Daily Archives: July 23, 2021

In the nine years since AlexNet spawned the age of deep learning, artificial intelligence (AI) has made significant technological progress in medical imaging, with more than 80 deep-learning algorithms approved by the U.S. FDA since 2012 for clinical applications in image detection and measurement. A 2020 survey found that more than 82% of imaging providers believe AI will improve diagnostic imaging over the next 10 years and the market for AI in medical imaging is expected to grow 10-fold in the same period.

Despite this optimistic outlook, AI still falls short of widespread clinical adoption in radiology. A 2020 survey by the American College of Radiology (ACR) revealed that only about a third of radiologists use AI, mostly to enhance image detection and interpretation; of the two thirds who did not use AI, the majority said they saw no benefit to it. In fact, most radiologists would say that AI has not transformed image reading or improved their practices.

Why is there such a huge gap between AIs theoretical utility and its actual use in radiology? Why hasnt AI delivered on its promise in radiology? Why arent we there yet?

The reason isnt because companies havent tried to innovate. Its because they were trying to automate away the radiologists job and failed, burning plenty of investors and leaving them reluctant to fund other projects aimed at translating AIs theoretical utility into real-world use cases.

AI companies seem to have mistaken Charles Friedmans fundamental theorem of biomedical informatics: it isnt that a computer can accomplish more than a human; its that a human using a computer can accomplish more than a human alone. Creation of this human-machine symbiosis in radiology will require AI companies to understand:

Together, these features, delivered as a unified cloud-based solution, would simplify and optimize the radiology workflow while augmenting the radiologists intelligence.

History Lessons

Modern deep learning dawned in 2012, when AlexNet won the ImageNet challenge, leading to the resurgence of AI as we think of it today. With the problem of image classification sufficiently solved, AI companies decided to apply their algorithms to images that have the greatest impact on human health: radiographs. These post-AlexNet companies can be viewed as falling into three generations.

The first generation approached the field with the assumption that AI know-how was sufficient for commercial success, and so focused on building early teams with knowledge around algorithms. However, this group drastically underestimated the difficulty of acquiring and labeling large-enough medical imaging data sets to train these models. Without sufficient data, these first-generation companies either failed or had to pivot away from radiology.

The second generation corrected for failures of their predecessors by launching with data partnerships in hand either with academic medical centers or large private healthcare groups. However, these startup companies encountered the twin problems of integrating their tools into the radiology workflow and building a business model around them. Hence they ended up building functional features without any commercial traction.

The third generation of AI companies in radiology realized that success required an understanding of the radiology workflow, in addition to the algorithms and data. These companies have largely converged on the same use case: triage. Their tools rank-order images based on their urgency for the patient, thereby sorting how work flows to the radiologist without interfering in the execution of that work.

The third generations solutions for the radiology workflow are a positive advancement that demonstrate there is a path towards adoption, but there is still much more AI could do beyond triage and worklist reordering. So where should the next wave of AI go in radiology?

Going For The Flow

To date, AI has demonstrated value in its ability to handle asynchronous tasks such as image triage and detection. Whats even more interesting is the potential to enhance real-time image interpretation by giving the computer context that lets it work with the radiologist.

There are many aspects of the radiologists workflow where radiologists want improvements and that AI-based context could optimize and streamline. These include, but are certainly not limited to: setting the radiologists preferred image hanging protocols; auto-selection of the proper reporting template for the case; ensuring the radiologists dictation is entered into the correct section of the report; and removing the need to repeat image measurements for the report.

Individually, a shortcut that optimizes any one of these workflow steps a micro-optimization would have a small impact on the overall workflow. But the collective impact of an entire compendium of these micro-optimizations on the radiologists workflow would be quite large.

In addition to its impact on the radiology workflow, the concept of a micro-optimization compendium makes a feasible and sustainable business possible; whereas it would be difficult, if not impossible, to build a business around a tool that optimized just one of those steps.

Radiology Tools for Thought

In other areas of software development, we are witnessing a resurgence in tools for thought technology that extends the human mind and in these areas, creating a product that improves decision making and user experience is table stakes. Uptake of this idea is slower in healthcare, where computers and technology have failed to improve usability and workflow and continue to lack integration.

