Home Blog Page 144

Key announcements at Google AI@ Event 2022

November 8, 2022

Google held its inaugural AI@ Event (Google AI@ Event 2022) last Wednesday at the company’s Pier 57 offices in New York City to highlight its latest work in AI technology. The event focused on new advances as well as early research from Google AI and research teams in climate change, Generative AI, Language Translation, Health AI, disaster management, responsible AI, accessibility, healthcare, and creativity. Here is a rundown of all the announcements made during the event.

Flood Hub

The business declared the opening of a brand-new improved platform called Flood Hub, which aims to analyze enormous meteorological data sets to show the potential for flooding in various counties. Google has been using AI to anticipate floods since 2018, warning consumers through Google search and Maps. It began utilizing AI to anticipate flood patterns in India’s Patna region in 2018. Three years later, an enlarged version of the technology-assisted in reaching an estimated 23 million individuals in India and Bangladesh with 115 million flood alerts using Google Search and Maps. As part of the latest update, this feature is available in 18 additional countries, viz., Brazil, Colombia, Sri Lanka, Burkina Faso, Cameroon, Chad, the Democratic Republic of Congo, Ivory Coast, Ghana, Guinea, Malawi, Nigeria, Sierra Leone, Angola, South Sudan, Namibia, Liberia, and South Africa. Google said that if FloodHub performs as planned, it may be able to anticipate floods in affected regions up to seven days in advance

Wildfire Detection System

Google began providing a map feature that instantly displays wildfire boundaries to users in the US in 2020. Google introduced a wildfire layer to maps in 2021, with the US still having more granular tracking of specific incidents. Since July, the company has kept tabs on more than 30 significant wildfire incidents in the United States and Canada.

During the Google AI@ Event 2022, Google said it is introducing an improved AI-powered wildfire monitoring system to the United States, Canada, Mexico, and some regions of Australia. The system uses machine learning algorithms built on satellite data from NASA and the National Oceanic and Atmospheric Administration to track and forecast the development of wildfires in real-time. The feature’s initial focus is assisting first responders in deciding how to put out the fire effectively. Since the largest change is on the back end, users might notice few differences in how they use the product.

AI Test Kitchen for Imagen

In May, Google released its AI text-to-image model, Imagen, which uses diffusion models to create high-resolution images mapping noise back to data. Google revealed at the AI@ Event that Imagen would be added to the AI Test Kitchen app for season 2, albeit only as a restricted Imagen demo. The tool can only be used to generate AI-generated content using the two demos: City Dreamer and Wobble. In contrast to City Dreamer, which lets users create photos of themed cities, Wobble lets users build artificial intelligence-created monsters using similar language prompts. Google uses the AI Kitchen app as a platform to test some of their AI models and solicit user input. It aims to minimize substantial inadvertent harm by letting consumers see its technological prowess and offer feedback through these incredibly constrained use cases. A condensed version of the company’s controversial LaMDA chatbot may also be found on the app.

AI Text Kitchen is now accessible in English for Android and iOS users in Australia, Canada, Kenya, New Zealand, the United Kingdom, and the United States.

Generative Video Content

For the first time, Google used Imagen Video and Phenaki to produce a long coherent video from text prompts for the Google AI@ Event 2022. Imagen Video is a text-to-video generative AI model that can create high-definition videos from text input. It is, to put it simply, an extension of Imagen. The text-conditioned video diffusion model is capable of producing movies with a maximum resolution of 1280×768 at a frame rate of 24 fps. Like Imagen Video, Phenaki is a language model capable of realistic video synthesis when given a series of textual prompts.

The video demo showed a single blue helium balloon that was the subject of Google’s AI-generated, super-resolution video, where viewers could see it move across a park before hitting a stray giraffe. The film was interspersed with a series of relevant written commands that were shown on an adjacent screen every few seconds.

Speaking at the AI@ Event, Google Brain Principal Scientist Douglas Eck claimed it’s quite challenging to produce videos with high quality and coherence in time. Movies, or really any other media aiming to employ pictures and videos to create a cohesive tale, rely heavily on that mix of visual quality and continuity over time.

