Big Data Archives - Business Intelligence Info

Category Archives: Big Data

Intel Geospatial is a cloud platform for AI-powered imagery analytics

October 28, 2020 Big Data

Intel today quietly launched Intel Geospatial, a cloud platform that features data engineering solutions, 3D visualizations, and basic analytics tools for geovisual workloads. Intel says it’s designed to provide access to 2D and 3D geospatial data and apps through an ecosystem of partners, addressing use cases like vegetation management, fire risk assessment and inspection, and more.

The geospatial analytics market is large and growing, with a recent Markets and Markets report estimating it will be worth $ 96.34 billion by 2025. Geospatial imagery can help companies manage assets, for example network assets prone to damage during powerful storms. Moreover, satellite imagery and the AI algorithms trained to analyze it have applications in weather prediction, defense, transportation, insurance, and even health care, namely because of their ability to capture and model environments over extended periods of time.

Using Intel Geospatial, which is powered by Intel datacenters, customers can ingest and manage geovisual data from a mobile- and desktop-accessible web portal. They’re able to view slope, elevation, and other data layers in a 3D environment with zoom, pan, and tilt controls and auto-updated time and date stamps. Moreover, they can analyze the state of various target assets as well as run analytics to extract insights that can then be passed to existing enterprise systems.

Intel Geospatial offers data from satellites, manned aircraft, and unmanned aerial vehicles (UAVs) like drones, with data from Mobileye — Intel’s autonomous vehicle subsidiary — available upon request. The platform’s user interface auto-populates with area-specific datasets and allows for search based on street addresses or GPS coordinates, which are standardized for analytics.

Intel Geospatial offers out-of-the-box algorithms for risk classification, object counting, distance measuring, and public and private record reconciliation. Intel says it’s leveraging startup Enview’s AI to power 3D geospatial classification for faster lidar analytics turnaround. Meanwhile, LiveEO is delivering algorithmic monitoring for railway, electricity, and pipelines.

Intel’s new service joins the list of geospatial products already offered by companies including Google, Microsoft, and Amazon. Google’s BigQuery GIS lets Google Cloud Platform customers analyze and visualize geospatial data in BigQuery. Microsoft offers Azure Maps, a set of geospatial APIs to add spatial analytics and mobility solutions to apps. Amazon provides a registry of open geospatial datasets on Amazon Web Services. And Here Technologies, the company behind a popular location and navigation platform, has a service called XYZ that enables anyone to upload their geospatial data — such as points, lines, polygons, and related metadata — and create apps equipped with real-time maps.

The audio problem:

Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here

Let’s block ads! (Why?)

Big Data – VentureBeat

AMD is buying Xilinx for $35 billion to compete with Intel in the datacenter

October 27, 2020 Big Data

Processor and graphics firm Advanced Micro Devices has agreed to buy programmable chip maker Xilinx for $ 35 billion in stock. AMD also reported better-than-expected third-quarter earnings and raised its forecast for fourth-quarter results.

The acquisition will help AMD stay competitive with archrival Intel in the datacenter, where programmable chips have become increasingly popular. It also follows a trend for increasing consolidation in the chip industry, where Nvidia is acquiring Arm for $ 40 billion. (Intel bought Xilinx’s rival in programmable chips, Altera, for $ 16.7 billion in 2015).

AMD reported third-quarter adjusted earnings per share of 41 cents on revenue of $ 2.8 billion, up from 18 cents and $ 1.8 billion for the same quarter a year earlier. Analysts had expected 35 cents per share in earnings and $ 2.56 billion in revenues.

Shares of AMD fell 3.5% in trading on Tuesday morning, while shares of Xilinx rose 9%. That pattern is typical in acquisition deals.

“Our acquisition of Xilinx marks the next leg in our journey to establish AMD as the industry’s high performance computing leader and partner of choice for the largest and most important technology companies in the world,” said Lisa Su, CEO of AMD, in a statement. “This is truly a compelling combination that will create significant value for all stakeholders, including AMD and Xilinx shareholders who will benefit from the future growth and upside potential of the combined company. The Xilinx team is one of the strongest in the industry and we are thrilled to welcome them to the AMD family. By combining our world-class engineering teams and deep domain expertise, we will create an industry leader with the vision, talent and scale to define the future of high performance computing.”

AMD is buying Xilinx for $35 billion to compete with Intel in the datacenter

Above: Xilinx’s Zynq programmable chips are used in the Subaru Levorg.

Image Credit: Xilinx

Xilinx shareholders will get 1.7234 shares of AMD common stock for each share of Xilinx common stock that they own. That amounts to $ 143 a share for Xilinx stock, which has been trading under $ 125. While AMD has a lot less cash than rival Intel, its stock price has had a big run as it gained market share during the past couple of years with its Zen and Zen 2 chip designs. AMD’s stock price has tripled since the spring of 2018.

This quarter, AMD is shipping its next-generation server chips, code-named Milan, to cloud customers. Xilinx is in the midst of a product renewal with its programmable Versal ACAP chips, which help datacenters deal with with a tsunami of computing demands from artificial intelligence.

Over time, AMD expects to save $ 300 million via streamlining costs in the first 18 months after the approval of the acquisition. That may mean lower costs through layoffs, but the company didn’t spell that out. AMD expects the deal will close by the end of 2021.