The number and complexity of medical images continues to increase as novel applications of imaging for screening and diagnosis emerge; but the total number of radiologists is not increasing at the same rate. The ongoing expansion of medical imaging therefore requires better tools for thought. Without them, we will eventually reach a breaking point when we cannot read all of the images generated, and patient care will suffer.

The next wave of AI must solve the workflow of real-time interpretation in radiology and we must embrace that technology when it comes. No single feature will address this problem. Only a compendium of micro-optimizations, delivered continually and at high velocity via the cloud, will solve it.

Photo: metamorworks, Getty Images

Follow this link:

Real-time Interpretation: The next frontier in radiology AI - MedCity News

Making hiring technology accessible means ensuring both that a candidate can use the technology and that the skills it measures dont unfairly exclude candidates with disabilities, says Alexandra Givens, the CEO of the Center for Democracy and Technology, an organization focused on civil rights in the digital age.

AI-powered hiring tools often fail to include people with disabilities when generating their training data, she says. Such people have long been excluded from the workforce, so algorithms modeled after a companys previous hires wont reflect their potential.

Even if the models could account for outliers, the way a disability presents itself varies widely from person to person. Two people with autism, for example, could have very different strengths and challenges.

As we automate these systems, and employers push to whats fastest and most efficient, theyre losing the chance for people to actually show their qualifications and their ability to do the job, Givens says. And that is a huge loss.

Government regulators are finding it difficult to monitor AI hiring tools. In December 2020, 11 senators wrote a letter to the US Equal Employment Opportunity Commission expressing concerns about the use of hiring technologies after the covid-19 pandemic. The letter inquired about the agencys authority to investigate whether these tools discriminate, particularly against those with disabilities.

The EEOC responded with a letter in January that was leaked to MIT Technology Review. In the letter, the commission indicated that it cannot investigate AI hiring tools without a specific claim of discrimination. The letter also outlined concerns about the industrys hesitance to share data and said that variation between different companies software would prevent the EEOC from instituting any broad policies.

I was surprised and disappointed when I saw the response, says Roland Behm, a lawyer and advocate for people with behavioral health issues. The whole tenor of that letter seemed to make the EEOC seem like more of a passive bystander rather than an enforcement agency.

The agency typically starts an investigation once an individual files a claim of discrimination. With AI hiring technology, though, most candidates dont know why they were rejected for the job. I believe a reason that we havent seen more enforcement action or private litigation in this area is due to the fact that candidates dont know that theyre being graded or assessed by a computer, says Keith Sonderling, an EEOC commissioner.

Sonderling says he believes that artificial intelligence will improve the hiring process, and he hopes the agency will issue guidance for employers on how best to implement it. He says he welcomes oversight from Congress.

Read more here:

Disability rights advocates are worried about discrimination in AI hiring tools - MIT Technology Review

PORTO, Portugal, July 22 (Reuters) - Chinese artist and dissident Ai Weiwei unveiled a 32-meter-tall (105 ft) tropical tree made of iron in the Portuguese city of Porto on Thursday, an artwork he hopes will raise awareness of the devastating consequences of deforestation.

Four years ago, Ai was in Brazil to investigate the threats faced by its forests when he stumbled upon an endangered ancient tree of the Caryocar genus in the northeastern Atlantic forest.

Using scaffolding, a team moulded the tree and shipped the mould to China, where it was cast before being sent to Portugal, Ai's new home, to be assembled and exhibited for the first time. read more

The exhibition, which also includes installations composed of iron tree roots, is taking place at Porto's Serralves museum and park, and will be open to visitors until next year.

"People should look at these works and think of what we could lose in the future," Ai, 63, told Reuters by telephone. "It's... a warning about what we are going to lose if we don't act."

Ai's tree stands leafless, has a hollow trunk and the iron looks rusty, reminding visitors of the environmental threats facing the planet.

In Brazil's Amazon, deforestation has surged since right-wing President Jair Bolsonaro took office in 2019. read more

Bolsonaro has called for mining and agriculture in protected areas of the Amazon and weakened environmental agencies.

"Brazil has a clear policy which sacrifices their best resource: their rainforest, their nature," Ai said. "And that's not just Brazil's best resource...it's planet earth's best resource."

