Connect with us

Social Networking

Facebook to Use AI in Predicting if COVID-19 Patients Need Better Healthcare

Avatar

Published

on

By ANI | Updated: 19 January 2021

American social media giant Facebook is publishing a research conducted by its artificial intelligence (AI) unit in an effort to help healthcare providers predict in advance if a COVID-19 patient may need more intensive care solutions and adjust resources accordingly.

Facebook in a recent detailed blog post said that it had developed two AI models, one based on a single chest X-ray, and another from a series of X-rays that could help forecast if a patient infected by COVID-19 is likely to get worse. A third model predicts the amount of extra oxygen a COVID-19 patient might need.

The research, which can help produce predictions based on chest X-rays, has been done in collaboration with NYU Langone Health’s Predictive Analytics Unit and Department of Radiology.

Facebook’s AI models in general did a better job than a human when it came to forecasting a patient’s need for more intensive care resources up to four days in advance.

Continue Reading
Advertisement
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Social Networking

Facebook Said to Be Probed by US Agency for ‘Systemic’ Racial Bias in Hiring, Promotions

Avatar

Published

on

By Reuters | Updated: 6 March 2021

A US agency investigating Facebook Inc for racial bias in hiring and promotions has designated the probe as “systemic,” attorneys for three job applicants and a manager who claim the company discriminated against them told Reuters on Friday.

A “systemic” probe means the agency, the Equal Employment Opportunity Commission, suspects company policies may be contributing to widespread discrimination.

The EEOC typically resolves disputes through mediation or allowing complainants to sue employers. But agency officials designate a few cases “systemic,” enabling investigators to rope in specialists to analyse company data and potentially bring a broader lawsuit representing entire classes of workers.

Facebook operations program manager Oscar Veneszee Jr. and two applicants denied jobs brought a charge last July to the EEOC, and a third rejected applicant joined the case in December. They have alleged Facebook discriminates against Black candidates and employees by relying on subjective evaluations and promoting problematic racial stereotypes.

The designation of the EEOC’s probe has not been previously reported.

The EEOC has not brought allegations against Facebook. Its investigation, which may last months more, may not result in findings of wrongdoing. The agency declined to comment.

Facebook spokesman Andy Stone declined to comment on the status of the probe or specific allegations but said that “it is essential to provide all employees with a respectful and safe working environment.”

“We take any allegations of discrimination seriously and investigate every case,” he said.

The EEOC brought in systemic investigators by last August and received detailed briefing papers from both sides over the last four months, said Peter Romer-Friedman, an attorney at Gupta Wessler representing Veneszee and the job candidates.

Employment law firms Mehri & Skalet and Katz Marshall & Banks also are helping the workers.

The EEOC’s Baltimore, Pittsburgh and Washington offices are involved, attorneys from the firms said.

Facebook’s counsel, Covington & Burling, did not respond to a request for comment.

Increasing racial and gender diversity has been a persistent challenge for the nation’s largest tech companies, which at times have blamed a shortage of qualified candidates from underrepresented groups. But tech workers have grown emboldened to publicly challenge that notion and allege in formal complaints that biased employment practices cause disparities.

Romer-Friedman said he and his colleagues told the EEOC in a submission last month that one such Facebook policy is awarding employees bonuses of up to $5,000 when a candidate they refer is hired. Referred candidates tend to reflect the makeup of existing employees, disadvantaging Black professionals, he said.

Facebook said about 3.9 percent of its US employees as of last June were Black.

David Lopez, a former EEOC general counsel now teaching at Rutgers University, said that systemic investigations are significant because of the additional resources involved. When they result in allegations of wrongdoing, multimillion-dollar settlements sometimes follow, he said, citing recent cases against Dollar General Corp and Walmart.

In the year ended last September 30, 13 of the 93 EEOC merit lawsuits were systemic, according to agency data.

Last December, the Justice Department accused Facebook of discriminating against US workers broadly, saying it gave hiring preference to temporary workers such as H-1B visa holders.

Alphabet’s Google last month agreed to spend $3.8 million (roughly Rs. 27.8 crores) to settle US government allegations that it underpaid women and unfairly passed over women and Asians for job openings.

