Category Archives: Ai

Decision algorithms are commonly infected with biases

Noel Sharkey, an expert in the field of artificial intelligence (AI), has urged the government to ban the use of all decision algorithms that impact on peoples' lives.

In an interview with theGuardian, Prof Sharkey said that such algorithms are commonly "infected with biases," and one should not expect them to make fair or trusted decisions.

"There are so many biases happening now, from job interviews to welfare to determining who should get bail and who should go to jail. It is quite clear that we really have to stop using decision algorithms," says the Sheffield University professor and robotics/AI pioneer, who is also a leading figure in a campaign against autonomous weapons.

"I am someone who has always been very light on regulation and always believed that it stifles innovation. But then I realised eventually that some innovations are well worth stifling, or at least holding back a bit. So I have come down on the side of strict regulation of all decision algorithms, which should stop immediately," he added.

According to Sharkey, all leading tech firms, such as Microsoft and Google, are aware of the algorithm bias problem and are also working on it, but none of them has so far come up with a solution.

Sharkey believes AI decision-making machines needs to be tested in the same way as any new pharmaceutical drug is before it is allowed on to the market.

By this, Sharkey means testing AI systems on at least hundreds of thousands, and preferably millions, of people, to eventually reach a point where the algorithm shows no major inbuilt bias.

To address this issue, US lawmakers introduced the Algorithmic Accountability Act in the lower House and Senate this April. The Act would require technology companies to ensure that their machine learning algorithms arefree of gender, race, and other biases before deployment.

If passed, the bill would require the Federal Trade Commission to create guidelines for assessing 'highly sensitive' automated systems. If companies find an algorithm implying the risk of privacy loss, they would take corrective actions to fix everything that is "inaccurate, unfair, biased or discriminatory" in the algorithm.

Last month Arvind Narayanan, an associate professor at Princeton, warned that most of the products or applications being sold today as AI are little more than "snake oil".

According to Narayanan, many companies have been using AI-based assessment systems to screen applicants. The majority of such systems claim to work not by analysing what the candidate said or wrote in their CV, but by speech pattern and body language.

"Common sense tells you this isn't possible, and AI experts would agree," Narayanan said.

According to Narayanan, the areas where AI use will remain fundamentally dubious include predicting criminal recidivism, forecasting job performance and predictive policing.

Arecent study by Oxford Internet Institute researchers suggested that current laws have largely failed to protect the public from biased algorithms that influence decision-making on everything from housing to employment.

The lead researcher of the study found many algorithms that were making conclusions about personal traits such as gender, ethnicity, religious beliefs and sexual orientation based on the individuals' browsing behaviour.

More here:

AI expert calls for end to use of 'racially biased' algorithms - http://www.computing.co.uk

Yona. (Courtesy of Auxuman)

Subscribe now for as little as $2 a month!

A mechanical way of thinking, an artificial kind of intelligence, has underlaid the making of popular music since the invention of the form. Thats why we can talk about pop as an invention with a given form. As early as the days of hit songs written on sheet music, before the advent of recordings and radio, Charles K. Harris, writer of the first million-selling song sheet, After the Ball, described the trade of tunesmithing in industrial terms: a matter of applying procedures derived from past successes, calibrating them for mass consumption. He wrote an instruction manual, How to Write a Popular Song, in 1906, with a checklist of essential rules. In his words:Ad Policy

Watch your competitors. Note their success or failure; analyze the cause and profit thereby.

Note public demand.

If you do not feel confident to write or compose a certain style of songadapt yourself to the others.

Over more than a century since Harriss era, the proposition that songs should be made according to rules based on precedent and shaped by market forces is taken as a given and, indeed, valued as a way of honoring tradition and pleasing the public in many spheres of music, from country and gospel to jazz and R&B. At the same time, talk of regimented production or subordination of the creative impulse is widely and freely employed as criticism. Music journalists and critics have few tools as piercing as the charge that a song is formulaic or mechanical, or that an artist is pandering or a sellout.

The growing use of artificial intelligence in music challenges us to consider not only songwriting but also singing and musicianship in ways that are essentially extensions of the machine-age thinking of Harris and, at the same time, startlingly new. A small but expanding group of tech innovators has been developing a range of musical applications of AI such as Boomy, which, so far, has allowed people to make more than 400,000 songs through a combination of machine learning and input from human users. Earlier this year, the start-up Endel released a series of albums of AI-generated ambient music on the major streaming services, through a partnership with Warner Music. The company is planning to release 20 albums by the end of this year, all in the vein of chill playlists, with candle-scent-like titles that signal their purpose of conjuring soothing atmospheres: Clear Night, Rainy Night, Cloudy Afternoon, Cloudy Night, and Foggy Morning. Wordless, nearly formless, and harmless, theyre perfectly functional background music, well-marketed to make an asset of the absence of anything warranting the listeners attention. A more recent project from another start-up, Auxuman, is an achievement on another level and may well be a watershed in pop music history.

Auxuman (brand shorthand for auxiliary human) is the brainchild of the British Iranian interdisciplinary artist Ash Koosha (aka Ashkan Kooshenaejad). He first established himself with a couple of pleasantly atmospheric synth-based pop albums, the first of which, Guud, included a single, I Feel That, whose official video starred a synthetic semihumanoid, created by the digital artist Hirad Sab. Before turning to AI, Koosha made some multisensory art for VR headsets. Now Koosha and a team of programmers have created a stable of digital music acts who (Ill use the personal pronoun for human beings in deference to the uman in Auxuman) will be releasing a full album of new material every month under the umbrella name Auxuman.

