Home Blog Page 222

Toyota to install Invisible AI’s tech in Factories

Toyota Invisible AI tech

Automobile manufacturing giant Toyota announces that it will install next-gen computer vision company Invisible AI’s tech in its factories across North America. 

The artificial intelligence-powered technology will enable the company to make more informed decisions about quality, safety, and productivity in its factories. 

According to the company, it has cracked a deal with Invisible AI to upgrade 14 of its factories with Invisible AI’s cutting-edge computer vision technology. 

Read More: EU accuses Apple of restricting rivals’ access to Payment tech

As per the terms of the agreement, Invisible will supply Toyota with the systems, beginning with a deployment of 500 AI edge devices at its Indiana factory. 

Group Vice President, Vehicle Production Engineering & Manufacturing Production Innovation Center, Stephen Brennan, told Forbes, “We observe our employees assembling vehicles to identify inefficiencies and bottlenecks in their standardized work.” 

He further added that Invisible AI systems would assist them in increasing the frequency and accuracy of process reviews, along with reducing the time required to identify inefficiencies across processes, allowing them to devote more time to improve. 

Invisible AI’s technology employs 500 edge AI devices equipped with an NVIDIA Jetson chipset, 1TB of storage, and a 3D camera to track all floor activity. The technology can effectively provide real-time insights and information that could considerably help Toyota increase its teams’ productivity and accuracy. 

“What we’re giving them is tools to be able to see more of what’s going on inside their facility, and they can then go in and correct any problems and any bottlenecks, making sure people are being safe,” said Co-founder and CEO of Invisible AI, Eric Danziger. 

He also mentioned that what they have is a running AI computer vision model that is constantly processing all of the incoming videos to provide real-time information and insights while processing that data efficiently.

Advertisement

NASA leverages AI-based CFD to Develop Hypersonic Missiles

NASA aRGONNE Vulcan cfd ai HYPERSONIC MISSILE
Image Source: BBC

According to sources, NASA is developing a new hypersonic missile that will allow missiles to be optimized for maximum range and damage. It is leveraging an AI-based technology to work around a scramjet missile design by using the results of computational fluid dynamics (CFD). Compared to typical, bulky rockets, the Scramjet, or ‘supersonic combustion ramjet,’ makes optimal fuel use, reaches hypersonic speeds, and produces maximum damage on impact. The scramjet achieves far quicker speeds because of very effective propulsion mechanisms, and it can be built in a much smaller, lighter form, and has a greater range than traditional rockets. This is where Artificial Intelligence enters the hypersonic missile development process.

Any aircraft or drone that can achieve Mach 5 or more, or five times the speed of sound at sea level, is considered hypersonic. To give an insight, passenger planes fly at around 600 mph (966 kph), whereas hypersonic systems move at speeds of up to 3,500 mph (5,633 kph) – roughly 1 mile (1.6 kilometers) every second. 

Achieving such velocity allows hypersonic aircraft to bypass existing air defenses and anti-ballistic missile systems. It also improves their ability to penetrate strongly protected facilities and kill targets with only kinetic energy, without the need for a payload of high explosives. The ability to move and maneuver at hypersonic speeds, on the other hand, poses considerable and diverse engineering hurdles.

Because of their high-velocity mobility throughout their trajectory, these new missiles must be followed throughout their journey since their flight trajectories might alter as they travel. 

On the fluid forces factor, when a missile or aircraft breaks beyond the sound barrier, a shock wave is created that is hotter, denser, and has greater pressure than the surrounding air. Friction and air resistance produce a huge amount of heat in the hypersonic phase, which might cause structural damage to commercial aircraft.

By studying the behavior of fluids like air and water, researchers have long used CFD to anticipate, among other things, how an aircraft in flight would interact with the forces surrounding it. In CFD the behavior of fluids such as air and water are expressed numerically and visually to better understand the fluid dynamics. On top of that, aerospace engineers must consider how air flows around the plane or weaponry payload and how it behaves as it passes through the engines and reacts with fuel.

Typically, hypersonic missiles require enormous power and fuel to operate effectively at high speeds, which was previously only conceivable with rockets. This motivated NASA to turn to ‘scramjets’ as an alternative.