© Thomson Reuters 2021

Continue Reading

Social Networking

Twitter CEO Jack Dorsey Auctions First-Ever Tweet as NFT

Avatar

Published

on

By Reuters | Updated: 6 March 2021

“just setting up my twttr” – the first ever tweet on the platform is up for sale after Twitter boss Jack Dorsey listed his famous post as a unique digital signature on a website for selling tweets as non-fungible tokens (NFTs).

The post, sent from Dorsey’s account in March of 2006, received offers on Friday that went as high as $88,888.88 (roughly Rs. 64.04 lakh) within minutes of the Twitter co-founder tweeting a link to the listing on ‘Valuables by Cent’ – a tweets marketplace.

—tw—

Old offers for the tweet suggest that it was put for sale in December, but the listing gained more attention after Dorsey’s tweet on Friday.

NFTs are digital files that serve as digital signatures to certify who owns photos, videos and other online media.

Dorsey’s 15-year old tweet is one of the most famous tweets ever on the platform and could attract bidders to pay a high price for the digital memorabilia. The highest bid for the tweet stood at $100,000 (roughly Rs. 73.1 lakh) at 0125 GMT (6:55am IST) on Saturday.

Launched three months ago, Valuables compares the buying of tweets with buying an autographed baseball card. “There is only one unique signed version of the tweet, and if the creator agrees to sell, you can own it forever.”

A tweet’s buyer will get an autographed digital certificate, signed using cryptography, that will include metadata of the original tweet, according to the Valuables website. The tweet will continue to be available on the Twitter website.

© Thomson Reuters 2021

Continue Reading

Social Networking

Facebook Asked by Russia to Provide Explanation on Blocked Accounts

Avatar

Published

on

By Agence France-Presse | Updated: 5 March 2021

Russia on Thursday demanded an explanation from Facebook after the social media giant said it had derailed a campaign to mislead Russians protesting the arrest of Kremlin critic Alexei Navalny.

The US-based social network said its automated systems detected and disabled 530 Instagram accounts being used in the campaign against protesters who took to the streets in Russia following Navalny’s arrest in mid-January.

“Roskomnadzor has sent Facebook management a letter containing a request to provide lists of accounts to which access has been limited and also to explain the reasons for blocking them,” the Russian communications watchdog said.

Roskomnadzor demanded that Facebook, which owns the image-centric service, also provide proof that the blocked accounts had been involved in “illegal activities”.

The network of Instagram accounts used hashtag and location “poisoning” typically associated with spam or financial scams to drown out posts by protesters, according to Facebook global threat disruption lead David Agranovich.

Some of the Instagram posts suggested people got COVID-19 and died as a result of attending protests, according to samples provided by Facebook.

Facebook reported that 55,000 people followed one or more of the Instagram accounts.

Tens of thousands of opposition supporters took to the streets in January and February to protest Navalny’s arrest and President Vladimir Putin’s two-decade rule.

Navalny was sentenced last month to two and a half years in a penal colony for breaching parole terms while in Germany recovering from a poisoning attack Novichok nerve agent.

Continue Reading

Social Networking

Zoom Recovers From Hours-Long Outage, Says Issues Were Likely Due to Local Internet Glitch

Avatar

Published

on

By Reuters | Updated: 4 March 2021

Zoom Video Communications said it had recovered from an outage where some users were unable to join meetings on the video-conferencing platform on Wednesday.

The company’s status page said all systems were operational.

Zoom said the glitch was likely due to a localised Internet service issue.

The platform has seen a meteoric rise in users, thanks to remote work and online learning against the backdrop of stay-at-home orders.

Outage tracking website Downdetector.com showed nearly 2,000 incidents of people reporting issues with Zoom at its peak at 9am ET (7:30pm IST).

Separately, telecom operator Verizon’s Internet services were also affected earlier, with nearly 5,400 users registering complaints at 10am ET (8:30pm IST), according to Downdetector.

Verizon said it had since resolved the issue. “Earlier today there was an issue impacting Internet routing for some of our customers. The issue has since been resolved,” a spokesperson for the telecom operator said.

In January, thousands of users across the US East Coast faced widespread Internet outage from providers including Verizon, which disrupted services offered by Google, Amazon, Zoom, Slack, and other technology firms.

