Category: Technology

The event marks the largest gathering of the world’s best data scientists in Australian history as professionals gather to transform Australia into an AI nation

Mountain view, California. & Sydney, November 13, 2022– (work wire) – H2O.ai, the leader in the AI ​​cloud, today announced a series of product improvements designed to further democratize AI by improving the ability of non-technical analysts to use the company’s powerful, award-winning technology with both structured and unstructured data. , such as award-winning Document AI, Hydrogen Torch, and autoML.

The updates were announced at H2O World, the company’s first conference in Australia. The event was held at the headquarters of the Commonwealth Bank of Australia in Sydney. Thousands of viewers joined the conference online. The event also marks the largest gathering of Kaggle Grand Masters (KGMs) in Australia. Kaggle is an international arena for data scientists to compete to solve complex data problems – H2O.ai is home to more than 30 KGMs, widely accepted as the world’s best data scientists, gathering in Sydney to advance and inspire AI across the country as noted by Sri Ambati, President CEO and Founder of H2O.ai last year.

The event featured world-class speakers including:

• Matt CominCEO of the Commonwealth Bank of Australia, engaged in a dynamic chat with Ampaty.

• Dr. Andrew McMullanChief Data and Analytics Officer, Commonwealth Bank of Australia

• Dan GermainChief Decision Scientist, Commonwealth Bank of Australia

• Sonal SoranaGeneral Manager, CommBank.ai

• Edward SantoProfessor at UTS, Co-Director of the Institute of Human Technology, former Australian Human Rights Commissioner

• Junko YamashitaDigital Strategy HQ Global Application and Data Management Center, Yokogawa

• Tania Berger Wolfdirector, co-founder and director of the AI ​​nonprofit Wild Me

• Additional partners from Dell, Accenture, McKinsey, Snowflake and SimplyAI

H2O.ai has announced the launch of the Driverless AI Wizard, an interactive tool designed to help both citizen data scientists and experienced users increase the efficiency and effectiveness of their AI models. The wizard is an essential part of democratizing the H2O.ai platform by looking at the data and recommending next steps the user can take. The main goal of the wizard is to help users avoid common pitfalls and instead use best practices that experts use to make the right choices during modeling to reduce the risk of inaccurate or poor models. The educational component of the wizard is designed to help people who do not normally work in machine learning understand basic data science concepts and uncover insights.

The story continues

The wizard is designed to assist four specific types of roles and users: data scientists, data engineers, business analysts, and domain experts. Safeguards ensure that even non-expert users can safely and securely develop and deploy AI models to help scale truly responsible AI across the enterprise.

In addition, H2O.ai has announced updates to its new tagging tool for deep learning use cases called Label Genie, including bullet-less learning, support for voice uses, and more. H2O.ai’s H2O Hydrogen Torch Engine has received new updates to compare experiments, automate network search, and enhance NLP interpretability.

These innovations come less than 12 months after the company launched its Document AI product. H2O Document AI helps organizations quickly and accurately process unstructured documents and text data to increase productivity and find hidden insights. H2O Document AI delivers automation not possible before by combining intelligent character recognition (ICR), natural language processing (NLP), computer vision and planning intelligence. The service has flexible processing of out-of-box and post-document, and integrates seamlessly with existing customer business processes and workflows for structured and unstructured data.

CBA currently has hundreds of employees leveraging AI using the H2O.ai platform across a wide range of use cases across the group. With Document AI, the bank has now expanded its machine learning capability to integrate digital documents, with the ability to process millions of files using a suite of custom-trained models that can be customized to any format or use case.

The bank explained how Document AI has been used to augment its existing Customer Intake (KYC) approach, a largely manual process that relies on large amounts of digital documents. Leveraging AI to process such large amounts of unstructured data could transform the process at scale, by reducing human error, improving accuracy, and delivering a better customer experience.

“In order to better serve our customers, we must continually strive to keep systems up to date and ensure we take advantage of best-in-class global technology,” said Dr Andrew McMullan, Director of Data and Analytics at Commonwealth Bank of Australia. “Through our partnership with H2O.ai, we are unlocking the huge potential of AI to deliver better services and experiences to our customers and communities.”