Traditional air-breathing jet engines, such as those seen in big passenger planes, actively take in and compress oxygen in order to burn fuel while flying especially using rotor blades. However, after the jet or weapon reaches Mach 3, this is no longer necessary because the passage of the jet or weapon through the air does this on its own. This is where scramjet outshines commercial aircraft. 

To help NASA in its mission USA’s Argonne National Laboratory will contribute its supercomputing capability to accelerate the development of hypersonic testing equipment, including experimental aircraft like the X-43A scramjet-powered plane created as part of its Hyper-X program. The company will enlist CFD to model and predict how an aircraft will react to the forces around it at hypersonic speeds. Earlier this year, the lab delivered a presentation at an American Institute of Aeronautics and Astronautics (AIAA) Forum, explaining how it is utilizing AI to better computer simulations and simplify the creation of cutting-edge aircraft.

With the help of Argonne, NASA has created a hypersonic computational fluid dynamics code called VULCAN-CFD, which models how combustion occurs in turbulent airflows of engines at sub-, super-, and hypersonic speeds with the aid of Argonne.

The software works by storing single, one-dimensional snapshots of flames called “flamelets” in enormous, multidimensional flamelet tables. Later, the flamelet table, which was created using Argonne software, was used to train an artificial neural network. This was critical because it minimizes the high memory needs and computing costs involved with simulating hypersonic fuel combustion.

Read More: NASA’s Perseverance Mars rover traverses Martian delta using AI

Thanks to the Argonne simulations, NASA aerodynamicists were able to better grasp how high temperatures, high pressures, and turbulent air flows influence aircraft at hypersonic speeds. That is, without putting experimental aircraft in dangerous real-world test scenarios at high altitudes, which can range from 100,000 to 300,000 feet at the very edge of outer space.

Meanwhile, Argonne believes that owing to the positive results of VULCAN-CDF, NASA now has a good handle on the effects of Mach 5 and higher on aircraft, allowing it to pursue the development of experimental and commercial hypersonic aircraft.

Advertisement

Last few days left to apply for Udacity and Cognizant Skills Accelerator

Udacity Cognizant Skills Accelerator

Udacity, in partnership with Cognizant, launched a new Skill Accelerator scholarship nanodegree program to help learners kick start their careers in Java programming. 

The Cognizant Skills Accelerator is designed to assist students in learning important foundational concepts in Java development. 

This unique scholarship program will allow selected learners to potentially start a career at one of the world’s leading tech companies as the top performers will get an opportunity for interviews or additional training with Cognizant. 

Read More: IIT-KGP and Virginia Tech launch Online Course in AI

Learners from multiple locations in the United States, such as Dallas, Tampa, Atlanta, Phoenix, Teaneck, and others, can apply for this skill accelerator program offered by Udacity and Cognizant. 

Students can take advantage of multiple hours of free training modules and tutorials on Java programming and its real-world applications, along with exciting opportunities to learn competencies in one of the fastest-growing fields in technology. 

Moreover, Learners get access to monthly ask me anything sessions with Java experts and curriculum developers (AMAs) along with one-to-one mentoring sessions from the advisors. 

It is a part-time online course that allows students to learn at their own pace and convenience. Learners will be eligible for a certificate if they submit and pass all of the given projects by the deadline. 

Udacity has already started accepting applications for this skill accelerator program, and learners can submit their applications by 11th May 2022. Interested candidates can visit the official website of Udacity to apply for this program jointly offered by Cognizant. 

Apart from this skill accelerator, Udacity had also partnered with Infosys to launch a similar scholarship program aimed at imparting valuable and relevant skills to learners to help them grow their careers in fields like Java programming and data operations analysis. 

Advertisement

Mastodon: What does Elon’s Twitter Acquisition mean for Fediverse and Decentralized Platforms?

Mastodon Twitter Elon Musk Fediverse
Image Credit: Analytics Drift Design Team

Last month on April 4, Musk revealed owing a 9.2% stake in the microblogging company’s free-floating stock, making him the company’s largest stakeholder. Though Twitter promised him a seat on its board of directors, Elon pulled a trump card by declining a seat on Twitter’s board of directors on April 10 in favor of attempting to purchase the entire company. 