Wordcraft

Google is also paying attention to the generative side of conversational AI despite the market for visual synthetic media’s explosive expansion. The software titan said that it had initiated early-stage experimentation with Wordcraft, a text generator built on its LaMDA dialog system. Wordcraft, in contrast to other text-editing programs like WordTune or Grammarly, aims to aid in the creation of fiction rather than merely enhancing spelling and grammar. On the Wordcraft Writers Workshop website, a group of 13 authors have been using Wordcraft to create brand-new stories that you can read.

1000 Languages Initiative

At the Google AI@ Event 2022, Google announced its ambitious 1,000 Languages Initiative, an effort to create a single AI language model that supports the top 1,000 languages spoken worldwide, including Luganda, Swahili, Lingala, and others. The company created a Universal Speech Model (USM) for this project that has been trained in more than 400 languages. To gather audio samples of various regional languages, Google is also collaborating with South Asian local governments, NGOs, and academic institutions.

Google already has a vast language portfolio, but it wants to continue. More than 7000 languages are spoken worldwide, but just a few are represented online today. The initiative will concentrate on improving representation while AI models are being trained.

Robot to write codes

The amount of work required to train a robot to carry out new jobs can be decreased thanks to an internally developed software tool that was debuted by Google at AI@ Event 2022. The tool, Code as Policies, or CaP, claims that developing code for large language models may be repurposed to produce robot policy code in response to directions in natural language. The objective is to enable robotic systems to develop their own code. The intention is to spare human developers the trouble of having to go in and reprogram things when new information comes in.

It’s available on GitHub under an open-source license.

CaP relies on Google’s PaLM-SayCan paradigm for robots to interpret open-ended human suggestions and reply appropriately and safely in a physical setting. It expands on the PaLM-SayCan research by allowing language models to carry out complex robotic tasks using the full expression of Python code for general-purpose jobs. With CaP, Google advocates leveraging language models to urge robots to create code directly. Apart from PaLM-SayCan, CaP also builds on previous work in automated code completion, like GitHub’s Copilot functionality and OpenAI’s GPT-3.

In addition to writing new code, the tool may use software libraries, which are pre-packaged collections of code that perform common activities. CaP also uses third-party libraries and APIs to create the most appropriate code for a given situation, including supporting instructions in non-English languages and even emojis.

CaP’s capabilities were evaluated by Google researchers in a series of internal tests. In one experiment, the researchers examined whether CaP could instruct a robot to move toy blocks around a table. When given the directive to “arrange the blocks in a square around the middle,” CaP was able to produce code that enabled the robot to do just that.

Healthcare AI

With a mission to offer accessible healthcare solutions, Google announced it is creating a low-cost ultrasonic tool in collaboration with Northwestern Medicine to help nurses and midwives in underdeveloped areas without access to sonographers. Using an Android app and a portable ultrasound monitor, nurses and midwives in the U.S. and Zambia are trialing a system that assesses a fetus’ gestational age and position in the womb. By detecting problems early in pregnancy, the AI application will assist medical professionals in collecting and interpreting ultrasound pictures and providing timely healthcare.

At the AI@ Event, Google also assured it would expand its collaboration with caregivers and public health authorities to provide access to diabetic retinopathy screening using its Automated Retinal Disease Assessment technology (ARDA). By clicking a photo of their eyes with their smartphone, more than 150,000 people have undergone screening.

Responsible AI

In an effort to reaffirm the company’s commitment to Responsible AI, Google Vice President of Engineering Research Marian Croak highlighted certain possible drawbacks offered by the technology on show at the Google AI@ Event 2022. These include concerns about prejudice and toxicity being amplified by algorithms, deep fakes further eroding faith in news, and false information that can make it difficult to tell what is true from what is false. According to Croak, part of that process include doing research to enable people to have more influence over AI systems so that they could collaborate with systems rather than having the machine take complete charge of situations.