Scientists say protection of the Amazon is vital to curbing climate change because of the vast amount of greenhouse gas its rainforest absorbs.

"The problem is that we never learn from our mistakes... we never really learn a lesson," Ai said, urging the world to prepare for "even bigger" environmental disasters.

Reporting by Catarina Demony and Violeta Santos Moura; Editing by Andrei Khalip, Alexandra Hudson

Our Standards: The Thomson Reuters Trust Principles.

Read the original:

Ai Weiwei unveils giant iron tree to warn people what they risk losing - Reuters

All the sessions from Transform 2021 are available on-demand now. Watch now.

Untether AI, a startup developing custom-built chips for AI inferencing workloads, today announced it has raised $125 million from Tracker Capital Management and Intel Capital. The round, which was oversubscribed and included participation from Canada Pension Plan Investment Board and Radical Ventures, will be used to support customer expansion.

Increased use of AI along with the technologys hardware requirements poses a challenge for traditional datacenter compute architectures. Untether is among the companies proposing at-memory or near-memory computation as a solution. Essentially, this type of hardware builds memory and logic into an integrated circuit package. In a 2.5D near-memory compute architecture, processor dies are stacked atop an interposer that links the components and the board, incorporating high-speed memory to bolster chip bandwidth.

Founded in 2018 by CTO Martin Snelgrove, Darrick Wiebe, and Raymond Chik, Untether says it continues to make progress toward mass-producing its RunA1200 chip, which boasts efficiency with computational robustness. Snelgrove and Wiebe claim that data in their architecture moves up to 1,000 times faster than is typical, which would be a boon for machine learning, where datasets are frequently dozens or hundreds of gigabytes in size.

Each RunA1200 chip contains a RISC-V processor and 511 memory banks, with the banks comprising 385KB of SRAM and a 2D array of 512 processing elements (PE). There are 261,632 PEs per chip, with 200MB of memory, and RunA1200 delivers 502 trillion operations per second (TOPS) of processing power.

One of Untethers first commercial products is the TsunAImi, a PCIe card containing four RunA1200s. App-specific processors spread throughout the memory arrays in the RunA1200s enable the TsunAImi to deliver over 80,000 frames per second on the popular ResNet-50 benchmark, 3 times the throughput of its nearest competitor. According to analyst Linley Gwennap, the TsunAImi outperforms a single Nvidia A100 GPU at about the same power rating, or about 400W of power.

Untether is shipping TsunAImi samples and aims for general availability this summer. The company says the cards can be used in a range of industries and applications, including banking and financial services, natural language processing, autonomous vehicles, smart city and retail, and other scenarios that require high-throughput and low-latency AI acceleration.

Untether AI has a scalable architecture that provides a revolutionary approach to AI inference acceleration. Its industry-leading power efficiency can deliver the compute density and flexibility required for current and future AI workloads in the cloud, for edge computing, and embedded devices, Tracker Capital senior advisor Shaygan Kheradpir said in a press release.

Theres no shortage of adjacent startup rivals in a chip segment market anticipated to reach $91.18 billion by 2025. California-based Mythic has raised $85.2 million to develop custom in-memory compute architecture. Graphcore, a Bristol, U.K.-based startup creating chips and systems to accelerate AI workloads, has a war chest in the hundreds of millions of dollars. SambaNova has raised over $1 billion to commercialize its AI acceleration hardware. And Baidus growing AI chip unit was recently valued at $2 billion after funding.

Toronto-based Untethers total raised now stands at $152 million.

See the original post here:

Untether AI nabs $125M for AI acceleration chips - VentureBeat

Theator, an Israeli-founded, Palo Alto, California-based startup that aims to bring artificial intelligence and computer vision technologies to surgical rooms, has set up a collaboration with Mayo Clinic in the US.

The partnership will allow Theator to work with the clinics urology and gynecology departments and test out its annotation and video analytics technology to improve the surgeons pre-operative preparation and post-operative analysis and debriefing, in order to give surgeons a chance to improve their performance.