What followed next was an interesting turn of events that started when Elon’s initial offer to buy Twitter at US$54.20 per share, citing that it was his “best and final,” in a Securities and Exchange Commission filing on April 13 That offer valued Twitter at about US$43 billion. The board initially rejected the proposal because, while the offer was significantly above Twitter’s existing share price, it was probably too low, considering that the stock had lately risen beyond it. The board also took steps to limit his power, including instituting a “poison pill” on April 15 that would diminish everyone’s shares and prevent a hostile takeover. This so-called poison pill, or shareholder rights agreement, allows other shareholders to buy the shares at a reduced price if Musk or another individual or group acquires more than 15% of the outstanding common stock. 

In addition, Musk’s funding for the acquisition was first unknown. However, on April 21 Musk explained in another Securities and Exchange Commission filing that he had obtained US$46.5 billion in commitments to help fund the deal. This includes around US$25.5 billion in loan funding from Morgan Stanley Senior Funding and other companies. Further, Elon stated that he had invested US$21 billion in equity.

After a roller coaster of rumors and speculations, on April 25, Elon Musk finally purchased Twitter. It is important to note that the buyout has not yet occurred, since the procedure will take some time to complete; nevertheless, the deal is expected to be completed by the end of the year.

Though Musk has frequently mentioned that his primary concern with Twitter is the company’s censorship of free speech, critics fear that the billionaire’s control of the platform will result in the silencing of their voices and those with whom he may disagree, given that he has frequently blocked critics from his account. Moreover, the ‘free speech’ bait also implies giving a platform to profiles that spew hate, spread misinformation, and fuel propaganda. Not only that, but since today’s social networking sites have the power to change the course of political opinion, trigger discourse on literally every trivial and important topic, and manipulate the crypto market with hearsay, the ‘free speech’ is a necessary gambit to bring chaos. This is mortifying especially when Twitter has been both fertile ground and echo chamber for anti-Islamic and antisemitic content.

To surmise, not everyone is happy with the takeover. Since the world was hit by the news of the then alleged buyout, many netizens have been looking for alternatives to Twitter. This development gave an influx of users to switch to Mastodon, a Germany-based open-source social network site. According to Mastodon’s CEO Eugen Rochko, the platform recently witnessed an influx of approx. 41,287 users with 30,000 new user signups in the wee hours post-acquisition. The unexpected surge of new users has also caused the servers of this modest social network to slow down. There was also discussion among the Twitterati about restarting their blogs and recommitting to RSS (Really Simple Syndication).

Rochko, who created Mastodon was inspired by a rumor about the potential selling of Twitter to a “controversial billionaire” back in 2016. In its essence, Mastodon is a decentralized social networking platform that allows users to build their own private networks. People can post 500-character messages known as “toots” (similar to Tweets on Twitter), repost or “boost” posts on their own timeline, and follow or privately chat with other users. It’s a software platform based on the open-source ActivityPub protocol, rather than a single company-run site. Anyone could establish their own Mastodon private Twitter-like networks, called instance thanks to the open-source functionality. While each of these instances can set up its own guidelines for users, users from different servers can also freely communicate with each other. When it comes to choosing your instance, users can join groups organized by area, LGBTQ+, and specialty hobbies such as art, music, activism, gaming, technology, humor, cuisine, and more. Also, because these networks are hosted on numerous servers, no one company can target them with adverts. Mastodon’s feed, which is chronological, and non-algorithmic, is another noteworthy contrast from Twitter.

Earlier Mastodon had only a web version, until last year when it launched an iOS app. To entice more users, it has recently launched an Android version. 

Mastodon accounts can communicate with ActivityPub-enabled accounts on other social networks. PixelFed, a decentralized and open-source Instagram equivalent, for example, leverages ActivityPub to allow Mastodon users to follow and engage with PixelFed profiles. This cross-network social following and engagement capabilities are contributing to the popularity of Fediverse. The phrase Fediverse is a portmanteau of the words “federation” and “universe.” It’s a cluster of federated (i.e. networked) servers used for online publishing (i.e. social networking, microblogging, blogging, or webpages) and file storage, and could communicate with one other while being hosted independently.