As Koosha has described the Auxuman process to the tech-fan site Digital Trends, the words, music, instruments, and singing voices on the tracks are generated by mining existing music on the Web and processing it to generate new songs. The synthetic artists who are the public face of the work are, in Kooshas words, a reflection of human life on the internet. Their music comes from stories we have told, ideas we have generated, and opinions we have shared. The principles are not far removed from Harriss rules for analyzing the music of other songwriters and factoring in public demand.

There are five digital performers in the Auxuman collective to date: Gemini, Hexe, Mony, Yona, and Zoya. Visually, in the avatar imagery that accompanies each song on YouTube, theyre a mix of racial and ethnic identities, in a few cases thoroughly and indefinitely mixed. Four (Gemini, Hexe, Mony, and Yona) are distinctly female in appearance, one (Zoya) male. The one front and center in group pictures, Yona, is, a pixieish white woman who looks like what she is: a computers idea of a pop star.

In September, the first music attributed to the five was released as a 10-track album, Auxuman #1. Taking in these recordings, at first I tried to shake the fact that they were generated by AI, and sought to listen with no preconceptions or expectations. Within a minute, I realized that was pointless and unfair to the work and its digital creators. With Auxuman #1, we are forced to confront a genuinely new type of music that works on its own terms.

All five voices, though somewhat distinct from one another, sound of a piece and appropriately artificialmetallic in tone and rigidly staccato in their diction and phrasing. The voices carry no warmth and have no flexibility. They sound inhuman, soulless, but fascinatingly so. To expect otherwise from them would be as wrongheaded as it was for early listeners of recordings to expect pioneers of the microphone such as Billie Holiday and Frank Sinatra to belt like Al Jolson, Sophie Tucker, and others who needed to bellow to reach the rafters of big concert halls. Yona, in particular, exudes impersonal detachment and superficiality. As well as any other artist I can think of, she gives voice to the sense of isolation that chills the air of the digital world.

The lyrics are an eerie jumble of phrases, mostly trite babbling, not unlike the words of a fair number of pop songs since Charles K. Harriss day. From time to time, though, the random juxtapositions come together in unnerving, accidental eloquence. In Oblivious, Yona sings:

Ive never felt warmIve never felt warmThrough the lensThrough your lensI feel warm

We know she cannot feel anything. And she communicates only coldness. Knowing this, as I listened, I found myself projecting onto her and started to feel bad for her. Through my lens, I gave her warmth.

The music on all 10 tracks is perfectly, unsettlingly synthetic. Every note is placed precisely on beat. The harmony in every chord is correct, technically. And yet, nothing sounds quite right. Theres something profoundly but fittingly disturbing about it all. Its utterly conventional and predictable in its formal structures and musical particulars, but wholly arbitrary, built of nothing but its own surface qualities. It sounds like what it is: code pretending to be life. I can think of nothing more relevant.

Read the original post:

Do Auxuman's AI Singers Herald the Shape of Music to Come? - The Nation

When it recently launched a new weather model called IBM GRAF, The Weather Company took a big supercomputing step forward.

IBM GRAF, with its ability to process weather data from a variety of sources worldwide, enables The Weather Company, an IBM Business, to deliver high-resolution, hourly-updated forecasts around the globeparticularly to regions that have never had them before.

But the full power of IBM GRAF to generate local forecasts for the entire world depends on technology that The Weather Company already had in place and is continually refining: artificial intelligence (AI) and cloud computing.

When a series of winter storms recently lashed much of the United States, millions of people used The Weather Channel mobile app and weather.com website to help plan their travel.

Sophisticated AI algorithms from The Weather Company turn troves of current and historic weather data into recommendations, for example, that can tell an electric utility company where to trim trees to prevent blackouts before the next storm hits. Or that can compute a two-week flu-risk forecast for a given locale.

And it is the mix of IBM Cloud and hybrid cloud networks that can deliver The Weather Companys forecasts anywhere in the worldto business customers computer systems and for free to millions of users via smartphones and web apps.

You have to know whats happening everywhere in the atmosphere, right now, said Cameron Clayton, general manager of IBM Watson Media and Weather. Supercomputing and AI have had a profound impact on our ability to map the atmosphere and predict the future. The cloud helps us share those forecasts any time, anywhere.

IBM GRAFshorthand for Global High-Resolution Atmospheric Forecasting systemruns on a purpose-built IBM POWER9 supercomputer known as Dyeus. It maps conditions at billions of points in the atmosphere to produce hyperlocal pictures of what the weather will look like up to 12 hours in advance. It also produces a fresh set of predictions every hour, rather than every 6 or 12, as with many weather models in other parts of the world.

But thats just the beginning. Those predictions feed into the multi-model forecasting engine that The Weather Company, part of IBM, has been refining for more than two decades. Using complex machine learning algorithms, that engine combines the IBM GRAF results with those of about 100 other weather forecasts from around the world, including the American modelthe Global Forecast System used by the National Weather Serviceand the European model, ECMWF.

All these other models have different opinions about what the forecast is, said Peter Neilley, an IBM distinguished engineer and director of weather forecasting sciences and technologies for The Weather Company. The AI or machine learning is the thing that binds them all together and figures out the right way to combine them to produce an optimized forecast.

To do that, the engine compares factors like temperature or precipitation from each model based on geography, time, weather type and recent forecast accuracy and assigns them relative weights and correction factors. The system then blends those weighted contributions to arrive at the final forecast. The calculations involve some 400 terabytes of data collected daily to provide forecasts for two billion locations around the world, according to Mr. Clayton, and can result in 25 billion forecasts every day.

Those end up, in one form or another, with consumers, whether delivered by a broadcaster as part of the local morning weather report; on an IBM website such as weather.com or wunderground.com or its familiar The Weather Channel app; or, for business customers, through a product tailored for specific markets and uses. But increasingly, in an approach Mr. Clayton calls cognitive computing, the company is using artificial intelligence and machine learning to combine meteorological observations with a variety of other types of data, and go beyond predicting the weather to enable specific responses.