Downdetector tracks outages by collating status reports from a series of sources, including user-submitted errors on its platform. The outages might be affecting a larger number of users.

© Thomson Reuters 2021

Continue Reading

Social Networking

Facebook Removes Information-Influencing Accounts Run by Military in Thailand

Avatar

Published

on

By Reuters | Updated: 4 March 2021

Facebook has taken down 185 accounts and groups engaged in an information-influencing operation in Thailand run by the military, the company said on Wednesday, the first time it has removed Thai accounts with ties to the government.

The Thailand-based network removed in the latest sweep of “coordinated inauthentic behaviour” on the platform included 77 accounts, 72 pages, and 18 groups on Facebook and 18 accounts on Instagram, Facebook said.

The company said the accounts were linked to the Thai military and targeted audiences in the southern provinces of Thailand, where conflict has flared on and off for decades as insurgent groups continue a guerrilla war to demand independence.

Thailand’s army spokesman declined to comment when contacted by Reuters, citing a matter of policy not to make comments outside of official news conferences.

Some 7,000 people have been killed during the past 15 years as a result of the insurgency in the Malay-speaking, largely Muslim southern region of predominantly Buddhist Thailand.

“This is the first time that we’ve attributed one of our takedowns to links to the Thai military,” Nathaniel Gleicher, Facebook’s head of Cybersecurity Policy, told Reuters in a briefing.

“We found clear links between this operation and the Thai military’s Internal Security Operations Command. We can see that all of these accounts and groups are tied together as part of this operation.”

The network, mainly active in 2020, used both fake accounts and authentic ones to manage groups and pages, including overt military pages and those that did not disclose their affiliations with the military, Gleicher said.

Posed as individuals

Some of the fake accounts posed as individuals from Thailand’s southern provinces, Gleicher said, adding that the network had spent about $350 (roughly Rs. 25,500) on Facebook and Instagram advertisements.

Some 700,000 accounts followed one or more of the pages and about 100,000 accounts joined at least one of the groups, he added.

Gleicher said Facebook took action on the network based on deceptive behaviour and not the content posted, which included support for the military and the monarchy, and allegations of violence and criticism of insurgent groups in southern Thailand.

The move was Facebook’s second takedown of information-influencing operations in Thailand, after one in 2019 involving 12 accounts and 10 pages that used “fictitious personas”.

In October, Twitter also took down 926 accounts it said were linked to the Thai army that promoted pro-army and pro-government content. The army denied it was behind the accounts.

Twitter in November also suspended a Thai pro-royalist account linked to the palace that a Reuters analysis found was connected to thousands of others that spread content in favour of Thailand’s monarchy.

Facebook on Wednesday said it has taken down four other networks from Iran, Russia, and Morocco engaged in such coordinated inauthentic behaviour.

The company said it has removed more than 100 networks engaged in inauthentic behaviour globally in recent years.

© Thomson Reuters 2021

Continue Reading

Social Networking

Facebook to Resume Political Advertisements, Joining Google

Avatar

Published

on

By Agence France-Presse | Updated: 4 March 2021

Facebook said Wednesday it would resume accepting US political advertisements, ending a ban imposed after the November presidential election to stem misinformation.

The leading social network said “political, electoral, and social issue advertisements in the United States” would be permissible again from Thursday.

“We put this temporary ban in place after the November 2020 election to avoid confusion or abuse following Election Day,” a Facebook blog post said.

“Unlike other platforms, we require authorisation and transparency not just for political and electoral advertisements, but also for social issue advertisements, and our systems do not distinguish between these categories.”

Facebook noted that political and social issue advertisements will require verification and disclaimers explaining that the messages are “paid for by” a specific person or group.

It will also keep the messages in its advertisement library to enable researchers and others to view them.

Facebook added it would “use the coming months to take a closer look at how these advertisements work on our service to see where further changes may be merited.”

Facebook’s move followed a similar one by Google last month.

The two big platforms, which dominate digital advertising, suspended paid political messages in part to stem the flow of hoaxes and misinformation, notably from those rejecting the results of the election, won by Joe Biden over Donald Trump.

Facebook in December partly lifted the ban to allow advertisements pertaining to Georgia’s special Senate runoff election in January.

Continue Reading

Trending