Using computer vision, software developed by the firm scans video footage of real-time procedures and their key moments. This allows surgeons to watch video recordings of procedures broken down into their various steps and fast forward or rewind to steps they wish to learn. Users can also get digital summaries of their own surgical performance, with analytics, graphics and rankings, allowing them to pinpoint additional training and skills that might be needed.

The strategic collaboration follows Mayo Clinics investment in Theators $15.5 million Series A funding round earlier this year.

Theator was co-founded in 2018 by Dotan Asselman and Tamir Wolf, and has an R&D center in Israel.

Get The Times of Israel's Daily Editionby email and never miss our top stories

The collaboration with Mayo Clinic will enable the firm to access a rich array of insights from world-class urology and gynecology departments, with the goal of broadening our experience in order to tackle the pressing problems of disparity and variability in surgical care today, said Dr. Tamir Wolf, CEO and co-founder of Theator.

The strategic collaboration will deepen our visual and contextual understanding of surgical best practices and enable us to refine and develop our preoperative preparation and postoperative debrief Surgical Intelligence platform, helping surgeons around the world upskill and raise the standards of care for even more patients, he added.

Youre serious. We appreciate that!

Were really pleased that youve read X Times of Israel articles in the past month.

Thats why we come to work every day - to provide discerning readers like you with must-read coverage of Israel and the Jewish world.

So now we have a request. Unlike other news outlets, we havent put up a paywall. But as the journalism we do is costly, we invite readers for whom The Times of Israel has become important to help support our work by joining The Times of Israel Community.

For as little as $6 a month you can help support our quality journalism while enjoying The Times of Israel AD-FREE, as well as accessing exclusive content available only to Times of Israel Community members.

Read this article:

Israel-founded Theator to work with Mayo Clinic to bring AI to surgical rooms - The Times of Israel

Three-part webinar series

Description:

Many legal professionals are familiar with eDiscovery and the benefits the technology provides. However, what may not be common knowledge is the power that artificial intelligence brings to the table and the techniques that can be applied to achieve legal success.

eDiscovery AI excels at sifting through massive quantities of data to identify specific terms or concepts, even when those concepts are expressed differently. Because an AI system can scan data faster than any human and doesnt get tired or distracted, it can evaluate data sets faster and more easily than a human while maintaining accuracy.

This three-part webinar series aims to help eDiscovery practitioners discover advanced techniques that offer higher levels of automation, completeness, and speed.

Reasons to attend:

Explore uncharted eDiscovery territory. Join us as we uncover the secret techniques used by eDiscovery experts. Learn how to:

Part 1: How to speed up protection and redaction of sensitive data using Auto-Classification and Auto-Redaction

Date: July 29, 2021 | 1pm to 2pm EST

Lawyers handle a tremendous amount of sensitive information every day: clients personal data, including both personally identifiable information (PII) and protected health information (PHI), intellectual property, trade secrets, financial information, and much more.

During the review process, identifying and redacting sensitive information can be a time-consuming, aggravating, and error-prone task. There are two major challenges around redaction: efficiently identifying the pieces of sensitive information that may be hiding within reams of disclosable data and thoroughly redacting that information prior to production.

In this session, Jeffrey Wolff, Director of Legal Technology Solutions at ZyLAB, will demonstrate some advanced techniques review teams can use in order to achieve a more efficient process.

Heres what well cover:

Next events:

Part Two Fact-finding in eDiscovery August 26th, 2021

Part Three Handling non-standard ESI September 30th, 2021

Register now

About the speaker:

Jeffrey Wolff has been serving as eDiscovery Director at ZyLAB since 2015 and currently oversees the product direction and marketing strategy for ZyLABs North American corporate market. Prior to joining ZyLAB, Mr. Wolff has over 25 years of professional experience in information systems and enterprise software. He has been involved in solution architecture, design, and implementation for a variety of major projects within the Department of Defense and Fortune 1000 corporations.

Here is the original post:

[Webinar] The Hidden Gems of eDiscovery AI: Advanced techniques to supercharge your process - July 29th, 1:00 pm - 2:00 pm EST | ZyLAB - JDSupra - JD...

AUSTIN, July 21 (Reuters) - Self-driving startup Argo AI, carmaker Ford Motor Co (F.N) and ride-hail company Lyft Inc (LYFT.O) on Wednesday said they partnered to offer robotaxi trips to Lyft customers in Miami and Austin.