H2O.ai is strongly committed to democratizing artificial intelligence to improve the world through its AI 4 Good initiative. The company focuses on education, conservation and healthcare to improve the world. As part of the global H2O agenda, Tanya Berger Wolf spoke for the conservation nonprofit she co-founded, Wild Me. She noted the use of artificial intelligence in support of the organization’s mission. The organization focuses on protecting the planet’s biodiversity through input from researchers, citizen scientists, robotic cameras and social media across users with varying degrees of data science expertise, demonstrating a commitment to democratizing AI for all.

With the company returning to in-person events, they also staged H2O World in Dallas, Texas on November 17 next week at AT&T.

For more information on the H2O World event organized by H2O.ai, the company’s wizard tool or the AI ​​4 Good work program, visit https://h2o.ai/.

About H2O.ai

H2O.ai is a leading cloud AI company, on a mission to democratize AI and lead an open AI movement around the world. They focus on extracting insights from structured and unstructured data such as video and documents through their award-winning products such as Hydrogen Torch and Document AI. Customers use H2O AI Cloud to quickly solve complex business problems and accelerate the discovery of new ideas. H2O.ai is the trusted AI provider for more than 20,000 global organizations, millions of data scientists and more than half of the Fortune 500, including AT&T, Commonwealth Bank of Australia, Citi, GlaxoSmithKline, Hitachi, Kaiser Permanente, and Procter & Gamble, PayPal, PwC, Reckitt, Unilever, Goldman Sachs, NVIDIA and Wells Fargo are not only customers and partners, but also strategic investors in the company. Over 30 Kaggle Grandmasters (the community of best machine learning practitioners and data scientists in the world) are the makers of H2O.ai. Strong AI for a good spirit to make the world a better place and responsible AI driving the corporate purpose. Please join our movement at www.h2o.ai.

View source version on businesswire.com: https://www.businesswire.com/news/home/20221113005082/en/

Contacts

Half:
Anthony Gomez
press@h2o.ai

there is new Samsung Galaxy S23 Leak to report: This time the alleged photo taken with the S23 Ultra has been compared to similar shots taken on Samsung Galaxy S22 Ultra and the Google Pixel 7 Pro.

This comes from a well-known tip Ice world (Opens in a new tab) (Across note book (Opens in a new tab)), who initially posted a series of three enlarged photos of the pumpkin without naming the phones involved, and asked to vote which of the photos was the best.

The Galaxy S23 Ultra won the vote, beating this year’s Ultra version and Google’s latest flagship—but You can report for yourself Which one do you prefer? The shot taken with the Galaxy S23 Ultra definitely looks the most detailed and vibrant.

Camera Specifications

We have already I saw another picture From the same leaker purportedly from the Samsung Galaxy S23 Ultra, so it looks like Ice Universe may have owned a unit beforehand – the phone isn’t expected to launch until February 2023.

This means that Samsung still has plenty of time to tweak its photo optimization algorithms on the S23 phones, making sure that photos and videos look the best they can be (in fact, this usually continues after the phone is released as well).

We expect the Galaxy S23 Ultra to come with a triple-lens rear camera that’s not much different than the Galaxy S22 Ultra — although There is talk It will greatly upgrade the capabilities of low-light photos.

Analysis: Cameras keep getting better

Best phones of 2022 – incl. iPhone 14 Pro And the Google Pixel 7 Pro – it can take great photos and pictures with quality you wouldn’t believe when cameras were first attached to the back of smartphones.

However, these cameras continue to improve, due to improvements in two areas: hardware (the actual sensors installed inside phones) and software (how the phone interprets and processes the information it gets from the cameras).

Google may be best known for the way its algorithms can create great-looking shots even with not-so-first-class camera sensors, but other phone makers are also busy improving their core code, including Samsung.

That’s why we’ll have to reserve judgment on the Galaxy S23 Ultra’s camera until we can actually test it out – camera specs alone won’t tell the full story, even if the main sensor gets a significant bump up to the 100MP mark.

Subtract from the big Windows 11 2022 Update (22H2) didn’t run very smoothly, and there’s now another issue to report: Microsoft says some users may experience “lower-than-expected performance in some games”.

This is the latest entry in List of known problems (Opens in a new tab) Managed by Microsoft (via computer games (Opens in a new tab)), and are described as games that slow or stutter when played. If it’s something you’ve noticed since installing the update, you’re not alone.

According to Microsoft, this is because some games and apps “unintentionally enable GPU performance-correction features that are not intended for use by consumers” — so it’s a feature built into Windows 11 that turns on when it’s not running.