However, ActivityPub is not alone in the game. Bluesky, Twitter’s version of a decentralized protocol, has already been in development. The main difference between ActivityPub and BlueSky is that it is still a vaporware project that is aiming to produce a Web3-based version. Vaporware is software that has been publicly publicized and actively pushed by a company despite the fact that it does not yet exist.

Towards April end, Mastodon welcomed the EU to join its social network and put up its own server, called EU Voice, in the midst of the record inflow of users. The European Commission also opened an account for PeerTube, another decentralized video-sharing network, on the same day.

According to European Data Protection Supervisor Wojciech Wiewiórowski, the EU hopes to offer alternative social media platforms that emphasize individuals and their rights to privacy and data protection with the trial launch of EU Voice and EU Video.

While it is true that there is no guarantee that extremists and others won’t find new platforms to spread hate and misinformation, Mastodon has a number of policies in place regarding content that is not permitted on the network. This includes the requirement for sexually explicit or violent media to be labeled. Racism, sexism, homophobia, transphobia, xenophobia, and casteism are all prohibited. The propagation of violent beliefs or incitement to violence is likewise forbidden. Users are also forbidden from harassing, dogpiling, or doxing one another. There’s also a ban on any content that’s regarded as illegal in Germany, thus Mastodon users can’t deliberately publish incorrect or misleading information.

Read More: New AI tool can identify Depressed Twitter Users

Apart from Mastodon, former Twitter users flocked to newcomers CounterSocial and Truth Social, pioneering social networks like Tumblr and MySpace. There are alternatives that also reminisce you of older Internet days or are designed using former social networking sites. These include Neocities, Spacehey, and Pillowfort.

CounterSocial resembles Twitter in appearance and claims to be free of bots, trolls, deepfake, and fake news. It also promises not to mine or sell user data and is ad-free. But its beginnings are murky, as it was formed by The Jester, an anonymous hacktivist who is suspected of carrying out hacks against WikiLeaks, and several Islamist websites. 

Unfortunately, this site soon started receiving negative reviews. Users complained about the experience after joining this ad-free social network or tweeted that they had joined CounterSocial but didn’t know how to utilize the platform and that the site crashed after many new registrations.

Former US President Donald Trump’s social media network, Truth Social, has likewise gained a lot of interest since its debut in February. Though there was conjecture that Elon might reinstate Trump following the buyout, it did not materialize, despite news of Trump’s supposed departure from Truth Social. Trump reiterated that he has no plans to return to Twitter, but the prospect of his leaving Truth Social was enough to send shares of Digital World Acquisition Corp., a special purpose acquisition firm hoping to join with Trump Media & Technology Group, plunging.

Mastodon still has a long way to go before it can compete with Twitter, especially on the basis of size, backed by its virtues. The Mastodon.social server has just 662,000 users, however, the federated network as a whole has nearly 3.5 million. Twitter, on the other hand, has 217 million daily active users. There are many areas where Mastodon needs work to keep the new users want to continue using it. For instance, though it is simple to use and the UX appears to be dynamic, it has a drab aesthetic and misses some of the functionality of the bird app. 

While the open-source feature of Mastodon keeps attracting new audiences, it may not be for long as there Elon has been considering to open-source Twitter and has been an advocate of the same for a long time.

Rochko believes that the decentralized nature of Mastodon guarantees that it would not disappear like MySpace, Friendfeed, and SchülerVZ, a German version of Facebook. Those who remain long enough could observe some significant changes. According to Rochko, “really wonderful groups functionality” and end-to-end encrypted chats are in the works.

This is also not the first time the social networking platform had surged in popularity. Back in 2019, many users in India raced to join Mastodon as they were enraged by moderation policies. 

In retrospect, though Mastodon is a drop in the ocean compared the Twitter, years later if Fediverse goes mainstream, this recent mass migration of users and Elon’s Twitter acquisition will be regarded as catalysts. 

