows immersion in different artistic styles and historical periods, making each selfie unique and educational.

But it's not just a fun gadget. Google Art Selfie 2 goes beyond entertainment by incorporating a fascinating educational aspect. Each generated portrait is with information about the style and artistic eraand thus enrich your knowledge of art, history and culture. This feature transforms the user experience into a fun and educational adventure.

Google Art Selfie 2, more than an application: a journey into art and culture

The Google Arts & Culture application has been continuously enriched since its launch in 2011. It offers various interactive tools such as: Create poems, postcards, or even search for artwork from images. These features make discovering art and culture an exciting game.

In addition to Google Art Selfie 2, the application also offers “Art Remix”, an extension that allows you to do this Customize themes and art styles. This option invites users to express their creativity and share their unique creations with their friends, making artistic exploration even more interactive and personal.

Since 2018, Google Arts & Culture has been celebrating overwhelming success. With each update, the application becomes more complete and enriches the cultural and artistic experience of its users. The application's popularity continues to grow, attracting an ever larger and more diverse audience.

Are you ready to turn your selfies into works of art? Download the app Google arts and culture on your smartphone or tablet. Head to the Art Selfie section to explore the available art styles, take your selfie, and watch the AI ​​work its magic. It's a fun and informative way to learn more about art and culture while playing with your own image.

In early December 2023, we learned that unidentified governments were forcing Apple, under the guise of American law, to tell them the contents of push notifications from iPhone users.

Senator Ron Wyden also sent a letter to the US Department of Justice on this issue, expressing the urgent need to repeal the laws that prevent transparency of these types of requests. The topic of spying on push notifications from Apple smartphones is back on the agenda today, but this time with different protagonists.

© iPhon.fr

Meta, Microsoft and LinkedIn in the searcher

In a video posted on his YouTube channel, cybersecurity researcher Tommy Mysk shows how certain web giants deliberately violate rules imposed by Apple to refine their advertising targeting. This time it's not about democratic governments, but about companies like TikTok, Facebook, Twitter, Bing and Linkedin. Tommy Mysk explains in his video how they hijack push notifications to collect user data.

The hypocrisy of confidentiality

Recently, many disturbing elements have increased the general distrust of large platforms. In particular, the investigation by Joseph Cox of 404 Media, which highlighted the offers of certain advertising agencies. The latter boasted of being able to spy on people using the microphone on their smartphones so that they could listen to their conversations. This with the aim of their Offer verbal promotion of the products or services in your daily conversations.

Despite constant promises from web giants regarding data protection, the temptation to profit seems to be far too great so that they decide to finally stop secretly collecting as much data as possible. As a reminder, Apple remains the only company that has not fallen into this trap. This is easily explained by the fact that the revenue does not come from advertising, and therefore our data is not as interesting for the company as for other GAFAM. However, it is good to note that Google is making the most effort in this area, although there are still many shortcomings.

See also:

i-nfo.fr – Official iPhon.fr app

From: Keleops AG