You can fix it?

Microsoft has paused the release of the 22H2 update for PCs it thinks may be vulnerable to the error, but if you’ve already installed it, the official advice is that updating affected software and games to newer versions “may” help the problem.

Other than that, Microsoft says it’s “working on a solution and will provide an update in an upcoming release” — in other words, you’ll need to wait for a fix to be included in a future Windows 11 patch.

You may still be able to roll back to your previous version of Windows 11 while Microsoft sorts its icon: From Settings, choose System, then Recovery, then Back to go back to the software version without applying update 22H2.

Analysis: bumps in the road

When you update millions of computers, it’s inevitable that some of them will have some problems – but the goal is always to make sure those issues are as simple as possible and affect as few devices as possible.

It’s hard to assess the prevalence of this problem, though, given that Reddit thread about the problem (Opens in a new tab) It has attracted hundreds of comments, and we believe a large number of games will be affected by this issue.

we Previously mentioned On a large portion of users who are experiencing gaming performance issues after applying the 22H2 update to their devices, it seems likely that this is the same set of issues, which have been officially recognized and acknowledged by Microsoft.

Looking into this major update also caused problems with printers And the WindowsMicrosoft hopes you can fix this gaming issue ASAP – otherwise take up Windows 11 is will remain low.

The scientific world has long recognized that proving mathematical theorems is an essential first step in the development of artificial intelligence. To prove the correctness or falsity of a conjecture, one must use symbolic reasoning and sort through an unlimited number of alternatives. These tasks are beyond the capabilities of even the most complex artificial intelligence systems.

The cutting edge of artificial intelligence today is the creation of machines that can “solve on the spot” or come up with a complete answer to a problem in one go. However, this is not the way most individuals deal with difficult situations. Mathematical reasoning is more difficult to formalize and measure.

Meta AI is an important development at the intersection of artificial intelligence and mathematics. The team’s neural theory installer completed five times more IMO problems as any other AI system before it, for a total of ten. In terms of miniF2F, a popular math test, the AI ​​model outperforms the latest model by 20% and outperforms Metamath by 10%.

The team’s HyperTree Proof Search (HTPS) approach is taught to generalize from a data set of valid mathematical proofs to entirely new challenges. Using some computational shorthand for a finite number of examples, he can deduce a suitable proof for the problem IMO.

Neurological theory has demonstrated that it is required to relate a “state” to an existing (incomplete) understanding of the problem in order to behave like a human being. Initially, the researchers used a reinforcement learning strategy combined with pre-existing proof tools such as Lean.

The “current state” of the (incomplete) proof is a node in the graph, and each subsequent step is an edge. This allows visualization of the directory as a whole. Proof aides use the process of deductive reasoning to make this method possible.

The AI ​​for playing chess requires a similar mechanism: the ability to assess the strength of a given chess position, and in this case, a proof state. To achieve this, the team adopted a strategy reminiscent of the Monte Carlo Tree Research (MCTS), in which the model iteratively estimates that:

1. A set of plausible counterarguments to use in a particular case of proof
2. The result of the proof after a specified number of counterarguments are presented.

This allows for online training to significantly improve the performance of a previously tested prototype on specific issues.

As a result, the proposed method outperforms the currently published latest by 20% over the accuracy of the Minif2f validation set and solves ten previously unresolved problems IMO. The team hopes their work will help the community build on their work so we can all take more rapid steps forward in this wonderful field.

check the paperAnd the ToolsAnd the reference article. All credit for this research goes to the researchers on this project. Also, don’t forget to join Our page on redditwhere we share the latest news on AI research, cool AI projects, and more.

Amazon has unveiled “Sparrow,” a new intelligent robotic system that aims to simplify the fulfillment process by moving individual products before they are packaged.

Unlike the retail giant’s previously deployed robotic arms, which include the likes of Cardinal and Robin, Amazon claims the Sparrow arm can identify about 65% of its product inventory without human assistance.

Sparrow uses computer vision and artificial intelligence (AI) to identify and manipulate millions of items, simplifying the process and saving human energy, and the company says Amazon employees around the world have picked, stored or packed nearly 5 billion packages — or more than 13 million packages per day.

Why is Amazon betting on bots?

An Amazon spokesperson said in Blog Posts Announcing the news.