This is often accomplished through IBMs Watsona suite of AI tools and apps named for the companys founder that famously beat two returning champions on Jeopardy! in 2011and in the IBM Cloud. The idea is to deliver specific insights to help people make better decisions, from whether to take along the umbrella for the day to whether to evacuate in advance of a storm.

For example, the company is working with Oncor, based in Texas and among the nations largest utilities, to help predict where vegetative growth is most likely to interfere with power lines which can cause blackouts and wildfiresallowing managers to better plan for preventative maintenance.

The Weather Company uses AI to analyze weather phenomena, location, season and time of day and other actors to provide 15-day local flu forecasts.

At Knight Transportation, a leading trucking company, a Watson-enabled service allows drivers to receive real-time audio alerts through their in-cab communication systems about how weather conditions will affect their planned routes. That means truckers, whose safety, delivery deadlines and cargo are regularly threatened by the effects of bad weather, can better avoid roads made dangerous by ice, fog or high winds, or seek shelter before hitting a storm.

The Weather Company is applying AI-powered predictive platforms to other industries as well, including agribusiness, insurance, aviation, retail and energy trading.

The Weather Company's flu tracker can help people take precautions during a local outbreak of illnesses.

On the consumer side, The Weather Company websites and apps can provide personalized information beyond, say, a prediction of a 75 percent chance of rain at 2 p.m. in a particular zip code. The Weather Channel app, for instance, is able to glean users interests from the alerts and services they opt into and determine what information they most want to see each day, like the best hours to go for a run or whether an allergy sufferer will have to contend with a high pollen count.

There is also a new tracker that analyses weather phenomena, location, season and time of day along with other factors to determine the likelihood that people are at risk for getting the flu. It then provides a 15-day flu forecast on the theory that people can take precautions like avoiding events or washing their hands more frequently during an outbreak.

It becomes a platform that you could use for pretty much any data type, any type of a health-type situationallergies, flu, arthritis flare-ups, said Chris Hill, chief information and technology officer for IBM Watson Media and Weather. Theres a lot of interesting dynamics of health and weather and weve built a generic engine that enables us to gain insights from different types of weather using data science.

See the original post here:

IBM GRAF Builds on The Weather Companys AI and Cloud Capabilities - Forbes

According to DataRobot report, nearly half of tech pros are concerned about AI bias, yet many organizations still use untrustworthy AI systems

As artificial intelligence (AI) continues its march into enterprises, many IT pros are beginning to express concern about potential AI bias in the systems they use.

A new report from DataRobot finds that nearly half (42%) of AI professionals in the US and UK are "very" to "extremely" concerned about AI bias.

The report, conducted last June of more than 350 US- and UK-based CIOs, CTOs, VPs, and IT managers involved in AI and machine learning (ML) purchasing decisions, also found that "compromised brand reputation" and "loss of customer trust" are the most concerning repercussions of AI bias. This prompted 93% of respondents to say they plan to invest more in AI bias prevention initiatives in the next 12 months.

SEE:The ethical challenges of AI: A leader's guide (free PDF)(TechRepublic)

Despite the fact that many organizations see AI as a game changer, many organizations are still using untrustworthy AI systems, said Ted Kwartler, vice president of trusted AI, at DataRobot.

He said the survey's finding that 42% of executives are very concerned about AI bias comes as no surprise "given the high-profile missteps organizations have had employing AI. Organizations have to ensure AI methods align with their organizational values, Kwartler said. "Among the many steps needed in an AI deployment, ensuring your training data doesn't have hidden bias helps keep organizations from being reactionary later in the workflow."

DataRobot's research found that while most organizations (71%) currently rely on AI to execute up to 19 business functions, 19% use AI to manage as many as 20 to 49 functions, and 0% leverage the technology to tackle more than 50 functions.

While managing AI-driven functions within an enterprise can be valuable, it can also present challenges, the DataRobot report said. "Not all AI is treated equal, and without the proper knowledge or resources, companies could select or deploy AI in ways that could be more detrimental than beneficial."

The survey found that more than a third (38%) of AI professionals still use black-box AI systems--meaning they have little to no visibility into how the data inputs into their AI solutions are being used. This lack of visibility could contribute to respondents' concerns about AI bias occurring within their organization, DataRobot said.

AI bias is occurring because "we are making decisions on incomplete data in familiar retrieval systems,'' said Sue Feldman, president of the cognitive computing and content analytics consultancy Synthexis. "Algorithms all make assumptions about the world and the priorities of the user. That means that unless you understand these assumptions, you will still be flying blind."

This is why it is important to use systems that include humans in the loop, instead of making decisions in a vacuum, added Feldman, who is also co-founder and managing director of the Cognitive Computing Consortium. They are "an improvement over completely automatic systems," she said.

SEE:Managing AI and ML in the enterprise 2019: Tech leaders expect more difficulty than previous IT projects(TechRepublic Premium)

Bias based on race, gender, age or location, and bias based on a specific structure of data, have been long-standing risks in training AI models, according to Gartner.

In addition, opaque algorithms such as deep learning can incorporate many implicit, highly variable interactions into their predictions that can be difficult to interpret, the firm said.

By 2023, 75% of large organizations will hire AI behavior forensic, privacy and customer trust specialists to reduce brand and reputation risk, Gartner predicts.

"New tools and skills are needed to help organizations identify these and other potential sources of bias, build more trust in using AI models, and reduce corporate brand and reputation risk," said Jim Hare, a research vice president at Gartner, in a statement.

"More and more data and analytics leaders and chief data officers (CDOs) are hiring ML forensic and ethics investigators," Hare added.