“AMD’s intent to acquire Xilinx is a bold move that I think makes sense and caps off an incredible run by AMD,” said Patrick Moorhead, an analyst at Moor Insights & Strategy, in an email to VentureBeat. “I believe AMD will continue to organically grow with or without this acquisition, as hopefully evidenced by its monster Q3, its upward guide, and roadmap.”

He added, “I do think AMD has a brighter, long-term future with Xilinx as it creates a larger entity that is more diversified across different markets and products while leveraging similar technologies. While I was more excited by its long-term possibilities, I am now equally impressed by its day one accretion and look forward to getting more details on its short-term leverage.”

For the third quarter, AMD reported $ 1.67 billion in revenue from computing and graphics, up 31% from a year earlier. It reported $ 1.13 billion in revenue for its enterprise, embedded, and semi-custom chip division. That’s up 116%, in part because AMD is supplying chips to the Microsoft and Sony game consoles launching in November.

The audio problem:

Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here

Let’s block ads! (Why?)

Big Data – VentureBeat

Researchers design virtual environment to spur development of helpful home robots

October 25, 2020 Big Data

Without much prior experience, kids can recognize other people’s intentions and come up with plans to help them achieve their goals, even in novel scenarios. By contrast, even the most sophisticated AI systems to date still struggle with basic social interactions. That’s why researchers at MIT, Nvidia, and ETH Zurich developed Watch-And-Help (WAH), a challenge in which embodied AI agents need to understand goals by watching a demonstration of a human performing a task and coordinating with the human to solve the task as quickly as possible.

The concept of embodied AI draws on embodied cognition, the theory that many features of psychology — human or otherwise — are shaped by aspects of the entire body of an organism. By applying this logic to AI, researchers hope to improve the performance of AI systems like chatbots, robots, autonomous vehicles, and even smart speakers that interact with their environments, people, and other AI. A truly embodied robot could check to see whether a door is locked, for instance, or retrieve a smartphone that’s ringing in an upstairs bedroom.

In the first phase of WAH, which the researchers call the Watch stage, an AI agent observes a humanlike agent perform a task and infers a goal from their actions. In the second stage — the Help stage — the AI agent assists the humanlike agent in achieving the same goal in a completely different environment. The researchers assert that this two-stage framework poses unique challenges for human-AI collaboration because the AI agent has to reason about the humanlike agent’s intention and generalize its knowledge about the goal.

To enable the kinds of interactions involved in WAH, the researchers had to extend the open source platform VirtualHome and build a multi-agent environment dubbed VirtualHome-Social. VirtualHome-Social simulates home settings so agents can interact with different objects and agents, for example opening a container or grabbing a utensil from a drawer. VirtualHome-Social also provides built-in agents that emulate human behaviors and an interface for human players. This enables testing with real humans and human activities displayed in semi-realistic environments.

The humanlike agent represents a built-in agent in VirtualHome-Social. It plans its actions based on a goal and its observation of the environment. During the Help stage, the AI agent receives observations from the system at each step and sends an action command back to control a virtual avatar. Meanwhile, the humanlike agent — which can also be controlled by a human — updates its plan based on its latest observation to reflect any state change caused by the AI agent.

The researchers designed an evaluation protocol and provided benchmarks for WAH, including a goal model for the Watch stage and multiple planning and machine learning baselines for the Help stage. The team says results indicate that to achieve success in WAH, AI agents must acquire strong social perception and generalizable helping strategies — as hypothesized.

“Our ultimate goal is to build AI agents that can work with real humans. Our platform opens up exciting directions of future work, such as online goal inference and direct communication between agents,” the researchers wrote. “We hope that the proposed challenge and virtual environment can promote future research on building more sophisticated machine social intelligence.”

The audio problem:

Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here

Let’s block ads! (Why?)

Big Data – VentureBeat

AI researchers urge tech to go beyond scale to address systemic social issues

October 24, 2020 Big Data

The definition of success for startups and Big Tech companies alike has long been summed up by three words: hockey stick growth. Speedy gains in terms of both users and revenue is the dream for any company looking to scale. But according to a paper recently published by Google senior research scientist Alex Hanna and independent researcher Tina Park, a growing number of AI researchers say companies interested in purpose beyond profit need to consider approaches beyond rapid growth.

The paper argues that scale thinking is not just a way to grow a business, but a method that impacts all parts of that business, actively inhibits participation in tech and society, and “forces particular types of participation to operate as extractive or exploitative labor.”

“Whether people are aware of it or not, scale thinking is all-encompassing. It is not just an attribute of one’s product, service, or company, but frames how one thinks about the world (what constitutes it and how it can be observed and measured), its problems (what is a problem worth solving versus not), and the possible technological fixes for those problems,” the paper reads.

The authors go on to say that companies rooted in scale thinking are unlikely to be as “effective at deep, systemic change as their purveyors imagine. Rather, solutions which resist scale thinking are necessary to undo the social structures which lie at the heart of social inequality.”

This kind of thinking runs counter to not only dogma at the heart of Big Tech companies like Facebook and Google, but also the way media and analysts typically assess the value of emerging startups.

Earlier this month, Congress released an antitrust investigation that found Big Tech companies rely on scale to maintain and strengthen monopolies across the digital economy. A Department of Justice (DOJ) lawsuit filed Tuesday against Google, the first against a major tech company in two decades, also points to the scale achieved through algorithms and collection of personal user data as a key factor in the government’s decision to sue the Alphabet company.

The opposition includes scale evangelists like Y Combinator cofounder Paul Graham, AWS CTO Werner Vogels, and former Google CEO Eric Schmidt, who is quoted in the DOJ lawsuit as saying “scale is the key” to Google’s strength in search.