The service is expected to launch in Miami later this year and in Austin next year with a safety driver inside the Ford Escape hybrid vehicles. Over the next five years, the companies want to deploy at least 1,000 robotaxis in multiple cities.

The first truly driverless cars are expected to launch in 2023, said Jody Kelman, head of Lyft's autonomous team.

The partnership marks the first large-scale U.S. collaboration between a carmaker, a self-driving developer and a ride-hailing company. The companies hope to gain valuable insights on how to turn robotaxis into a commercially viable business - a challenge no company has yet answered.

As part of the agreement, Argo AI, which is backed by Ford and Volkswagen AG (VOWG_p.DE), will receive anonomized data on passenger trips and safety incidents. That will allow Argo to optimize its technology and routing to avoid unsafe streets, Argo CEO Bryan Salesky said in a blog post.

In exchange, Lyft will receive a 2.5% stake in the company. At Argo's most recent valuation of $7.5 billion, that equity slice would be worth $187.5 million. Argo, which is currently testing autonomous vehicles in several U.S. cities, in June said it plans to list publicly within the next year. read more

Ford will fuel, service and clean the robotaxi fleets under the partnership.

In traditional ride-hailing services, human drivers make up an estimated 80% of the total per mile cost, according to research firm Frost & Sullivan, underscoring the companies' interest in a driverless future. But self-driving vehicles need to recoup their expensive development costs and still need to be managed and maintained.

"Our job is to generate the maximum revenue out of each of these vehicles by getting the highest utilization," said Lyft's Kelman.

Lyft in April sold its own self-driving technology unit to Toyota Motor Corp (7203.T) for $550 million to focus instead on providing services such as routing, consumer interface and fleet management. read more

Lyft already allows consumers to book rides in self-driving vehicles in select cities in partnership with Alphabet Inc's (GOOGL.O) Waymo and Motional, the joint venture between Hyundai Motor Co (005380.KS) and Aptiv (APTV.N).

Reporting by Tina Bellon in Austin, additional reporting by Paul Lienert in Detroit, editing by Chris Reese

Our Standards: The Thomson Reuters Trust Principles.

Read more from the original source:

Argo AI, Ford and Lyft to launch self-driving ride-hail service in Miami and Austin - Reuters

Anthony Bourdain was a singularly beloved cultural figure. His death by suicide in 2018 at 61 while filming an episode of his CNN travel series Parts Unknown remains, for many, one of the most tragic and baffling public losses of the past few years. Given the intense connection to the chef turned TV personality felt by his fans, and the shock of his death, Roadrunner: A Film about Anthony Bourdain, a bracing, graceful new documentary on Bourdain that features his inner circle, was going to be met sensitively.

On Friday, news broke that the film-makers used artificial intelligence to simulate the television hosts voice for three lines of synthetic audio. In interviews with the New Yorker and GQ, the films director, Morgan Neville, revealed that he fed 10 hours of Bourdains voiceovers into an AI model for narration of emails Bourdain wrote, totaling about 45 seconds. Reaction to the news was startlingly, if perhaps predictably angry. Some outright dismissed the film, which grapples with Bourdains less palatable qualities his obsessiveness, his flakiness, his gnawing impostor syndrome.

When I wrote my review I was not aware that the film-makers had used an AI to deepfake Bourdains voice, tweeted Sean Burns, a film critic from Bostons WBUR who reviewed the film negatively. I feel like this tells you all you need to know about the ethics of the people behind this project.

Others on Twitter, where it was a trending topic, called it ghoulish, a freakishly bad idea, awful.

Neville did not exactly help matters with his comment to the New Yorker that We can have a documentary ethics panel about it later, which felt more flippant than considered. Said panel seems pertinent, but not as a verdict of the film. While much of the discussion has focused on the ethics of reanimating a dead persons voice, the question of AI, in this case, seems like a misdirection. Bourdain wrote the words; we dont know if he read them aloud, but its not synthetic material, nor is it akin to hawking a deceased pop stars hologram performance for money.