Organizations such as Facebook, Google, Bank of America, MassMutual, and NASA are hiring or have already appointed AI behavior forensic specialists to focus on uncovering undesired bias in AI models before they are deployed, Gartner said.

If AI is to reach its potential and increase human trust in the systems, steps must be taken to minimize bias, according to McKinsey. They include:

The DataRobot study found that to combat instances of AI bias, 83% of all AI professionals say they have established AI guidelines to ensure AI systems are properly maintained and yielding accurate, trusted outputs. In addition:

The latter stat surprised Kwartler. "I am concerned that only about half of the executives have algorithms in place to detect hidden bias in training data."

There were also cultural differences discovered between US and UK respondents to the DataRobot study.

While US respondents are most concerned with emergent biaswhich is bias resulting from a misalignment between the user and the system design UK respondents are more concerned with technical biasor bias arising from technical limitations, the study found.

To enhance AI bias prevention efforts, 59% of respondents say they plan to invest in more sophisticated white box systems, 54% state they will hire internal personnel to manage AI trust, and 48% say they intend to enlist third party vendors to oversee AI trust, according to the study.

The 48% figure should be higher, Kwartler believes. "Organizations need to own and internalize their AI strategy because that helps them ensure the AI models align with their values. For each business context and industry, models need to be evaluated before and after deployment to mitigate risks," he said.

Besides those AI bias prevention measures, 85% of all global respondents believe AI regulation would be helpful for defining what constitutes AI bias and how it should be prevented, according to the report.

Be in the know about smart cities, AI, Internet of Things, VR, autonomous driving, drones, robotics, and more of the coolest tech innovations. Delivered Wednesdays and Fridays

Image: iStockphoto/PhonlamaiPhoto

Go here to read the rest:

Why you need to pay more attention to combatting AI bias - TechRepublic

How businesses can use artificial intelligence (AI) to their advantage, perhaps even in a ... [+] transformative way, without turning the pursuit of AI advantage into a quixotic quest

Frankly, I was hoping an artificial intelligence (AI) algorithm would write this column for me, because who knows more about AI than the mysterious little gremlins that make machine learning possible? That, alas, didnt happen; so Im on my own.

Like most people in business, I dont need any convincing that artificial intelligence (for most companies in many areas of their operations) will become a game-changer.

Still, it remains a fluid, if not amorphous, concept in many respects. What, exactly, can we expect AI to do for us that were not already doingor how will it improve what were doing by doing it better, faster, cheaper, with greater insight or fewer errors?

As an important article (Winning With AI) in the MIT Sloan Management Review put it back in October, AI can be revolutionary, but executives must act strategically. [And] acting strategically means deciding what not to do. Thats not as easy as it sounds.

The problem I have with most discussions of artificial intelligence is that they assume the reader or listener already understands the promises and perils of AI. But based on my conversations with a lot of very intelligent people I dont think thats always the case.

As Amir Husain, founder and CEO of the Austin, Texas-based machine learning company, SparkCognition, explained to Business News Daily last spring, Artificial intelligence is kind of the second coming of software. Its a form of software that makes decisions on its own, thats able to act even in situations not foreseen by the programmers.

AI is so ubiquitous we hardly even think about it. As the Business News Daily article pointed out: Most of us interact with artificial intelligence in some form or another on a daily basis. From the mundane to the breathtaking, artificial intelligence is already disrupting virtually every business process in every industry.

Examples abound: online searches, spam filters, smart personal assistants, such as Alexa, Echo, Google Home and Siri, the programs that protect our information when we buy (or sell) something online, voice to text programs, smart-auto technologies, programs that automatically sound alarms or shut down operating systems when problems are identified, security alarm systems, even those annoying pop-up ads that follow us throughout the day. To one degree or another, theyre all based on or impacted by AI.

In other words, most of us are far more familiar with AIintimately sothan we give ourselves credit for.

The business question is (as the Sloan article correctly put it): How can executives exploit the opportunities, manage the risks, and minimize the difficulties associated with AI? Put another way, how can they use it to their advantage, perhaps even in a transformative way, without turning the pursuit of AI advantage into a quixotic questkeeping in mind that acting strategically involves deciding what not to do as well as pushing ahead and taking chances in some areas?

Some suggestions from the MIT Sloan Management Review article:

First: Dont treat AI initiatives as everyday technology gambits. Theyre more important than that. Run them from the C suite and closely coordinate them with other digital transformation efforts.

Second: Be sure to coordinate AI with the companys overall business strategy. One of the surest ways to come up shortas most AI initiatives do [from 40% to 70%, according to the Sloan article]is to focus AI narrowly on one set of priorities while the company is equally or more concerned with others. While AI can help companies reduce costs, for example, by identifying waste and inefficiencies, growing the business may be a higher priority.

The Hartford, Conn.-based insurance company, Aetna (now a subsidiary of CVS), for example, has been using AI to prevent fraud and uncover overpaymentstypical insurance company concerns. Its also been using AI to design products and increase customers and customer engagement. In one Medicare-related Aetna product, the article notes, designers used AI to customize benefits, leading to 180% growth in new member acquisition. More long term, Aetnas head of analytics, VP Ali Keshavarz, told the authors Aetnas goal is to use AI to become the first place customers go when they are thinking about their health.

Third: This may be obvious to the geeks among us, but perhaps less so to the more technology-challenged: Be sure to align the production of AI with the consumption of AI.

Fourth: Invest in AI talent, data and process change in addition to (and often more so than) AI technology. Recognize that every successful AI undertaking is the product of a great group of people. While some of this talent should be home grown, youll also have to hire from the outside: bring people in to develop and enhance your internal capabilities. Thats a fact of modern business life.

As with everything else in business, all companies are different. Their needs are different. Their available resources (financial, talent, patience) are different. And their goals and expectations should be different.