Embedded in scale thinking, Hanna and Park argue, is the idea that scalability is morally good and solutions that cannot scale are morally impoverished. Authors say that’s part of why Big Tech companies place such a high value on artificial intelligence.

“Large tech firms spend much of their time hiring developers who can envision solutions which can be implemented algorithmically. Code and algorithms which scale poorly are seen as undesirable and inefficient. Many of the most groundbreaking infrastructural developments in Big Tech have been those which increase scalability, such as Google File System (and subsequently the MapReduce computing schema) and distributed and federated machine learning models,” the paper reads.

Hanna and Park also characterize scale thinking as shortsighted because it requires companies to treat resources and people as interchangeable units and encourages the datafication of users in order to “find ways to rationalize the individual into legible data points.” This approach can lead to systems that are not made to serve everyone equally and that negatively impact the lives of those who fall outside their scaled solutions.

The paper also notes that scale thinking is an inefficient way to increase hiring or retention of employees from diverse backgrounds. Since the deaths of Black Americans like Breonna Taylor and George Floyd led to calls for racial justice earlier this year, a number of major tech companies have recommitted to diversity goals, but for years now progress has been virtually undetectable. Examples in the paper include a tendency to focus on bias workshops or other inclusion metrics rather than the experiences of marginalized people within the company.

Rather than making scale a company’s North Star, the authors suggest approaches like “mutual aid,” in which businesses adopt an interdependent model and take responsibility for meeting the direct material needs of individuals. The idea of mutual aid arose in part from the kinds of support systems that sprung up in the wake of the COVID-19 pandemic.

“While scale thinking emphasizes abstraction and modularity, mutual aid networks encourage concretization and connection,” the paper reads. “While mutual aid is not the only framework through which we can consider a move away from scale thinking-based collaborative work arrangements, we find it to be a fruitful one to theorize and pursue.”

In addition to exploring mutual aid, the paper encourages developers to ask questions about any system they create, such as whether it legitimizes or expands social systems people are trying to dismantle, whether it encourages broad participation, and whether it centralizes power or distributes power among developers and users.

Recommendations are in line with a range of ethically centered technology models proposed by members of the AI community in recent months. Other approaches include the idea of anticolonial AI, which rejects algorithmic oppression and data colonization, queering machine learning, data feminism, and building AI based on the African philosophy of Ubuntu, which focuses on the interconnectedness of people and the natural world.

There’s also “Good Intentions, Bad Inventions,” a Data & Society primer published earlier this month that attempts to dispel common myths about the best ways to build technology and improve user well-being.

Titled “Against Scale: Provocations and Resistances to Scale Thinking,” the paper was highlighted this week at a Computer-Supported Cooperative Work and Social Computing (CSCW) conference workshop. Before writing critically about scale, Hanna and colleagues at Google published a paper in late 2019 that argues the algorithmic fairness community should look to critical race theory as a way to interrogate AI systems and their impact on human lives.

The audio problem:

Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here

Let’s block ads! (Why?)

Big Data – VentureBeat

AI Weekly: Constructive ways to take power back from Big Tech

October 23, 2020 Big Data

Facebook launched an independent oversight board and recommitted to privacy reforms this week, but after years of promises made and broken, nobody seems convinced that real change is afoot. The Federal Trade Commission (FTC) is expected to decide whether to sue Facebook soon, sources told the New York Times, following a $ 5 billion fine last year.

In other investigations, the Department of Justice filed suit against Google this week, accusing the Alphabet company of maintaining multiple monopolies through exclusive agreements, collection of personal data, and artificial intelligence. News also broke this week that Google’s AI will play a role in creating a virtual border wall.

What you see in each instance is a powerful company insistent that it can regulate itself as government regulators appear to reach the opposite conclusion.

If Big Tech’s machinations weren’t enough, this week there was also news of a Telegram bot that undresses women and girls; AI being used to add or change the emotion of people’s faces in photos; and Clearview AI, a company being investigated in multiple countries, allegedly planning to introduce features for police to more responsibly use its facial recognition services. Oh, right, and there’s a presidential election campaign happening.

It’s all enough to make people reach the conclusion that they’re helpless. But that’s an illusion, one that Prince Harry, Duchess Meghan Markle, Algorithms of Oppression author Dr. Safiya Noble, and Center for Humane Technology director Tristan Harris attempted to dissect earlier this week in a talk hosted by Time. Dr. Noble began by acknowledging that AI systems in social media can pick up, amplify, and deepen existing systems of inequality like racism or sexism.

“Those things don’t necessarily start in Silicon Valley, but I think there’s really little regard for that when companies are looking at maximizing the bottom line through engagement at all costs, it actually has a disproportionate harm and cost to vulnerable people. These are things we’ve been studying for more than 20 years, and I think they’re really important to bring out this kind of profit imperative that really thrives off of harm,” Noble said.

As Markle pointed out during the conversation, the majority of extremists in Facebook groups got there because Facebook’s recommendation algorithm suggested they join those groups.

To act, Noble said pay attention to public policy and regulation. Both are crucial to conversations about how businesses operate.

“I think one of the most important things people can do is to vote for policies and people that are aware of what’s happening and who are able to truly intervene because we’re born into the systems that were born into,” she said. “If you ask my parents what it was like being born before the Civil Rights Act was passed, they had a qualitatively different life experience than I have. So I think part of what we have to do is understand the way that policy truly shapes the environment.”