Croak asserted that she believes Google’s AI Principles prioritize people, the avoidance of damage, and safety over its standard economic objectives. She claims that Google conducts adversarial testing on a continuous and constant basis. Then, their researchers make sure that they are establishing quantitative standards across all of the aspects of their AI technologies that can be evaluated and confirmed. These initiatives are being carried out by a wide group of researchers, including social scientists, ethicists, and engineers.

Bengaluru Metro launched a QR ticketing service on WhatsApp

Manjiri Gaikwad

November 8, 2022

WhatsApp and Bangaluru Metro Rail Corporation Limited (BMRCL) have recently launched Namma Metro’s first WhatsApp-based chatbot QR ticketing service.

Namma Metro is happy to give additional facility for commuters- QR Code based Single Journey Metro Travel Tickets can be downloaded from WhatsApp from 1st Nov 2022. Commuters are requested to make use of this facility pic.twitter.com/4j6oeDMopb
— ನಮ್ಮ ಮೆಟ್ರೋ (@cpronammametro) October 31, 2022

The WhatsApp-based chatbot is integrated with Unified Payments Interface (UPI) enabled payments on Whatsapp, allowing Namma metro passengers to purchase tickets and recharge their travel passes within Whatsapp. As per BMRCL, the Whatsapp-based chatbot is the first transit service to enable end-to-end QR ticketing on WhatsApp. The chatbot is available in both English and Kannada languages for passengers.

The Namma Metro official app is readily available on the Google play store. It consists of services like purchasing QR tickets, recharging metro cards, providing feedback, using travel planners to get the nearest metro stations, fare information, and train departure times, and purchasing single journey tickets with WhatsApp payments through a UPI pin. Easy cancel and refund services are available on the QR tickets. These QR tickets are available at a 5% discount off the token fee.

The managing director of BMRCL, Anjum Parewz, said that the Whatsapp-based QR ticketing service is to provide more comfort during traveling. Passengers are already using the app to plan their trips with QR tickets from 1st November. BMRCL has collaborated with WhatsApp to avail of the QR tickets service, and no additional charge is applicable for purchasing these tickets on both platforms.

Password attacks rise to 921 every second globally: Microsoft

Manjiri Gaikwad

November 8, 2022

As per Microsoft’s digital defense report of 2022, the volume of password attacks has risen to an estimated 921 attacks every second globally, which is a 74 % increase in just one year.

According to the same report, attacks against remote management devices have increased, with more than 100 million attacks detected in May 2022. These attacks have increased to five times more than the last year.

Microsoft last year’s defense report stated that it synthesized over 24 trillion security signals daily, which changed to 43 trillion signals in 2022. Microsoft uses sophisticated data analytics and AI algorithms to understand and protect against digital threats, attacks, and criminal activity.

In 2021, Microsoft blocked 9 billion endpoint threats, 31 billion identity threats, and 32 billion email threats. To date, Microsoft has removed more than 10,000 domains used by cyber criminals and 600 used by nation-state actors. As per the report, the nation’s state actors have targeted 22% of the IT sector, 17% of NGOs, 14% of the education sector, and 10% of the government sectors.

The report claimed that 93% of Microsoft’s ransomware incident response engagements revealed insufficient privilege access and lateral movement controls. The most common factors leading to weak protection against ransomware are weak identity controls, limited data protection, ineffective security operations, and multi-factor authentication.

Microsoft comprises more than 8500 security and threat intelligence experts, including engineers, researchers, data scientists, geopolitical analysts, investigators, frontline responders, cyber security experts, and threat hunters across 77 nations.

Mayo Clinic Researchers Introduce a Novel ML-based Diffusion Model for Medical Imaging of the Brain

Disha Chopra

November 8, 2022

mayo clinic introduces diffusion model for brain imaging

Mayo Clinic researchers have been looking beyond the standard generative ML models for realistic medical imaging. They have come to introduce DDPMs (denoising diffusion probabilistic models), an ML-based diffusion technique to generate medical images. DDPMs are a relatively new class of generative ML models that enables the generation of labeled synthetic images.