“By working with our employees, Sparrow will take on repetitive tasks, enabling our employees to focus their time and energy on other things, while also enhancing safety.”

Sources indicated that the retail giant is in dire need of automated intervention.

Leaked internal search, originally reported by vox (Opens in a new tab)He claimed that Amazon could run out of employees willing to work in its warehouses by 2024 if it did not change its hiring practices.

“If we continue business as usual, Amazon will exhaust the available labor supply in the US network by 2024,” the research states.

Amazon has always bet on robots, and in 2012 it made a major investment to improve productivity in the supply chain by acquiring Kiva, a robotics company whose technology is still in heavy use ten years later.

Robots don’t take human jobs, they also create them if Amazon’s claims are to be believed.

The company estimates that more than 700 new job categories have been created across its operations to date thanks to the deployment of automated technology

• Interested in making your own foray into the online retail world? Check out our guide to Best e-commerce platforms

New research indicates that the majority of today’s apps are built with open source software (OSS), which once again demonstrates its importance

according to Octoverse 2022 Report from GitHub (Opens in a new tab) In the case of open source software, 90% of Fortune 100 companies use OSS for some capacity, compared to the results of the company’s first report in 2012 where companies were only using OSS to power their web servers.

This year alone, there were 413 million contributions from OSS to GitHub from 94 million users of the platform, the company indicated.

Open source software in 2022

Commercially backed open source software projects are on the rise, the report added, indicating that about a third of Fortune 100 companies now have an Open Source Software Office (OSPO) to coordinate their OSS strategies.

Commercially backed projects have proven particularly popular among first-time shareholders, who make up about half of their business.

Not all news is good, though Summary of the Open Source Security and Risk Analysis (OSSRA) Report (Opens in a new tab) For 2022, it found that despite a steady 3% year-over-year decrease in vulnerabilities, more than 80% of the analyzed code bases still contain at least one vulnerability.

The results showed a promising reduction in high-risk vulnerabilities, showing a reduction of 11%.

This may in part be due to a lack of maintenance, as 88% of the code bases investigated have not shown any signs of updating in the past 24 months.

According to GitHub VP of Developer Relations, Martin Woodward, who claims that OSS is the foundation of 99% of the world’s software, “there are a number of benefits to open source, from providing an environment to work quickly and flexibly, to enabling collaboration from developers around the world.”

Across VentureBeat (Opens in a new tab)

There are two main types of companies that investors are avoiding this year: those that operate in the technology sector, and those that serve consumers. Unfortunately Lemonade (LMND 7.14%)Both artificial intelligence-driven insurance company.

Inflation is rising, forcing the US Federal Reserve to raise interest rates aggressively to try to push them back toward a more natural rate. As a result, consumers are affected by a combination of higher prices for everyday goods and services and higher borrowing costs. Along the same lines, tech companies struggle when the cost of capital rises because money is hard to come by to fund growth.

But these headwinds won’t last forever. Lemonade just posted another strong quarterly result for the period ending September 3. 30, and believes it can be profitable with the money it already has. With its stock price down 88% from its all-time high, that’s why investors may want to buy now.

Lemonade difference

Most consumers can point out at least two things they don’t like about dealing with traditional insurance companies, whether it’s the customer experience or the time it takes to process and pay a claim. The industry isn’t known for innovation – but then lemonade came along.

Lemonade has been developed Artificial intelligence (AI) to revamp customer interactions and the way premiums are paid. Its online bot, Maya, can quote an insurance policy in 90 seconds and pay claims within three minutes without human intervention. The company now operates in five insurance markets: Home Owners, Renters, Life, Pets and Auto, its latest addition.

Lemonade is also opening up new ground outside of America, having launched Content Insurance (Tenants) in the UK during the last third quarter. It follows the company’s previous entry into Germany, the Netherlands and France.

Overall, Lemonade’s unique technology-centric approach resonates with customers, perhaps because it brings insurance in line with other goods and services consumers use to purchase online with speed, ease and a high level of convenience. In the third quarter, the company had more than 1.77 million customers, which is a 30% jump over the same period last year.

In addition, its premiums per customer rose to another all-time high as more Lemonade customers purchased multiple products.