When it comes to misinformation, Noble said people would be wise to advocate in favor of sufficient funding for what she called “counterweights” like schools, libraries, universities, and public media, which she said have been negatively impacted by Big Tech companies.

“When you have a sector like the tech sector that is so extractive — it doesn’t pay taxes, it offshores its profits, it defunds the democratic educational counterweights — those are the places where we really need to intervene. That’s where we make systemic long-term change, is to reintroduce funding and resources back into those spaces,” she said.

Forms of accountability make up one of five values found in many AI ethics principles. During the talk, Tristan Harris emphasized the need for systemic accountability and transparency in Big Tech companies so the public can better understand the scope of problems. For example, Facebook could form a board for the public to report harms; then Facebook can produce quarterly reports on progress toward removing those harms.

For Google, one way to increase transparency could be to release more information about AI ethics principle review requests made by Google employees. A Google spokesperson told VentureBeat that Google does not share this information publicly, beyond some examples. Getting that data on a quarterly basis might reveal more about the politics of Googlers than anything else, but I’d sure like to know if Google employees have reservations about the company increasing surveillance along the U.S.-Mexico border or which controversial projects attract the most objections at one of the most powerful AI companies on Earth.

Since Harris and others released The Social Dilemma on Netflix about a month ago, a number of people criticized the documentary for failing to include the voices of women, particularly Black women like Dr. Noble, who have spent years assessing issues undergirding The Social Dilemma, such as how algorithms can automate harm. That being said, it was a pleasure to see Harris and Noble speak together about how Big Tech can build more equitable algorithms and a more inclusive digital world.

For a breakdown of what The Social Dilemma misses, you can read this interview with Meredith Whittaker, which took place this week at a virtual conference. But she also contributes to the heartening conversation about solutions. One helpful piece of advice from Whittaker: Dismiss the idea that the algorithms are superhuman or superior technology. Technology isn’t infallible, and Big Tech isn’t magical. Rather, the grip large tech companies have on people’s lives is a reflection of the material power of large corporations.

“I think that ignores the fact that a lot of this isn’t actually the product of innovation. It’s the product of a significant concentration of power and resources. It’s not progress. It’s the fact that we all are now, more or less, conscripted to carry phones as part of interacting in our daily work lives, our social lives, and being part of the world around us,” Whittaker said. “I think this ultimately perpetuates a myth that these companies themselves tell, that this technology is superhuman, that it’s capable of things like hacking into our lizard brains and completely taking over our subjectivities. I think it also paints a picture that this technology is somehow impossible to resist, that we can’t push back against it, that we can’t organize against it.”

Whittaker, a former Google employee who helped organize a walkout at Google offices worldwide in 2018, also finds workers organizing within companies to be an effective solution. She encouraged employees to recognize methods that have proven effective in recent years, like whistleblowing to inform the public and regulators. Volunteerism and voting, she said, may not be enough.

“We now have tools in our toolbox across tech, like the walkout, a number of Facebook workers who have whistleblown and written their stories as they leave, that are becoming common sense,” she said.

In addition to understanding how power shapes perceptions of AI, Whittaker encourages people to try to better understand how AI influences our lives today. Amid so many other things this week, it might have been easy to miss, but the group AIandYou.org, which wants to help people understand how AI impacts their daily lives, dropped its first introductory video with Spelman College computer science professor Dr. Brandeis Marshall and actress Eva Longoria.

The COVID-19 pandemic, a historic economic recession, calls for racial justice, and the consequences of climate change have made this year challenging, but one positive outcome is that these events have led a lot of people to question their priorities and how each of us can make a difference.

The idea that tech companies can regulate themselves appears to some degree to have dissolved. Institutions are taking steps now to reduce Big Tech’s power, but even with Congress, the FTC, and the Department of Justice — the three main levers of antitrust — now acting to try to rein in the power of Big Tech companies, I don’t know a lot of people who are confident the government will be able to do so. Tech policy advocates and experts, for example, openly question whether factions Congress can muster the political will to bring lasting, effective change.

Whatever happens in the election or with antitrust enforcement, you don’t have to feel helpless. If you want change, people at the heart of the matter believe it will require, among other things, imagination, engagement with tech policy, and a better understanding of how algorithms impact our lives in order to wrangle powered interests and build a better world for ourselves and future generations.

As Whittaker, Noble, and the leader of the antitrust investigation in Congress have said, the power possessed by Big Tech can seem insurmountable, but if people get engaged, there are real reasons to hope for change.

For AI coverage, send news tips to Khari Johnson and Kyle Wiggers and AI editor Seth Colaner — and be sure to subscribe to the AI Weekly newsletter and bookmark our AI Channel.

Thanks for reading,

Khari Johnson

Senior AI Staff Writer

The audio problem:

Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here

Let’s block ads! (Why?)

Big Data – VentureBeat

Intel revenues drop 4% to $18.3 billion for Q3 2020 as competition heats up

October 23, 2020 Big Data

Intel reported earnings that matched Wall Street’s expectations during what was another tumultuous quarter for the world economy.

For the third quarter ended September 30, the big PC chipmaker posted non-GAAP earnings per share (EPS) of $ 1.11 a share on revenue of $ 18.3 billion, compared with non-GAAP EPS of $ 1.42 a share on revenue of $ 19.2 billion a year earlier.

CEO Bob Swan described the results as “solid” and said they exceeded Intel’s own expectations “despite pandemic-related impacts in significant portions of the business.”