Traditionally, generative ML models are used to learn from medical imaging data and generate realistic images that are not patient-specific. Researchers use these synthetic images to study medical conditions and abnormalities without compromising patient privacy. The only drawback is that such models generate unlabeled imaging data, which is selectively helpful in real-world applications.

However, the volume of medical imaging data with specific abnormalities is considerably less than that with common pathologies. This results in insufficiently large and imbalanced imaging datasets used for training the models, making them less accurate.

Diffusion models are based on the Markov chain theory and generate synthetic output by ‘gradual denoising’ of an image having Gaussian noise. The process is repeated for all images, making these models run significantly slower than others. Nevertheless, diffusion models outperform other generative models as they can extract more representative features from the input medical imaging data.

In proposing this research, the researchers created a tool that can retrieve 2D axial image slices from the FLAIR sequence of a brain MRI and inpaint a pre-defined area of that slice with a realistic image. The inpainted image can represent components, including the surrounding edema, tumors, or tumor-less brain tissues.

The diffusion model presented in ‘MULTITASK BRAIN TUMOR INPAINTING WITH DIFFUSION MODELS: A METHODOLOGICAL REPORT‘ will enable medical practitioners to induce/remove tumoral and tumor-less tissues using brain MRI slices with limited data. The researchers have also provided the code online for people to use.

Twitter loses over one million followers since Musk’s takeover

Sahil Pawar

November 8, 2022

Twitter loses over one million followers

One week after Elon Musk bought Twitter, the social media platform has reportedly lost more than one million followers.

“We have seen an uptick in people deactivating their accounts and Twitter suspending accounts,” said the founder of the anti-disinformation platform Bot Sentinel, Christopher Bouzy.

MIT Technology Review reported that Bot Sentinel has analyzed 3.1 million accounts on Twitter. It believes that around 877,000 accounts were deactivated, and 497,000 were suspended between October 27 and November 1. That’s more than double the usual number.

Bot Sentinel’s analysis overlooked a percentage of users with suspended or deactivated accounts before applying that data to the overall 237 million daily active users on Twitter. Based on the 5,958 accounts deleted or suspended the week before Musk’s takeover, it indicates a 208% increase in lost accounts.

Bouzy told MIT Technology Review that the increase in suspensions is from Twitter taking action on accounts purposely violating Twitter’s rules to see if they can push the limits of free speech. He believes some users are testing what can and will not stay posted, such as posts using hate speech.

The Infinite Conversation: Can AI chatbots converse for hours like humans?

Preetipadma K

November 8, 2022

A website called The Infinite Conversation, built using artificial intelligence and featuring a perpetual conversation between virtual avatars of Slovenian philosopher Slavoj Žižek and German director Werner Herzog, was launched last week by Italian artist and programmer Giacomo Miceli.

Through warning from virtual Žižek, Miceli shares that the initiative seeks to increase awareness about the simplicity of utilizing technologies for synthesizing a human voice. He believes any motivated user without in-depth technical knowledge can accomplish this feat today from their bedroom with a laptop. Miceli warns that this can alter how we interact with the online content we consume while posing issues with the value of reliable sources, betrayal of trust, and gullibility.

Miceli points out that by the end of 2022, it will be cheap and simple to create AI-generated content that makes it more appealing on the surface and is remarkably similar to the “real thing.” This holds true for recordings that seem like famous people (often referred to as Deepfakes) or speech, as in the instance of the Infinite Conversation.

The talented artist is said to have built the website using “open source tools available to anybody,” denying providing technical details but hinting that he could publish an explanation piece this week. In the site’s FAQ, he explains that the script was generated using a popular language model that has been improved on interviews and content created by each speaker.

Visitors will view AI-generated charcoal pictures of the two guys in their profile when they visit the website. A transcript of AI-generated text is marked in yellow and read aloud between them by voices that imitate Herzog or Žižek. You may jump between each part by clicking the arrows underneath the portraits as the conversation moves back and forth between them with their individual accents.