Its important to take the time to understand how to maximize the promise and minimize the pitfalls of AI. If you do, youre more likely to succeed.

See the rest here:

Maximize The Promise And Minimize The Perils Of Artificial Intelligence (AI) - Forbes

The principal tasks of artificial intelligence (AI) are training and inferencing. The former is a data-intensive process to prepare AI models for production applications. Training an AI model ensures that it can perform its designated inferencing tasksuch as recognizing faces or understanding human speechaccurately and in an automated fashion.

Inferencing is big business and is set to become the biggest driver of growth in AI. McKinsey has predicted that the opportunity for AI inferencing hardware in the data center will be twice that of AI training hardware by 2025 ($9 billion to 10 billion vs. $4 billion to $5 billion today). In edge device deployments, the market for inferencing will be three times as large as for training by that same year.

For the overall AI market, the market for deep-learning chipsets will increase from $1.6 billion in 2017 to $66.3 billion by 2025, according to Tractica forecasts.

I believe Nvidia NVDA, -3.46% will realize better-than-expected growth due to its early lead in AI inferencing hardware accelerator chips. That lead should last for at least the next two years, given industry growth and the companys current product mix and positioning.

In most server- and cloud-based applications of machine learning, deep learning and natural language processing, the graphics processing unit, or GPU, is the predominant chip architecture used for both training and inferencing. A GPU is a programmable processor designed to quickly render high-resolution images and video, originally used for gaming.

Nvidias biggest strength and arguably its largest competitive vulnerability lies in its core chipset technology. Its GPUs have been optimized primarily for high-volume, high-speed training of AI models, though they also are used for inferencing in most server-based machine learning applications. Today, that GPU technology is a significant competitive differentiator in the AI inferencing market.

Liftr Cloud Insights has estimated that the top four clouds in May 2019 deployed Nvidia GPUs in 97.4% of their infrastructure-as-a-service compute instance types with dedicated accelerators.

While GPUs have a stronghold on training and much of the server based inference, for edge-based inferencing, CPUs rule.

Whats the difference between GPUs and CPUs? In simple terms, a CPU is the brains of the computer and a GPU acts as a specialized microprocessor. A CPU can handle multiple tasks, and a GPU can handle a few tasks very quickly. CPUs currently dominate in adoption. In fact, McKinsey projects that CPUs will account for 50% of AI inferencing demand in 2025, with ASICs, which are custom chips designed for specific activities, at 40% and GPUs and other architectures picking up the rest.

The challenge: While Nvidias GPUs are extremely capable for handling AIs most resource-intensive inferencing tasks in the cloud and server platforms, GPUs are not as cost-effective for automating inferencing within mobile, IoT, and other edge computing uses.

Various non-GPU technologiesincluding CPUs, ASICs, FPGAs, and various neural network processing unitshave performance, cost, and power-efficiency advantages over GPUs in many edge-based inferencing scenarios, such as autonomous vehicles and robotics.

The opportunity: The company no doubt recognizes the much larger opportunity resides in inferencing chips and other components optimized for deployment in edge devices. But it has its work cut out to enhance or augment its current offerings with lower-cost, specialty AI chips to address that important part of the market.

Nvidia continues to enhance its GPU technology to close the performance gap vis--vis other chip architectures. One notable recent milestone was the recent release of AI industry benchmarks that show Nvidia technology setting new records in both training and inferencing performance. The companys forthcoming new AI-optimized Jetson Xavier NX hardware module will offer server-class performance, a small footprint, low cost, low power, high performance and flexible deployment for edge applications.

With an annual revenue run rate nearing $12 billion, Nvidia retains a formidable lead over other AI-accelerator chip manufacturers, especially AMD AMD, -1.07% and Intel INTC, -0.67%.

Intel, however, has upped its game in AI inference with the recent release of multiple specialty AI chips and the recent announcement that Ponte Vecchio, the companys first discrete GPU, should hit the market in 2021. There is also a range of cloud, analytics and development tool vendors who have flocked into the AI space over the past several years.

Nvidias early lead can be attributed to the companys focus, as well as the deep software integration that enables developers to rapidly develop and scale models on its hardware. This is why many of the hyperscalers (Alphabets GOOG, -1.15% GOOGL, -1.17% GoogleCloud, Microsofts MSFT, -1.21% Azure, Amazons AMZN, -1.07% AWS) also deliver AI inference capabilities on their infrastructure based upon Nvidia technology.

In edge-based inferencing, where AI executes directly on mobile, embedded, and devices, no one hardware/software vendor is expected to dominate, and Nvidia stands a very good chance of pacing the field. However, competition is intensifying from many directions. In edge-based AI inferencing hardware alone, Nvidia faces competition from dozens of vendors that either now provide or are developing AI inferencing hardware accelerators. Nvidias direct rivalswho are backing diverse AI inferencing chipset technologiesinclude hyperscale cloud providers AWS, Microsoft, Google, Alibaba BABA, -1.84% and IBM IBM, -1.15% ; consumer cloud providers Apple AAPL, -1.16%, Facebook FB, -0.96% and Baidu BIDU, -0.92% ; semiconductor manufacturers Intel, AMD, Arm, Samsung, Qualcomm QCOM, -1.32%, Xilinx XLNX, -2.71% and LG; and a staggering number of China-based startups and technology companies such as Huawei.

The significant opportunities tied to the growth of AI inferencing will drive innovation and competition to develop more powerful and affordable solutions to leverage AI. With the deep resources and capabilities of most of the aforementioned competitors, there is certainly a possibility of a breakthrough that could rapidly shift the power positions in AI inferencing. However, at the moment, Nvidia is the company to beat, and I believe this strong market position will continue for at least the next 24 months.