Advertisement

IIT-KGP and Virginia Tech launch Online Course in AI

IIT-KGP Virginia Tech Online Course AI

The Indian Institute of Technology Kharagpur (IIT-KGP) partners with Virginia Tech to launch a new online course in artificial intelligence (AI). 

The online certification program aims to address the challenges involved in the intelligent computation and interpretation of data domains. 

According to IIT-KGP, its new online course prepares analytics talent for today’s market, which requires an understanding of AI and its applications in data analytics. 

Read More: META announces new AI project to build AI-Language Models akin to Human Brain

The business analytics program takes an interdisciplinary approach, assisting students in identifying the utility of data so that they can improve the customer experience, diagnose problems, impact efficiency, satisfy customers, and improve quality of life. 

IIT-KGP and Virginia Tech’s new program, which will run for nine months from May 30, 2022, to January 7, 2023, will be open to students and industry executives worldwide. A total of 100 applicants would be chosen based on their academic credentials and work experience. 

IIT-KGP meticulously designed the course curriculum that includes modules like Big Data, structured & unstructured databases, statistics for business analytics, artificial intelligence, Classification and Clustering Methods, etc. 

“Our lead faculty members have Ph.Ds from highly reputed universities and have a vast amount of experience. Each one of them is considered an expert in their field and has achieved excellence in their academic pursuits,” mentioned IIT-KGP. 

Furthermore, graduate assistants from IIT Kharagpur will help learners of the program with assignments, projects, and tests. 

Individuals with an undergraduate degree in engineering or related subjects having an industry experience of two to five years can apply for this program. Interested candidates can apply for this IIT-KGP and Virginia Tech’s online certification program from the official website of IIT Kharagpur. Indian applicants need to pay an application fee of Rs 500, and foreign nationals need to pay $10. 

Advertisement

Last few days to apply for Udacity and Infosys Free Tech Scholars Program

Udacity Infosys Free Tech Scholars Program

Online learning platform Udacity has partnered with Infosys to launch a new free nanodegree Tech Scholars program. 

The scholarship program is aimed at imparting valuable and relevant skills to learners to help them grow their careers in fields like Java programming and data operations analysis. 

According to Udacity, the Infosys-Udacity Tech Scholars Program aims to assist students who do not yet hold a four-year college degree in learning valuable concepts. 

Read More: META announces new AI project to build AI-Language Models akin to Human Brain

Students from multiple localities in the United States, including Tempe, Richardson, Raleigh, Jersey City, Atlanta, Chicago, Seattle, and others, can apply for this scholarship program. 

This scholarship program will provide students with Hundreds of hours of free training modules and tutorials on Java programming and data analysis, along with their real-world applications. Moreover, learners can also access 1:1 mentoring from these advisors. 

Udacity will start this course on 25th May 2022 and run till 15th September 2022. However, as it is a self-paced course, learners can decide the timeframe within which they choose to finish the program. 

Candidates must be above 18 years old with a high school diploma or other equivalent qualification and have the right to work in the US without sponsorship are eligible to apply for this scholarship program offered by Infosys and Udacity. 

The application process has already begun, and applicants must submit their online application forms before 8th May 2022. Interested candidates can apply for this scholarship from the official website of Udacity. 

Advertisement

EU accuses Apple of restricting rivals’ access to Payment tech

EU accuses Apple Payment tech

The European Union (EU) accuses Apple of restricting its rivals’ access to iPhone’s payment technology. 

The allegation put on Apple is regarding the company’s act of blocking competitors from its popular “tap-as-you-go” iPhone payment system. 

According to early findings, the world’s most valuable company Apple may have infringed competition law by preventing competitors from utilizing its technology. If the charges are sustained, the company may face penalties of up to 10% of its global turnover, which was $36.6 billion last year. 

Read More: Amazon to Create 2500 Corporate and Tech Jobs in California

“Apple has unfairly shielded its Apple Pay wallets from the competition. If proven, this behavior would amount to an abuse of a dominant position, which is illegal under our rules,” said Margrethe Vestager, the EU’s antitrust chief. 

Vestager further added that the preliminary conclusion that they reached relates to mobile payments in shops by excluding others from the game. 

According to the Commission, this behavior by Apple has an “exclusionary effect” on competitors and “leads to less innovation and fewer choices for consumers for mobile wallets on iPhones.” 