UK bank Santander to block crypto exchanges next year

Sahil Pawar

November 8, 2022

UK bank Santander block crypto exchanges

UK bank Santander will block real-time payments for crypto exchanges next year. According to an email to customers, the move is intended to protect customers from scams. Santander has not disclosed when the change will take effect in 2023. The bank will enforce a more limited set of restrictions in the short term.

Payments for cryptocurrency exchanges using online banking will be limited to £1,000 per transaction from November 15 onwards, with a limit of £3,000 in total in any rolling 30-day period. The new rules will not affect the ability of consumers to make withdrawals.

“In recent months, we have seen a rise in UK customers becoming victims of crypto fraud,” said a Santander spokesperson. “We will do everything we can to protect our consumers, and we feel that limiting payments to crypto exchanges is the best way to ensure your money stays safe.”

Santander will continue to block payments that are being sent to Binance in light of the UK Financial Conduct Authority’s (FCA) harsh stance on the exchange, which was banned from operating in the UK in 2021 by the watchdog group.

The FCA claimed the firm is “incapable of being effectively supervised” and its “complex and high-risk financial products” pose a significant risk to consumers. But not all UK banks are pulling back from crypto. Neobank Revolut, operating in the UK since 2015, recently launched a card that allows users to pay in crypto for their goods and services.

Researchers Propose a Novel AR Localization and Mapping Technique, ‘LaMAR’

Disha Chopra

November 8, 2022

Researchers from Microsoft and ETH Zurich, one of the leading universities in research and innovation, have proposed a novel AR localization and mapping technique called ‘LaMAR: Benchmarking Localization and Mapping for Augmented Reality .’ The proposed technique appears to overcome the general challenges researchers encounter while mapping in AR.

AR technology has been around for a while and is simply about placing a virtual object in the real world and tracking its location and shape over time. However, visualization and mapping have yet to be done for the same.

There are a few challenges that researchers face while mapping AR. Firstly, the AR-specific devices are mostly smartphones equipped with multiple cameras and sensors, making it difficult to map in a single-camera environment. Additionally, AR devices offer spatially-posed sensor streams as they can locally track in real-time. However, objects in an AR scenario may change over time, requiring more than just local tracking.

Large-scale AR mapping requires robust algorithms and devices that can keep up with quickly changing data in AR scenarios. With LaMAR, researchers developed a robust system and benchmark for AR localization and mapping.

LaMAR introduces a large-scale dataset of captured AR images, including indoor and outdoor scenes with/without illumination. Secondly, LaMAR provides a pipeline to produce accurate AR trajectories and handle crowd-sourced data from multiple devices, overcoming many standard challenges in AR mapping and localization.

LaMAR is highly scalable and precise in mapping augmented reality scenarios, as seen from the performance metrics mentioned in the paper. Researchers plan to continue developing the technique further.

Meta’s new AI solves International Math Olympiad problems

Manjiri Gaikwad

November 7, 2022

Meta introduced a new AI system that solves complexes International Math Olympiad (IMO) problems. The new model acquires 67 percent of the miniF2F validation set accuracy and is 5x more than the previous AI system.

In December 2019, Facebook built its first AI system for solving advanced mathematical equations with the help of symbolic reasoning. The first AI system solves integration problems, first-order, and second-order integration problems. The AI model demonstrated 99.7 percent accuracy for integration problems and 94 and 81.2 percent for first and second-order integration problems, respectively.

Since solving complex equations requires precision rather than an approximation, the launch of Meta’s new AI in 2022 is a milestone. Meta has used the HyperTree Proof Search (HTPS) method in its new AI system, which is trained on a dataset of successful mathematical proofs and is generalized to new or different problems. This method can collect correct proofs for the IMO problem, consisting of some arithmetic reduction to a finite number of cases.

Meta has mentioned the detailed work of the new AI system in its new research paper on HTPS that will be presented in NeurIPS 2022. The new AI system is made available with the Lean Visual Studio Code (LVSC) plugin, which allows other researchers to explore the system. Lean is a functional programming language to write correct and maintainable codes.