Intel’s shares are down 8% this year, and the company’s market value at $ 229.2 billion remains lower than its biggest U.S. rival, Nvidia, which is valued at $ 330.1 billion. Intel’s stock price declined 10% to $ 48.30 a share in after-hours trading, as a number of the company’s business lines were weak. Intel faces tough competition from Nvidia in AI and graphics chips, while Advanced Micro Devices is much more competitive in central processing units (CPUs).

Analysts expected Intel to post adjusted earnings of $ 1.11 a share, down from $ 1.42 a share reported a year ago. Revenue was expected to be $ 18.24 billion, down from $ 19.19 billion a year ago.

“Nine months into 2020, we’re forecasting growth and another record year, even as we manage through massive demand shifts and economic uncertainty,” Swan said in a statement. “We remain confident in our strategy and the long-term value we’ll create as we deliver leadership products and aim to win share in a diversified market fueled by data and the rise of AI, 5G networks and edge computing.”

Intel revenues drop 4% to $18.3 billion for Q3 2020 as competition heats up

Above: Intel has some new Core processors coming in Q1 2021.

Image Credit: Intel

In July, Intel made the embarrassing disclosure that its new generation of 7-nanometer manufacturing has been delayed. This prompted Intel to say it was considering outsourcing some manufacturing to a contract chip manufacturer.

One of Intel’s advantages has been its internal manufacturing with plants in the United States, but the company has stumbled twice now on transitions to new generations, allowing rivals who use external manufacturer TSMC and others to make gains on Intel. Intel will likely offer an update on the manufacturing in its earnings call.

And Intel announced this week it would sell most of its flash memory business to SK Hynix for $ 9 billion. Intel’s memory business has been a mixed one with losses over a number of years.

“This is the first quarter I have seen COVID-19 negatively impact the company,” said Patrick Moorhead, an analyst at Moor Insights & Strategy. “Also, I believe it sold more 10nm parts with a faster ramp which are more costly to manufacture initially than 14nm parts. COVID-19 appears to have impacted the mix as PC demand shifted to lower margin education [PCs] and enterprise/datacenter demand looks to have dried up, replaced or weighed down by lower-margin cloud business. I do think [the CPU codenamed] Tiger Lake is strong for thin and light notebooks and believe it has a very good assortment coming into the holidays.

He added, “Looking at the longer-term picture, I do believe CEO Bob Swan is focusing (AI, GPUS, networking) and disinvesting (NAND, modems) in the right things in growing markets. I don’t think Intel needs to apologize for anything at this point. Its stock is trading at 10 times earnings and looks cheap.”

Swan said the company is raising its full-year revenue and earnings expectations from its July guidance. It now expects 5% revenue growth in 2020, with the full-year number coming in at $ 75.3 billion, while non-GAAP EPS will be $ 4.90 a share.

To date, Intel has generated $ 25.5 billion in cash from operations. Of Intel’s groups, Mobileye was up 2% at $ 234 million, and the PC client group was up 1% at $ 9.8 billion compared to a year ago. But Intel’s other four groups were down.

The datacenter group came in at $ 5.9 billion, down 7% from a year ago. The internet of things group was down 33% at $ 677 million. The NSG memory group was down 11% at $ 1.2 billion, and the programmable solutions group was down 19% at $ 411 million.

Intel said that continued strength in laptop sales was offset by COVID-19 headwinds affecting parts of the business. Cloud revenue grew 15% from a year ago with demand for supporting work-from-home environments, but the weaker economy affected enterprise and government sales, which was down 47% after two quarters of 30% growth.

The pandemic also hurt the internet of things and memory groups. Mobileye returned to growth as vehicle production recovered. Intel said more than 150 designs are in the works for new laptops with Intel processors.

For the fourth quarter, Intel expects non-GAAP earnings per share of $ 1.10 a share on revenue of $ 17.4 billion.

You can’t solo security

COVID-19 game security report: Learn the latest attack trends in gaming. Access here

Let’s block ads! (Why?)

Big Data – VentureBeat

Artificial intelligence and the antitrust case against Google

October 22, 2020 Big Data

Following the launch of investigations last year, the U.S. Department of Justice (DOJ) together with attorney generals from 11 U.S. states filed a lawsuit against Google on Tuesday alleging that the company maintains monopolies in online search and advertising, and violates laws prohibiting anticompetitive business practices.

It’s the first antitrust lawsuit federal prosecutors filed against a tech company since the Department of Justice brought charges against Microsoft in the 1990s.

“Back then, Google claimed Microsoft’s practices were anticompetitive, and yet, now, Google deploys the same playbook to sustain its own monopolies,” the complaint reads. “For the sake of American consumers, advertisers, and all companies now reliant on the internet economy, the time has come to stop Google’s anticompetitive conduct and restore competition.”

Attorneys general from no Democratic states joined the suit. State attorneys general — Democrats and Republicans alike — plan to continue on with their own investigations, signaling that more charges or backing from states might be on the way. Both the antitrust investigation completed by a congressional subcommittee earlier this month and the new DOJ lawsuit advocate breaking up tech companies as a potential solution.

The 64-page complaint characterizes Google as a “monopoly gatekeeper for the internet” and spells out the reasoning behind the lawsuit in detail, documenting the company’s beginning at Stanford University in the 1990s alongside deals made in the past decade with companies like Apple and Samsung to maintain Google’s dominance. Also key to Google’s power and plans for the future is access to personal data and artificial intelligence. In this story, we take a look at the myriad of ways in which artificial intelligence plays a role in the antitrust case against Google.