In its defense, an Apple spokesperson said, “Apple Pay is only one of many options available to European consumers for making payments and has ensured equal access to NFC while setting industry-leading standards for privacy and security.” 

The spokesperson also mentioned that they would continue to work with the Commission to ensure that European customers have access to their preferred payment method in a safe and secure environment. Apple is also concerned that tweaking its payment system may make transactions less secure.

Advertisement

Guardforce AI announces 4800 Robot deployments Worldwide

Guardforce AI 4800 Robot deployments Worldwide

Integrated security solutions providing company Guardforce AI announces the successful deployment of its over 4800 robots at multiple locations across the globe. 

The company said that the locations include nine countries worldwide, such as Hong Kong, Macau, Thailand, Malaysia, Singapore, Dubai, the USA, Guangzhou, and Shenzhen. 

Government buildings, hospitals, hotels, tourist sites, restaurants, sports centers, supermarkets, and transportation facilities are among the sectors where Guardforce AI robots have been deployed. 

Read More: Traceable raises $60 million in Series B Funding Round

Moreover, Guardforce AI is expecting to pump the number of robots to 10000 by the end of this year. 

Chairman of Guardforce AI, Terence Yap, said, “In just February of this year, we announced we had surpassed 1,400 robot deployments. I am pleased to announce we have now surpassed 4,800 robot deployments worldwide, achieving an important milestone in our global expansion strategy.” 

He further added that their Robotics-as-a-Service (RaaS) business model constantly changes in response to customer input. Customers are offered early pilot programs as part of Guardforce AI’s rollout strategy, which allows the company to gather feedback and upsell new features and applications. 

According to the company, Guardforce AI will be able to quickly establish a substantial presence in the booming robotics sector. Terence Yap founded Guardforce AI, which is an integrated security solutions provider trusted to secure and move high-value items for both the public and commercial sectors. 

The solutions offered by the company’s robots include features such as disinfection, hotel check-in/check-out services (under testing), patrol, temperature screening, and advertising. Guardforce AI successfully tested ad placements on its robots in Macau, where clients could purchase ad space. While in Hong Kong, it is now experimenting with hotel self-service check-in and check-out robot integration. 

“We are also further developing our Intelligent Cloud Platform (ICP) to integrate AI solutions with our robots that we believe will create value by enhancing the customer experience while improving operational efficiency.,” added Yap. 

Advertisement

META announces new AI project to build AI-Language Models akin to Human Brain

META language model
Image Source: KDNuggets

Meta, the parent company of Facebook, Instagram, and WhatsApp, has unveiled a new AI research initiative aimed at better understanding how the human brain processes language. Meta AI is examining how AI language models and the brain respond to the identical spoken or written words in partnership with Neurospin (CEA) and INRIA. 

Meta AI has spent the last two years analyzing how the brain interprets words and phrases using deep learning algorithms applied to public neuroimaging data sets. Several academic organizations, including the Max Planck Institute for Psycholinguistics and Princeton University, gathered and shared the data sets. Each institution acquired and shared the data sets with the volunteers’ informed consent in compliance with the legal regulations established by their respective ethics committees, including the authorization sought from the study participants.

Language Model is an artificial intelligence (AI) model that has been trained to anticipate the next word or groups of words based on the preceding words or phrases. It is part of the technology that predicts the next word you want to type on your phone, allowing you to finish your message sooner. The early versions of language models used classic statistical approaches such as N-grams, Hidden Markov Models (HMM), and followed rule-based linguistic criteria to learn the probability distribution of words. In contrast, language models of today are created using neural networks that have been trained on vast amounts of textual data. These AI models that most closely resemble the human brain to anticipate the next word using natural language processing (NLP). 

Despite having profound scope, these models fall short of performing at par with the human brain. For instance, while a human baby can simultaneously parse that ‘orange’ can refer to both a fruit and a color, machines fail to draw such a correlation. Hence researchers at Meta are on a mission to understand the working of the human brain that can help them develop better language models. The believes that insights from this study can offer enough ideas and guidance in the pursuit of developing AI that processes speech and text as efficiently as people. 