With Nvidia placing an increased focus on low-cost edge-based inferencing accelerators as well as high-performance hardware for all AI workloads, the company provides widely adopted algorithm libraries, APIs and ancillary software products designed for the full range of AI challenges. Any competitor would need to do all of this better than Nvidia. That would be a tall task, but certainly not insurmountable.

Daniel Newman is the principal analyst at Futurum Research. Follow him on Twitter @danielnewmanUV. Futurum Research, like all research and analyst firms, provides or has provided research, analysis, advising, and/or consulting to many high-tech companies in the tech and digital industries. Neither he nor his firm holds any equity positions with any companies cited.

Go here to read the rest:

Nvidia will dominate this crucial part of the AI market for at least the next two years - MarketWatch

Artificial intelligence is being adopted by companies in numerous industries and will displace jobs. It will also likely worsen gender and racial gaps. Nearly 11 percent of jobs held by women may be eliminated because of AI, warns the International Monetary Fund. McKinsey, meanwhile, estimates that 20 percent of women employed today could experience AI-driven job erosion in the coming decade. Preparing women to succeed in the age of AI must be a global priority as we work towards one key United Nations Sustainable Development Goal: to achieve gender equality, and empower women and girls.

Despite the gloomy statistics, theres reason to be hopefulif business, governments and individuals strategically work together. It helps to know that nearly 20 percent more women could be employed by 2030 than todayif theyre able to maintain their current representation level in each sector of the economy. The good news is that women tend to be concentrated in fields that are growing, like health care.

But then there are industries like financial services, where women make up nearly half of the U.S. sectors workforce. Women hold only 25 percent of the financial services senior management jobs that experts say are less vulnerable to AIs impact. Frontline workers are particularly vulnerable to AI disruption, and in the U.S. 85 percent of bank tellers are women.

Some companies are proactively developing solutions. Synchrony, a Connecticut-based financial services firm, has created a deliberate AI workforce strategy and invested aggressively in training programs for its 16,500 employees. In todays technology revolution, leaders must upskill their workforce and findor trainthe right talent,wroteCEO Margaret Keane. Nearly half of Synchronys employees are in frontline roles such as customer service. In one program, the company pays up to $20,000 for an employee to earn a higher education degree. This potentially helps employees transition from, say, customer service into a role thats more technicalpossibly in a field thats growing.

Between 40 and 60 million women around the world may need to change occupations to remain employed.

More companies should follow Synchronys lead. Globally, business and government must boost the number of women in high school, college and graduate schoolespecially those who focus on science, technology, engineering and math-related fields, all of which are growing. McKinsey estimates that between 40 and 60 million women around the world may need to change occupations to remain employed, often stepping into higher-skilled roles.

The unfortunate truth is that women tend to have smaller professional networks than men. So we must develop mechanisms for women to expand their base of contactswhich could ease their transition between fields. We must also train more women who can joinand leadthe teams that actually write the algorithms for AI. A critical mass of women in these roles should reduce the likelihood such systems will have built-in biasagainst women, a known problem. Theres no shortage of need: Only 22 percent of AI professionals are women, according to an analysis by the World Economic Forum and LinkedIn. By a variety of measures, the number of AI-related jobs are growing.

Shannon Eusey The Nations Largest Female-Owned RIA Started as a Business School Class Project

This isnt an issue just for womenall of business and society must participate. Thats why companies are moving quickly to create an infrastructure for job transformation. Bank of America is another financial services company taking concrete steps to help women adapt to an increasingly AI-driven workforce. Says Cathy Bessant, the banks chief operations and technology officer: Its an area where we still have a lot of work to do.

This article originally appeared in Techonomys Winter 2020 magazine.

The post How Women Can Win in the Age of AI appeared first on Worth.

See the original post here:

How Women Can Win in the Age of AI - Yahoo News

By Bob Sperber

When deployed in tandem, artificial intelligence (AI) and the Internet of Things (IoT) can bring powerful new capabilities and competitive advantagesa net effect that is greater than the sum of its constituent parts.

This is the central finding of a new study conducted by SAS, Deloitte and Intel with research from IDC based on survey of 450 global business leaders. Entitled AIoT: How Leaders Are Breaking Away, the survey report indicates that this combination of technologies, dubbed the Artificial Intelligence of Things, represents a key competitive advantage that already has passed from pilot-scale tests to early rollouts.

As companies grow into a fuller implementation of IoT, they begin to realize that the tremendous volumes of data generated are difficult to tame. In this context, the combination of AI with IoT is a natural fit for gaining insights that can help advance not just operational goals but business strategy.

Consider some of the findings the research brought to light:

99% of respondents said,in aggregate, the benefits of using AI together with their IoT solutions met or exceeded expectations. 90% of respondents who reported heavy use of AI for IoT operations said it exceeded their expectations for value. 35% of senior leaders cited increased revenue as the single most important area of improvement they expected to achieve from their IoT efforts.

Overall, projects that combine IoT with AI are having a greater-than-expected impact in operations, the enterprise and ultimately, the bottom line.

The expectations game

It came as a bit of a surprise to IDCs Maureen Fleming, program vice president for intelligent process automation, that leaders value the addition of AI to IoT projects as highly as the do. She confessed to Smart Industry that in her travels and client encounters, shes been barraged with negative feedback to the point where it seems like everywhere I go people are talking about the high failure rate of digital transformation efforts.

Naturally, she expected lower engagement among respondents. But what we found true is the exact opposite.

One possible explanation for the surprising, healthy attitude toward this thing called AIoT is that its the IT and operations leaders who fret over the details more than the CEOs office. According to the research, 56% of senior leaders believe their AIoT projects significantly exceeded expectations, a margin 18% greater than operations-related teams and 31% greater than data scientists and IT leaders. Interestingly, operational leaders were the greatest proponents of IoT alone (Figure 1).