Search

The best place to begin when examining the role AI plays in Google’s antitrust case is online search, which is powered by algorithms and automated web crawlers that scour webpages for information. Personalized search results made possible by the collection of personal data started in 2009, and today Google can search for images, videos, and even songs that people hum. Google dominates the $ 40 billion online search industry, and that dominance acts like a self-reinforcing cycle: More data leads to more training data for algorithms, defense against competition, and more effective advertising.

“General search services, search advertising, and general search text advertising require complex algorithms that are constantly learning which organic results and ads best respond to user queries; the volume, variety, and velocity of data accelerates the automated learning of search and search advertising algorithms,” the complaint reads. “The additional data from scale allows improved automated learning for algorithms to deliver more relevant results, particularly on ‘fresh’ queries (queries seeking recent information), location-based queries (queries asking about something in the searcher’s vicinity), and ‘long-tail’ queries (queries used infrequently).”

Search is now primarily conducted on mobile devices like smartphones or tablets. To build monopolies in mobile search and create scale insurmountable to competitors, the complaint states, Google turned to exclusionary agreements with smartphone sellers like Apple and Samsung as well as revenue sharing with wireless carriers. The Apple-Google symbiosis is in fact so important that losing it is referred to as “code red” at Google, according to the DOJ filing. An unnamed senior Apple employee corresponding with their counterpart at Google said it’s Apple’s vision that the two companies operate “as if one company.” Today, Google accounts for four out of five web searches in the United States and 95% of mobile searches. Last year, Google estimated that nearly half of all search traffic originated on Apple devices, while 15-20% of Apple income came from Google.

Data at scale

Exclusive agreements that put Google apps on mobile devices effectively captured hundreds of millions of users. An antitrust report referenced these data advantages, stating that “Google’s anticompetitive conduct effectively eliminates rivals’ ability to build the scale necessary to compete.”

In addition to the DOJ report, the antitrust report Congress released earlier this month frequently cites the network effect achieved by Big Tech companies as a significant barrier to entry for smaller businesses or startups. The incumbents have access to large data sets that give them a big advantage, “especially when combined with machine learning and AI,” the report reads. “Companies with superior access to data can use that data to better target users or improve product quality, drawing more users and, in turn, generating more data — an advantageous feedback loop.”

Network effects often come up in the congressional report in reference to mobile operating systems, public cloud providers, and AI assistants like Alexa and Google Assistant, which improve their machine learning models through the collection of data like voice recordings.

One potential solution the congressional investigation suggested is better data portability to help small businesses compete with tech giants.

Google Assistant

One part of maintaining Google’s search monopoly, according to the congressional report, is control of emerging search access points. While Google searches began on desktop computers, mobile is king today, and fast emerging are devices like smartwatches, smart speakers, and IoT devices with AI assistants like Alexa, Google Assistant, and Siri. Virtual assistants are using AI to turn speech into text and predict a user’s intent, becoming a new battleground. An internal Google document declared voice “will become the future of search.”

The growth of searches via Amazon Echo devices is why a Morgan Stanley analyst previously suggested Google give everyone in the country a free speaker. In the end, he concluded, it would be cheaper for Google to give away hundreds of millions of speakers than to lose its edge to Amazon.

The scale afforded by Android and native Google apps also appears to be a key part of Google Assistant’s ability to understand or translate dozens of languages and collect voice data across the globe.

Search is primarily done on mobile devices today. That’s what drives the symbiotic relationship between Apple and Google, where Apple receives 20% of its total revenue from Google in exchange for making Google the de facto search engine on iOS phones, which still make up about 60% of the U.S. smartphone market.

The DOJ suit states that Google is concentrating on Google Nest IoT devices and smart speakers because internet searches will increasingly take place using voice orders. The company wants to control the next popular environment for search queries, the DOJ says, whether it be wearable devices like smartwatches or activity monitors from Fitbit, which Google announced plans to acquire roughly one year ago.

“Google recognizes that its ‘hardware products also have HUGE defensive value in virtual assistant space AND combatting query erosion in core Search business.’ Looking ahead to the future of search, Google sees that ‘Alexa and others may increasingly be a substitute for Search and browsers with additional sophistication and push into screen devices,’” the DOJ report reads. “Google has also harmed competition by raising rivals’ costs and foreclosing them from effective distribution channels, such as distribution through voice assistant providers, preventing them from meaningfully challenging Google’s monopoly in general search services.”

In other words, only Google Assistant can get microphone access for a smartphone to respond to a wake word like “Hey, Google,” a tactic the complaint says handicaps rivals.

AI like Google Assistant also features prominently in the antitrust report a Democrat-led antitrust subcommittee in Congress released, which refers to AI assistants as efforts to “lock consumers into information ecosystems.” The easiest way to spot this lock-in is when you consider the fact that Google prioritizes YouTube, Apple wants you to use Apple Music, and Amazon wants users to subscribe to Amazon Prime Music.

The congressional report also documents the recent history of Big Tech companies acquiring startups. It alleges that in order to avoid competition from up-and-coming rivals, companies like Google have bought up startups in emerging fields like artificial intelligence and augmented reality.

Finishing thoughts

If you expect a quick ruling by the DC Circuit Court in the antitrust lawsuit against Google, you’ll be disappointed — that doesn’t seem at all likely. Taking the 1970s case against IBM and the Microsoft suit in the 1990s as a guide, antitrust cases tend to take years. In fact, it’s not outside the realm of possibility that this case could still be happening the next time voters pick a president in 2024.