Will Shanghai Upgrade bring any fee reduction for Ethereum users after The Merge ‘Disappointment’?

Preetipadma K

November 7, 2022

Ethereum Merge gas fees shanghai upgrade

In September, the second-biggest cryptocurrency by market cap, Ethereum, underwent a software update called “The Merge.” The Merge is a meticulously planned changeover from a Proof of Work consensus mechanism to a long-awaited Proof of Stake consensus mechanism. Using a consensus process, numerous anonymous participants can agree on a transaction so that it can be recorded in a ledger. This Merge signifies a substantial shift in how the Ethereum blockchain verifies a block, eventually changing how transactions are handled, lowering the ether token’s supply and considerably increasing its price.

Every user who wants their transaction to be included in a block on Ethereum must pay transaction costs, sometimes known as “gas fees.” The space on the block is awarded to the bidder, who offers the maximum amount of gas. As there is greater demand for Ethereum block space, fees have skyrocketed in recent years. While many anticipated that the gas fees would reduce post The Merge, unfortunately, that is not the case. This is because the purpose of the update was to strengthen network security, making it nearly difficult to attack the network. It will also use less energy, by around 10,000% less energy, making it more energy-efficient.

Contrary to popular belief, gas prices will most likely stay the same as a result of The Merge since the network won’t become significantly quicker or able to handle more transactions per second. In reality, this upgrade would result in blocks being generated 10% more frequently, which is a rather negligible difference and is unlikely to be seen by users. Additionally, neither The Merge nor any other parameters affecting network capacity increase block size. In other words, no more block space will be made available by The Merge. As a result, the network continues to charge users more fees to access block space, with the amount of competition increasing as a result.

But there is a possibility of reducing the transaction fees in the next phase after The Merge.

Withdrawing staked ETH will be made feasible by the Ethereum network’s upcoming significant upgrade, “Shanghai,” which shall wrap up unfinished business from the historic “Merge” update. In October, the Shanghai testnet, code-named Shandong, was announced. The testnet will allow developers to test new protocols before the “Shanghai” update is expected to happen in 2023. Several Ethereum Improvement Proposals (EIPs) will be tested in Shandong by Ethereum’s core developers, who will then create, modify, and narrow down the number of upgrades to be included in Shanghai when it eventually goes online. These EIPs include ideas for network effectiveness, scalability, and gas prices.

According to Bloomberg, the Shanghai upgrade is anticipated to let users who pledge or stake the network’s native Ether cryptocurrency eventually remove their money from the specialized digital wallets used to place network transaction orders. Before that occurs, a few minor tweaks would be made. For example, one of them is WARM Coinbase, which will drastically lower some costs paid by prominent ecosystem participants known as builders, who already have significant influence over Ethereum.

Transactions sent via Ethereum are packaged into blocks by builders like Flashbots and BloXroute, then forwarded to validators who put them in the proper order on the blockchain. At present, Flashbots is the largest builder and distributes more than 81% of these blocks.

Some critics are concerned that this may lead to it using its influence to gain an advantage or demand more money.

88% of validators have chosen to collaborate with the builders since September, when the new method utilizing builders was introduced into The Merge to get additional fees from traders attempting to execute a range of profitable tactics. The builders are compensated for organizing transactions in a certain sequence, which can let a trader, for instance, purchase a coin before another party and resale it to them for a greater price.

One rationale for the WARM Coinbase change is that it would enhance the builders’ industry’s profitability. According to Matt Nelson, a product manager at ConsenSys, an Ethereum infrastructure provider, some fees builders pay to the network could decrease by 26 times once it is deployed.

To accept new tokens on the network and serve as a conduit to validators, some users, such as builders, must have access to coinbase, a specialized blockchain application. There may be several interactions between each of these transactions and the coinbase. The moniker “warm” the coinbase comes from the fact that it costs more to access the coinbase for the first time. Once it has warmed up, further memory accesses are less expensive. Coinbase now starts out warm and is put into memory for a much lower gas fee upfront.

1...143144145...358 Page 144 of 358