Figure 1

In my experience, senior executives tend to be a lot more optimistic than those at other levels in the organizationits kind of a requirement for the job, says Shak Parran, partner at Deloitte Canada and analytics leader for its Omnia AI practice. Below the top floor, he says, the practical reality of putting these capabilities to work can make data scientists a little more pessimistic. They know that their data has to be cleaned up, they have to teach machines to do the right things, their processes have to be optimized, and so on. They see the obstacles, because thats what theyre responsible for navigating.

The good news is that this attitudinal gap may close over time, if an observation by Melvin Greer, Intels chief data scientist, comes to fruition. Over the past 24-36 months, were seen ample evidence of chief data officers moving into the CEO suite.

A competitive AI-vantage

As implementation teams have matured, so has the likelihood of success with digital transformation initiatives. For successful projects, focus shifts from connecting devices and collecting new and different data to the next step of the journey, analytics. From analytics to the use of AI is another step forward in the ability to filter, correlate and uncover complex relationships. The researchers confirm that industrial firms are indeed moving from proofs of concept and pilot tests to production systems with analytical approaches that incorporate AI.

The key to driving long-term, sustainable value with AIoT lies in building experience with large-scale rollouts, with higher levels of automation, throughout the organization. And the only way to reach that scale with AIoT is to increase the level of automation, according to Oliver Schabenberger, COO and CTO at SAS. So many CIOs I talk with say automation is a primary focus, to make IoT-related analytics insights consumable by business analysts and others, not just the data scientists.

In turn, the reason to scale-up and automate is to gain a competitive advantage. And, the report shows, companies that use AI and IoT together are more competitive than those using only IoT.

When asked about their success across six major initiativesfrom speeding-up operations to improving productivity and reducing coststhose respondents who used AI in conjunction with IoT said they were significantly more successful than counterparts who used only IoT. For instance, 53% of leaders reported significant value in using AIoT for speeding-up operations as compared to 32% using IoT without AI. Roughly similar numbers hold for initiatives to improve employee productivity, streamline operations and provide new digital services and innovations. In all cases, there was a double-digit gap between users of AIoT and IoT alone (Figure 2).

Figure 2

Of the six initiatives examined, AI was seen to be least important in the area of reducing costs/expenses.

This is not unexpected, according to Jason Mann, vice president of IoT, SAS. Companies are primarily focused on three core business objectives: achieving higher levels of operational efficiency, improving top-line growth and enhancing customer engagement, he says. While cost-cutting is important, its typically not a strategic business driver.

Companies who refresh their data at least once a day were asked about the role AI plays in rapid, tactical planning. When AI isnt in play, IoT data is overwhelmingly applied to operational decisions (68%); only 12% use IoT for day-to-day planning-oriented decisions. But with the introduction of AI, the number of respondents using this data for day-to-day planning nearly triples, increasing to 31%.

IDCs Fleming, calling this the most interesting finding of the study, explained that improving the speed of sensor data refresh combined with AI expands an organizations ability to focus on immediate planning, while also quickly identifying and resolving operational problems. The combination produces greater agility and more efficiency.

More generally, those who use AI this way broaden their toolset to address issues of supply and demand, product quality, merchandising and more, says Intels Gadgil: Theyre focusing on issues like productivity, but theyre also looking for the next opportunities for transformation in their businesstheyre pushing their organizations to connect the dots and see how some of these new technologies can contribute.

Increased revenue is job one

Leading companies that use AI to leverage data beyond its own operations and into the supply chain are better able to drive value back to your customer, and build a portfolio of data-driven services, says Bill Roberts, senior director in SAS IoT division. Further, after 12-24 months of using AI + IoT, users reported decreased costs or expenses (85%), improved employee productivity (87%) and streamlined operations (86%).

Logic and intelligence are now going to be distributed across the architecture, right back into the service center, onto the device or truck or piece of equipment, Roberts says.

Among the benefits sought for their IoT efforts, increased revenue topped the list for senior leaders across geographies, industries, and companies of all sizes. And if the improved results reported by those who have begun to add AI to their IoT connectivity projects is any indication, the AIoT has a bright future indeed.

More:

Will AI liberate the IoT's potential? - Smart Industry

Artificial intelligence is disrupting industries across the board. In healthcare, AI technologies are outperforming humans in diagnosing disease--particularly when it comes to spotting malignant tumors. In marketing, artificial intelligence analyzes users behavioral patterns, enabling businesses to target customers with highly personalized content

AI could disrupt the design industry in a variety of ways.

The design industry is yet another area where AI is making huge strides. Just recently, an AI tool pulled from a database of dozens of patterns and colors to create 7 million unique packaging designs for Nutella.

Lets take a closer look at how artificial intelligence is shaping the future of design.

In addition to generating its own designsas it did for NutellaAI is the driving force behind modern web design. Artificial design intelligence (ADI) systems are democratizing website development, generating functional, attractive websites from the bottom up.

Wix and Bookmark both offer AI platforms that allow websites to intelligently design themselves; the customer is responsible for choosing the sites name and answering a few quick questions, but AI will do the rest. Designers and developers no longer need to build websites completely from scratch, and they can create attractive sites regardless of their level of experience with web building or design.

Quality assurance (QA) is an essential area of concern to companies. If the public thinks a company or service does not care about keeping its quality level high, theyll likely do business elsewhere. AI for quality assurance can help. Read here

In addition to designing the face of web pages, AI builds the image of a brand. Artificial intelligence tools like Tailor Brands can gather data from around the web to design customized logos in seconds, requiring from the user nothing but their businesss name and a quick description of their company or industry. In this sense, AI has made brand design more accessible for emerging entrepreneurs with small budgets and little to no design experience.