What does seem clear from language used in both US v Google and the congressional antitrust report is that both Democrats and Republicans are willing to consider separating company divisions in order to maintain competitive markets and a healthy digital economy. What’s also clear is that both the Justice Department and antitrust lawmakers in Congress see action as necessary based in part on how Google treats personal data and artificial intelligence.

The audio problem:

Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here

Let’s block ads! (Why?)

Big Data – VentureBeat

Data privacy startup Mine raises $9.5 million and expands to the U.S.

October 21, 2020 Big Data

Mine, which is developing a platform that helps discover and manage personal data on the web, today closed a $ 9.5 million funding round led by Gradient Ventures, Google’s AI-focused venture firm. The cash infusion comes as Mine prepares to launch in the U.S. following debuts in Israel and Europe earlier in the year. According to a spokesperson, 100,000 people have used Mine’s app to send over 1.3 million data-reclaim requests to more than 150,000 companies to date.

A majority of people in the U.S. (60%) believe their online and offline activities are being tracked and monitored by companies and the government with some regularity, according to a Pew Research Center survey. It’s a sentiment reflected in reports like PricewaterhouseCoopers’ Consumer Intelligence Series, which found that only 10% of consumers felt they had complete control over their personal information.

Mine’s technology discovers companies that are holding data by analyzing email subject lines, as opposed to the actual content of emails (e.g., “Welcome to Facebook” or “Your Amazon Receipt”). Once connected with an inbox, Mine maps all companies a user has interacted with via email, generating visualizations that show the type of data collected and the corresponding risk level. A dashboard provides a detailed overview of this data, along with shortcuts to remove the data from services no longer in use.

Mine’s algorithms identify companies most likely to have collected personal data based on the sender and the number of times they appear in an inbox. (According to Mine, almost 90% of traces can be found through email subject lines.) The platform independently locates and scans these companies’ privacy policies and data structures to determine what kinds of information they’re collecting from users.

An initial scan takes about 30 seconds, after which the system continues in the background. Mine sends a notification and an email once the audit is completed, and the app continuously provides updates about new traces of companies in a footprint. Mine claims users typically have a 40% cleaner footprint by the end of a session. The company also claims 15,000 Mine users have narrowly avoided data breaches since 2018 by deleting their data right before the breach occurred.

Mine, which has offices in Tel Aviv and New York, is free to use in 10 languages for Gmail, Yahoo, and Microsoft Outlook accounts, though the company plans to eventually charge a subscription fee. For now, it will focus on products that enable businesses to automatically process “right-to-be-forgotten” requests, helping people demand the deletion of certain data pertaining to them. In May 2018, the European Union’s General Data Protection Regulation (GDPR) codified right-to-be-forgotten protections into law. This was followed in January 2020 by the California Consumer Privacy Act (CCPA), which similarly requires that people be able to access and delete certain data.

“Having been part of the tech world from different angles, I noticed how people, myself included, started to lose trust in how companies are collecting and treating personal data. My cofounders and I created Mine to do what we know best — use technology as a solution to shape a new future of equal choice and control to the internet,” CEO Gal Ringel told VenureBeat via email. “Our smart non-intrusive AI makes it easy for people to take back ownership of their online data and minimize digital risks without changing their online behavior.”

Mine’s Gradient Ventures-led round included participation from E.ventures, MassMutual Ventures, and existing investors Battery Ventures and Saban Ventures. It brings the company’s total raised to over $ 12.5 million.

Let’s block ads! (Why?)

Big Data – VentureBeat

Salesforce’s Simulation Cards spell out uses, risks, and bias to make AI models more transparent

October 20, 2020 Big Data

Salesforce recently open-sourced Foundation (formerly AI Economist), an AI research simulation for exploring tax policies. To accompany its release, the company this week published what it’s calling a Simulation Card, a file to document the use, risks, and sources of bias in published versions of the simulation.

Simulation Cards join ongoing efforts to bring transparency to historically black-box systems. Over the past year, Google launched Model Cards, which sprang from a Google AI whitepaper published in October 2018. Model Cards specify model architectures and provide insight into factors that help ensure optimal performance for given use cases. The idea of Model Cards emerged following Microsoft’s work on “datasheets for data sets,” or datasheets intended to foster trust and accountability through documenting datasets’ creation, composition, intended uses, maintenance, and other properties. Two years ago, IBM proposed its own form of model documentation in voluntary factsheets called “Supplier’s Declaration of Conformity” (DoC) to be completed and published by companies developing and providing AI.

The objective of Simulation Cards is similar to that of Model Cards and Data Sheets. However, Simulation Cards reflect the fact that simulations differ from trained models and datasets because they’re designed to create scenarios of interest, according to Salesforce. These scenarios can contain bias, which might be purposefully built-in or an unexpected side effect of the design choices made during creation. Because simulations create many datasets of various shapes and sizes, the potential for misuse is greater than that of a single fixed dataset that might contain bias.

Salesforce’s Simulation Cards spell out uses, risks, and bias to make AI models more transparent

Above: The Simulation Card for Foundation.