Speaking language, according to Jean-Rémi King, a senior research scientist at Meta AI, makes humans completely unique, and knowing how the brain works is still a problem and a work in progress. The fundamental question, according to King, is “What makes humans so much more powerful or so much more efficient than these machines? We want to identify not just the similarities, but pinpoint the remaining differences.”

The researchers modeled hundreds of brain scans while also using magnetoencephalography (MEG) scanner to take pictures every millisecond, using public neuroimaging datasets extracted from images of brain activity in magnetic resonance imaging (MRI) and computed tomography (CT) scans of participants. Working with INRIA, they compared a number of language models to the brain reactions of 345 volunteers while they listened to complicated narratives using functional magnetic resonance imaging (fMRI).

Next, AI systems were given the same narratives that were read or presented to human beings. The researchers then examined the two sets of data to identify where they overlapped and where they didn’t. Meta researchers inferred from their findings that language models that closely reflect brain activity are the best at predicting the next word from context like “on a dark and stormy night…” or “once upon a time…”. Self-supervised learning (SSL) in AI is focused on prediction based on partially visible inputs, and it might be crucial to how individuals acquire the language. 

The results of Meta AI revealed that particular brain areas, such as the prefrontal and parietal cortices – situated in the front and middle of the brain, respectively – better-represented language models with far-off future word predictions. In other words, specific parts of the brain anticipate words and concepts months quite in advance, whereas most language models today are trained to predict the next word. Unlocking this long-term predicting power might aid in the advancement of contemporary AI and language models.

The researchers also discovered that the human brain can learn with just a few million phrases and can adapt and retain knowledge in its trillions of synapses on a constant basis. Conversely, AI language models can parameterize up to 175 billion artificial synapses after being trained on billions of phrases.

Read More: Meta launches Large Language Models for AI Researchers

Meta AI researchers and NeuroSpin are presently working on an original neuroimaging dataset to refine their research. This will be open-sourced, along with code, deep learning models, and academic papers, to facilitate further AI and neuroscience research. The intent, according to King, is to create a set of tools that other peers in academia and other fields can utilize and benefit from.

He added that researchers can significantly enhance contemporary AI language models by researching long-term predicting capacity in more depth. The META team believes that by incorporating long-term forecasts into algorithms, they can become more parallel to the brain.

Advertisement

Amazon to Create 2500 Corporate and Tech Jobs in California

Amazon 2500 corporate tech jobs

Technology giant Amazon announces its plans to create 2500 additional corporate and tech jobs in California for its effort to expand in Santa Monica, Irvine, and San Diego. 

Amazon has signed a lease for a 200,000-square-foot space with J.P. Morgan Asset Management at the Water Garden, managed by CBRE, to house the new employees over the coming years.

As part of its expansion strategy, the company is looking to hire local talent and continue developing and innovating on behalf of its clients. 

Read More: BMU announces BSc CS with specialization in Cyber Security, Data Analytics, Data Science, and AI

Amazon is hiring for positions in software development, engineering, game design, user experience, and other areas across Amazon teams such as retail, games, operations, and Amazon Web Services. 

This new expansion comes after Amazon launched more than 15 locations around Southern California in 2021, resulting in the creation of over 17,000 employees. Candidates from California can now take advantage of Amazon Returnship, a program designed to assist professionals who have lost or quit their jobs, including those displaced due to the COVID-19 pandemic. 

Holly Sullivan, Vice President of Worldwide Economic Development at Amazon, said, “Communities across California have welcomed Amazon over the past decade and have seen firsthand how our investments can unlock new opportunities for their neighbors and local economies. We’ve created more than 170,000 jobs across the state and are committed to continue investing here.” 

Sullivan further added that building cloud infrastructure, refining the Alexa experience, and inventing cutting-edge video games are among the 2,500 new opportunities. The company claims that these jobs offer an incredible opportunity for Californians from all walks of life to join Amazon and create a great career. 

Moreover, people interested in working in technology can also use Amazon’s newest recruiting initiative, Best Fit. This unique program allows software engineers to apply once and be considered for multiple jobs across several Amazon teams.

Advertisement