The bigger issue is one of disclosure, both to the audience and to Bourdains loved ones. (In response to Nevilles claim in GQ that he checked with Bourdains widow and literary executor just to make sure people were cool with that. And they were like, Tony would have been cool with that, Bourdains ex-wife Ottavia Busia tweeted: I certainly was NOT the one who said Tony would have been cool with that.) If you know which lines to look for, you can hear how the AI voice is a touch stiffer, and a twitch higher, than the real one. But to an average viewer, the difference is fudged. You cant tell a point that may ultimately come to not matter, as audience comfort levels with synthetic audio shift. As Sam Gregory, a film-maker turned non-profit director on ethical applications of video and technology, pointed out in an interview with Helen Rosner in the New Yorker on the ethics of the AI voice, no one blinks an eye when a narrator in a documentary reads a letter written in the civil war.

The queasy part here was the blending of truth with interpretation of truth, fact with simulation, archive with embellishment. Nevilles obfuscation of the AI voice feels deceptive. But then again, all documentaries bend the lines of reality; audiences just often conveniently forget or ignore the artifice in favor of cohesion and momentum. This is, ironically, so much the subject of Roadrunner the blurring of person and persona, the bounded portrait on camera and the ambiguous messiness off it, the persistent burden of fame. The AI model of Bourdains voice, for three lines, is a questionable artistic choice, for sure. But its not an outright transgression that should overshadow a challenging, deeply emotional film.

Roadrunner starts not with Bourdains childhood, which is almost entirely glossed over, but his rebirth, of sorts: fame at middle age, thanks to his bestselling memoir Kitchen Confidential, published at 43 in 2000. Bourdain was, we see, at first an awkward, ungainly but eager student of the camera and the world around him. The bulk of the film traces the swift change of his life post-fame. He left cooking, his partner of nearly 30 years, his anonymity. He started hosting for CNN, married again, had a daughter, shifted his addictive personality an intimidating, exhaustive relentlessness that was once hooked on heroin to jiujitsu, among other things. He chafed against 250 days a year on the road, against the gulf between the ruggedness of his storytelling and the red carpets of promoting the story, against the role of a traditional TV dad he played in spurts at home.

The last third is dominated by his personal unraveling pushing away friends, almost leaving the show and his death. Arguably the more pressing ethical concern is Nevilles decision to not reach out to Asia Argento, the Italian actor/director and Bourdains final romantic partner, who is portrayed as the agent of undoing for both the show and Bourdain. Friends and crew recall an emotionally tense shoot in Hong Kong after Bourdain installed Argento as director at the last minute and fired a longtime cinematographer, how his infatuation with Argento seemed manic and adolescent, how angered he was by paparazzi photos of her and another man shortly before his death.

Neville has said that speaking to Argento would have been painful for a lot of people. Digging into the final days of Bourdains life instantly just made people want to ask ten more questions, he told Vulture. It became this kind of narrative quicksand of Oh, but then what about this? And how did this happen? It just became this thing that made me feel like I was sinking into this rabbit hole of She said, they said, and it just was not the film I wanted to make.

Still, its a lot of focus, words and images on a person who isnt given the opportunity to speak for herself. Its also clear that including Argentos interviews, if she agreed, would have dragged the film into a litigation of Bourdains death rather than an exploration of his life. There are no easy answers here, the type Bourdain eschewed while he was alive.

Roadrunner is ultimately an inviting, haunting, unsettling film, which doesnt hesitate to name the stars frustrating multitudes. Theres Bourdain the preternaturally magnetic host, Bourdain the giddily infatuated boyfriend, Bourdain the exacting boss and unreliable partner, the unbridled friend who once told confidant David Chang that he wouldnt be a good father. Theres his unbent curiosity, his zest for a hit of any experience, and a foreboding emptiness. The scene I cant stop thinking about is a scrap of footage from 2006, while filming a pivotal episode of No Reservations in Lebanon as war erupted between Israel and Hezbollah. Savoring pleasure amid devastation, lost for words, Bourdain simply shakes his head theres no neat way to sum it all up. Theres no fully cohesive portrait of a person in Roadrunner, maybe never any sense to be made. But theres plenty to sit with, to consider, and that feels like a tribute well done.

Read the rest here:

The Bourdain AI furore shouldnt overshadow an effective, complicated film - The Guardian