Image Credit: Salesforce

The Simulation Card for Foundation is divided into several sections: Simulation Details, Basic Information, Intended Use, Factors, Metrics, Quantitative Analyses, Ethical Considerations, and Caveats and Recommendations. Simulation Details provides the date of the simulation and the name of the publishing organization, as well as any keywords, licenses, contact information, and relevant version numbers. The Basic Information and Intended Use sections cover top-level info about the simulation and the applications for it that the coauthors had in mind. Factors canvases the modeling assumptions the simulations make, while Metrics and Quantitative Analyses outline the metrics used to measure the results. Finally, Ethical Considerations and Caveats and Recommendations provide guidelines for (or warnings against) applying the outputs to real-world systems.

It remains to be seen what sort of third-party adoption Simulation Cards might gain, if any, but Salesforce itself says it’s committed to releasing cards alongside future simulations. “We encourage researchers and developers to publish similar Simulation Cards for software releases, to broadly promote transparency and the ethical use of simulation frameworks. AI simulations offer researchers the power to generate data and evaluate outcomes of virtual economies that capture a part of the real world,” Salesforce wrote in a blog post. “An unethical simulation poses an order-of-magnitude larger ethical risk. As a result, our commitment to transparency is all that much more critical.”

The audio problem:

Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here

Let’s block ads! (Why?)

Big Data – VentureBeat

DeepMind open-sources the FermiNet, a neural network that simulates electron behaviors

October 20, 2020 Big Data

In September, Alphabet’s DeepMind published a paper in the journal Physical Review Research detailing Fermionic Neural Network (FermiNet), a new neural network architecture that’s well-suited to modeling the quantum state of large collections of electrons. The FermiNet, which DeepMind claims is one of the first demonstrations of AI for computing atomic energy, is now available in open source on GitHub — and ostensibly remains one of the most accurate methods to date.

In quantum systems, particles like electrons don’t have exact locations. Their positions are instead described by a probability cloud. Representing the state of a quantum system is challenging, because probabilities have to be assigned to possible configurations of electron positions. These are encoded in the wavefunction, which assigns a positive or negative number to every configuration of electrons; the wavefunction squared gives the probability of finding the system in that configuration.

The space of possible configurations is enormous — represented as a grid with 100 points along each dimension, the number of electron configurations for the silicon atom would be larger than the number of atoms in the universe. Researchers at DeepMind believed that AI could help in this regard. They surmised that, given neural networks have historically fit high-dimensional functions in artificial intelligence problems, they could be used to represent quantum wavefunctions as well.

DeepMind open sources the FermiNet, a neural network that simulates electron behaviors

Above: Simulated electrons sampled from the FermiNet move around a bicyclobutane molecule.

By way of refresher, neural networks contain neurons (mathematical functions) arranged in layers that transmit signals from input data and slowly adjust the synaptic strength — i.e., weights — of each connection. That’s how they extract features and learn to make predictions.

Because electrons are a type of particle known as fermions, which include the building blocks of most matter (e.g., protons, neutrons, quarks, and neutrinos), their wavefunction has to be antisymmetric. (If you swap the position of two electrons, the wavefunction gets multiplied by -1, meaning that if two electrons are on top of each other, the wavefunction and the probability of that configuration will be zero.) This led the DeepMind researchers to develop a new type of neural network that was antisymmetric with respect to its inputs — the FermiNet — and that has a separate stream of information for each electron. In practice, the FermiNet averages together information from across streams and passes this information to each stream at the next layer. This way, the streams have the right symmetry properties to create an antisymmetric function.

Above: The FermiNet’s architecture.

The FermiNet picks a random selection of electron configurations, evaluates the energy locally at each arrangement of electrons, and adds up the contributions from each arrangement. Since the wavefunction squared gives the probability of observing an arrangement of particles in any location, the FermiNet can generate samples from the wavefunction directly. The inputs used to train the neural network are generated by the neural network itself, in effect.

“We think the FermiNet is the start of great things to come for the fusion of deep learning and computational quantum chemistry. Most of the systems we’ve looked at so far are well-studied and well-understood. But just as the first good results with deep learning in other fields led to a burst of follow-up work and rapid progress, we hope that the FermiNet will inspire lots of work on scaling up and many ideas for new, even better network architectures,” DeepMind wrote in a blog post. “We have … just scratched the surface of computational quantum physics, and look forward to applying the FermiNet to tough problems in material science and condensed matter physics as well. Mostly, we hope that by releasing the source code used in our experiments, we can inspire other researchers to build on our work and try out new applications we haven’t even dreamed of.”

The release of the FermiNet code comes after DeepMind demonstrated its work on an AI system that can predict the movement of glass molecules as they transition between liquid and solid states. (Both the techniques and trained models, which were also made available in open source, could be used to predict other qualities of interest in glass, DeepMind said.) Beyond glass, the researchers asserted the work yielded insights into general substance and biological transitions, and that it could lead to advances in industries like manufacturing and medicine.

The audio problem:

Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here

Let’s block ads! (Why?)

Big Data – VentureBeat

Category Archives: Big Data

AMD is buying Xilinx for $35 billion to compete with Intel in the datacenter

Researchers design virtual environment to spur development of helpful home robots

AI Weekly: Constructive ways to take power back from Big Tech

Intel revenues drop 4% to $18.3 billion for Q3 2020 as competition heats up

Artificial intelligence and the antitrust case against Google

Search

Data at scale

Google Assistant

Finishing thoughts

Data privacy startup Mine raises $9.5 million and expands to the U.S.

Salesforce’s Simulation Cards spell out uses, risks, and bias to make AI models more transparent

DeepMind open-sources the FermiNet, a neural network that simulates electron behaviors

Recent Posts

Categories

Archives