Designers working with AI can create designs faster and more cheaply. Artificial intelligence tools can analyze vast amounts of data within minutes and suggest designs accordingly. Airbnb is already taking advantage of this function, feeding wireframe sketches to AI machines which, in turn, can generate complete images. This capability can be used to create several different prototype designs that can then be A/B tested with users.

Domos VP Data and Curiosity Ben Schein discusses integrating AI into your company practices without comprimising agility, speed or control. Read here

AI saves designers time by automating mundane tasks, allowing designers to instead focus on higher-level work. Rather than designing banners and product labels in multiple languages, for instance, designers can outsource the work to AI. The designers, in turn, can approve or reject the suggested graphics, saving them enormous amounts of time.

Artificial intelligence not only designs two-dimensional graphics like web pages and branding materials, but it also builds three-dimensional graphics. AI is being used to build three-dimensional architectural models, facilitating the work of architects by enabling them to create a detailed, lifelike blueprint of their building plan. AI is also being used to build worlds for virtual reality, artificial reality, and mixed reality. Artificial intelligence tools not only design the graphics for these worlds, but they also react intelligently to user interaction and behavior.

Some people worry that, with the advanced design capabilities of artificial intelligence, AI will eventually replace human designers altogether. The reality, however, is that AI will play a more complex and nuanced role in design. Artificial intelligence tools will facilitate the work of human designers while remaining a tooland not a replacementfor human designers.

Think about it: standard machines dont replace workers altogether but, instead, automate basic tasks and help workers operate more efficiently. Artificial intelligence programs will operate the same way, automating mundane tasks and even making original suggestions while still requiring a human designer to oversee the process and make the most important decisions.

Jens Krueger, chief technology officer, Workday, explains why a culture of inclusion, innovation and upskilling is crucial in order to reap the benefits of automation. Read here

So while AI isnt going to replace designers, it is going to replace something else: traditional methods of doing business. AI is changing the way people lead, manage, and make decisions.

In the world of design, this means using market data to make informed hypotheses about which designs work best for business. AI platforms can analyze which colors, shapes, typefaces, and other visuals perform best across different industries. That way, rather than coming up with a design based on limited research and intuition, human designers can use artificial intelligence tools to make design choices based on real, hard data.

In the next few years, designers will increasingly harness the power of artificial intelligence to inform their design decisions. AI tools wont replace designers but, on the contrary, will facilitate designers work so they can focus on bigger picture tasks. Artificial intelligence is already strong across the design industry, generating graphics for websites, brands, virtual reality and more. As AI becomes stronger, it wont just show us what we already know; itll also present us with novel ideas and open new avenues for thought.

Written by Tailor Brands, an automated, AI-powered logo design and branding platform.

Read more:

AI and the future of design - Information Age

AI's Impact in 2020: 3 Trends to Watch

The popularity of AI and ML have wide-reaching effects on your enterprise. Here are three important trends driven by AI to look out for next year.

[Editor's note: Upside asked executives from around the country to tell us what top three trends they believe data professionals should pay attention to in 2020. Ryohei Fujimaki, Ph.D., founder and CEO of dotData, focused on AI and ML.]

The Rise of AutoML 2.0 Platforms

As the need for additional AI applications grows, businesses will need to invest in technologies that help them accelerate the data science process. However, implementing and optimizing machine learning models is only part of the data science challenge. In fact, the vast majority of the work that data scientists must perform is often associated with the tasks that preceded the selection and optimization of ML models such as feature engineering -- the heart of data science.

This means that organizations will need to look for new, more sophisticated automated machine learning platforms. These "AutoML 2.0" tools will need to provide end-to-end automation, from automatically creating and evaluating thousands of features (AI-based feature engineering) to the operationalization of ML and AI models -- and all the steps in between.

The Shift to Automation Will Intensify Focus on Privacy and Regulations

As AI and ML models become easier to create using advanced "AutoML 2.0" platforms, data scientists and citizen data scientists will begin to scale ML and AI model production in record numbers. This means organizations will need to pay special attention to data collection, maintenance, and privacy oversight to ensure that the creation of new, sophisticated models does not violate privacy laws or cause privacy concerns for consumers.

As a result, in 2020 we will see an emergence of new tools that will enable data scientists to have greater transparency without sacrificing accuracy. This shift to a more "white box" approach to data science will deliver more transparent and accurate models thereby empowering businesses to make data-centric decisions and accelerating their digital transformations.

More Citizen Data Scientists Doing Data Science

Big data will continue to be on the upsurge in 2020 with a growing demand for skilled data scientists and a continued shortage of data science talent -- creating ongoing challenges for businesses implementing AI and ML initiatives. Although AutoML platforms have alleviated some of the pressure on data science teams, they have not resulted in the productivity gains organizations are seeking from their AI and ML initiatives. As such, companies need better solutions to help them leverage their data for business insights.

In 2020, we will see a swift adoption of new, broader, "full-cycle" data science platforms that will significantly simplify tasks that formerly could only be completed by data scientists and boost the productivity of citizen data scientists -- business analysts and other data experts who have domain expertise but are not necessarily skilled data scientists. This continued democratization will lead to new use cases that are closer to the needs of business users and will enable faster time-to-market for AI applications in the enterprise.

About the Author

Dr. Ryohei Fujimaki is the founder and CEO of dotData. In his career, Dr. Fujimaki was the youngest research fellow ever in NEC Corporations 119-year history, the title was honored for only six individuals among over 1000 researchers. During his tenure at NEC, Ryohei was heavily involved in developing many cutting-edge data science solutions with NECs global business clients, and was instrumental in the successful delivery of several high-profile analytical solutions that are now widely used in industry. You can reach the author via email or LinkedIn.

Go here to read the rest:

AI's Impact in 2020: 3 Trends to Watch - TDWI