Premium strength lemonade is on the rise

Perhaps the most significant benefit from Lemonade’s third-quarter report was a 76% jump in an effective premium. It’s the dollar value of premiums for all active policies, and it has risen to a record high of $609 million over the period. This was helped by Lemonade’s recent acquisition of AI-powered insurer MetroMile, which added 32 percentage points to the score.

The explosive growth rate is also a result of Lemonade’s rapidly expanding customer base, as well as the fact that each customer continues to spend more money with the company over time.

One thing investors are concerned about is Lemonade’s total loss ratio — the percentage of premiums it pays to fund claims — which was 94% in the third quarter. This is well above the 75% level that it is striving to achieve, and it will be difficult for the company to turn a profit if it remains at this high level. It was 77% a year ago, and the jump is mainly attributed to short-term factors such as Hurricane Ian in Florida and Lemonade’s acquisition of MetroMile.

Lemonade reported a net loss of $91 million for the quarter, bringing its total net loss to $304 million over the past four quarters. But the company sees a light at the end of the tunnel because it believes it can be profitable with the money it has. This would be a huge positive for investors as it means they won’t have to put up with a diluted capital increase, especially at Lemonade’s currently pent-up share price.

Why is lemonade a buy when it dips

Lemonade’s upcoming fourth-quarter results may not be as strong as Q3, because the company pushed some of its marketing efforts forward to get a better return on investment. She says Q4 is usually weak due to seasonality.

But this is a long-running story. Take, for example, Lemonade Auto Insurance, which entered only last year. It’s a $316 billion annual opportunity in the US alone, which leaves plenty of room for growth compared to the company’s current total premiums in place.

In addition, Lemonade has just agreed to a deal with rubbery, a retailer of pet products, offers Lemonade pet insurance to 20 million customers. The partnership is set to begin in 2023, and could provide a positive boost to Lemonade’s customer base.

All things considered, Lemonade has an exceptional growth path into the future. The risk is that it fails to turn a profit in the near term as expected, but this weak economy and A pessimistic tone towards technology companies It won’t last forever. If she can withstand these conditions, she will likely find herself in a better position a year or two from now.

With Lemonade stock down 88% from its all-time high, the risk-reward equation might make sense for investors here.

Teradyne Companies Mobile Industrial Robots (MiR) AutoGuide Mobile Robots has merged to become a single supplier of Autonomous Mobile Robots (AMRs), one of the fastest growing segments of automation today. At the end of September, the integrated company became officially known as Mobile Industrial Robotics (MiR), with Teradyne CEO Walter Vahey taking over as president. It will be headquartered in Odense, Denmark, where MiR has run its global operations since its launch in 2013.

MiR develops and markets secure, collaborative mobile robots that can quickly, easily and cost-effectively manage internal logistics, freeing employees to perform higher value activities. Prior to the merger, MiR offered a wide range of AMRs capable of carrying payloads and platforms up to 3,000 lb (1,350 kg). By integrating with AutoGuide, the portfolio will expand to include high-tonnage AMR winches and towings that will run on the MiRFleet software.

According to Vahey, who led Teradyne’s acquisition of MiR in 2018 and AutoGuide a year later, customers in the global inland logistics marketplace demand a full line of AMRs spanning into low, medium and high tonnage vehicles to automate material transportation throughout the value chain. “By joining forces, we will uniquely meet our customers’ needs for in-house logistics automation, from a single resource that provides easy-to-deploy, reliable and scalable AMRs with combined fleet management software,” he says.

As part of Teradyne, these two companies have already worked together to develop a way to run AMRs on the same fleet management software.

“This close collaboration accelerated the integration process,” Fahy adds. “Being a single organization helps ensure that we deliver an extensive AMR product line running on a common fleet program to our customers.”

The company maintains a huge global presence
The new combined company employs 450 employees, including 250 engineers. MiR is already well established in the AMR market with over 7,000 AMRs sold in more than 60 countries. AMR pumps can automatically and safely pick up, move and deliver pallets or other loads in highly dynamic environments, which is why they constitute a safe and efficient alternative to traditional Automated Guided Vehicles (AGVs), forklifts and pallet trucks.

The combined MiR company now has a global distribution network of more than 200 partners worldwide. In addition to its headquarters in Odense, Denmark, MiR has regional offices in Boston. Holbrook, New York; San Diego; Chelmsford, Massachusetts; Georgetown, Kentucky; Singapore ; Frankfurt. Barcelona. Tokyo; Seoul and Shanghai.