Categories
Ξ TREND

Meta presents an AI tool capable of generating images or texts imitating human reasoning


Meta has presented a new tool that allows the generation of images and texts with artificial intelligence (AI) using the prediction of certain parts of the content and imitating human reasoning.

The company has explained that this solution is the result of an idea by Meta’s chief AI scientist, Yann LeCun, who proposed “a new architecture aimed at overcoming the main limitations of the most advanced AI systems,” as stated in a statement. .

The result of their work is the Image Joint Embedding Predictive Architecture (I-JEPA), a tool that collects data from the outside world, creates an internal model of it, and compares abstract representations of images, instead of comparing the pixels themselves.

The company has recalled that humans “learn an enormous amount of prior knowledge about the world by passively observing it”, an aspect that it considers “key to enabling intelligent behavior”.

For this reason, the objective of this model is to predict the representation of a part of a content, such as an image or a text, based on the context offered by other parts of the composition.

Once I-JEPA collects all this information, it is in charge of predicting the missing pixels of an image or the words that do not appear in a certain text, to give it a natural and realistic meaning.

Meta has also commented that, unlike other generative AIs, theirs uses “abstract prediction goals” for which unnecessary detail is removed at the pixel level, allowing the model to learn additional semantic features.

The company has finally indicated that it continues to work on expanding the focus of this tool so that it learns “more general” models based on more specific modalities. For example, allowing spatial and temporal predictions about future events with a video from a simple context.

Categories
Ξ TREND

Fae Farm Magic Pickaxe Use: Magic Tool Hacks


How to use the iron pickaxe magical ability in the new Fae Farm? What exactly do you have to do in the “Magic Tool Hacks” quest and how do you even get the magic pickaxe? In the new Fae Farm from Phoenix Labs, all important tools are unlocked in the first minutes of the game. The tools include an axe, a shovel, a watering can and also a pickaxe. With the pickaxe you can mine the stones and ores such as copper and iron in the game. The tools can then be improved and upgraded as the game progresses. As part of the quest “Magic Tool Hacks” you are supposed to improve the pickaxe and use the magical ability of the tool. In this guide we show how this works.

Unlock Iron Pickaxe

The first step is to upgrade the normal pickaxe to the copper pickaxe. To do this you have to visit the blacksmith Cinder in the docks. To upgrade the iron pickaxe you first need the copper pickaxe. The copper pickaxe can be made with a copper ingot, which can be made from copper ore at the stone forge. You can get the copper in the saltwater mine from level 3. You can then go to Cinder and make the copper pickaxe.

The same thing happens with the iron pickaxe, for which you need an iron ingot that you can make from the iron ore. The iron ore can be found in the saltwater mine from level 12. The iron ore is then taken back to the stone smithy to produce the iron ingot. You can then use the iron ingot to craft the iron pickaxe from Cinder.

Use iron pickaxe’s magical ability

The iron pickaxe has a magical ability. To use the special ability, you must hold down the F button (PC) or the X button (Nintendo Switch) in front of a stone or other boulder. The character then strikes and destroys the stones in a small area.

Categories
Ξ TREND

NordVPN launches a free tool to check the security of any website: this is how its link checker works

Nowadays we have numerous ways to browse the Internet safely and privately. Whether through VPNs, ad blockers and other ways to avoid suspicious links and malicious files, the truth is that we have multiple tools to maintain our anonymity and guarantee our security on the network.

Among the companies that offer multiple products of this nature, NordVPN has just launched another very useful service for check the security of a website just by typing the URL manually in this service. Below these lines we tell you how it works.

A free link checker for any device

NordLabs has just launched a link checker that you can access completely free and from any device. Basically, we can add any URL to the box and the tool will tell us if the website is trustworthy.

Lately, various typographical tricks have been used in the URL to try to deceive users. Some tricks are more sophisticated than others, such as replacing the letter ‘a’ in ‘Amazon’ with its Cyrillic equivalent, something almost undetectable to the naked eye since the letters are practically identical.

Luckily, there are ways to avoid these types of scams and an alternative is the one recently launched by NordLabs. When we put the URL in the checker, the tool will tell us if it is a trustworthy website or not. To do this, this tester Compare the website to a list of sites known to include scams or malicious elements of some kind.

To collect information about malicious websites, NordVPN uses machine learning, specifically trained to detect zero-day phishing patterns implemented on websites. Along with this, the model makes use of a Nord database to identify fake websites that intend to scam users.

“One of the biggest advantages of the link checker is its dual nature. Combining proprietary machine learning techniques with the Nord Security Intelligence database, the link checker offers one of the most complete data sources to detect harmful URLs. In fact , this link checker incorporates information on the 95% of the most popular domains imitated by cybercriminals, who often use fake variations to disguise online scams or malware,” said Vykintas Maknickas, head of product strategy at Nord Security.

The tool It can be used freely and without the need to register on NordVPN, so it is a very easy way to check if the website we are trying to access is safe or not.

This product is the second experimental development created by NordLabs. In September last year, NordLabs introduced Sonar, an AI-powered browser extension to help users detect phishing emails and protect themselves from cybercrime.

Categories
Ξ TREND

Google launches virtual try-on tool powered by generative AI


American Internet users who are looking for a female top on Google Shopping can now view it on a model with their morphology and their hair and skin color. This tool based on generative image AI will be extended to other wardrobe pieces in the near future.

E-commerce players are beginning to take advantage of the tools offered by generative artificial intelligence, capable of producing text, images or sounds on demand. After Amazon, which uses AI to summarize the comments of its customers on products, Google in turn presented on June 14, 2023 a feature called Try on: the Internet giant wants to allow Internet users to view clothes on hundreds models with different sizes, weights (from XXS to 4XL), skin and hair colors. The objective: to find a model that looks like them to find out if these burgundy velvet pants would fall well on their hips or if the color of this fuchsia jacket would suit their complexion. Several images are available for each garment. The models take different poses, which allows the user to imagine how his crop top would behave if he raised his arms in the air.

Solve the issue of product returns

This feature is already active in the USA for women’s tops from a few brands, such as H&M, Anthropologie, Everlane, and even LOFT. It will gradually be offered to other companies and extended to other pieces of the wardrobe, starting with shirts and other men’s T-shirts, which can be scrutinized by e-consumers from every angle thanks to Try on by the end of 2023. This tool will not fail to interest e-merchants who promote their clothes on Google Shopping, because it could allow them to limit the costly returns of their products (in France, 30% of clothes purchased online are on average returned). Google does not give details on its cost for the moment.

This issue had been identified by start-ups such as the Israeli Zeekit, acquired by Walmart, or the French Veesual AI. Over the past few years, they have developed virtual fitting tools, based on AI or augmented reality, but which sometimes lacked realism. Generative artificial intelligence comes to solve this problem: the Try on models are very real, but the images of them wearing the clothes are generated by algorithms and seem more real than life, as shown in the screenshot below . They were trained internally by Google from two tools for automatically generating images from text, DALL-E 2 (created by the Open AI company, which is also behind the text generative AI ChatGPT) and Stable Diffusion (developed by the company Stability AI). Next step ? Allow the customer to post a photo of themselves to Google so that the product is projected directly onto their body.