Connect with us

OTHER

Google, Facebook Parent Meta Battle It Out to Create Ultimate AI Translator

Published

on

google,-facebook-parent-meta-battle-it-out-to-create-ultimate-ai-translator

A man from South Africa speaks Sepedi to a Peruvian woman who knows only Quechua, yet they can understand each other. The universal translator is a staple of science fiction, but Google, Meta and others are locked in a battle to get as many languages as possible working with their AI models.

Meta chief Mark Zuckerberg announced on Wednesday that his firm now had a block of 200 languages that could be translated into each other, doubling the number in just two years.

Meta’s innovation, trumpeted in 2020, was to break the link with English — long a conduit language because of the vast availability of sources.

Instead, Meta’s models go direct from, say, Chinese to French without going through English.

In May, Google announced its own great leap forward, adding 24 languages to Google Translate after pioneering techniques to reduce noise in the samples of lesser-used languages.

Advertisement
free widgets for website

Sepedi and Quechua, of course, were among them — so the Peruvian and the South African could now communicate, but so far only with text.

Researchers warn that the dream of a real-time conversation translator is still some way off.

Quantity vs quality

Both Google and Meta have business motivations for their research, not least because the more people using their tools, the better the data to feed back into the AI loop.

They are also in competition with the likes of Microsoft, which has a paid-for translator, and DeepL, a popular web-based tool that focuses on fewer languages than its rivals.

The challenge of automatic translation is “particularly important” for Facebook because of the hate speech and inappropriate content it needs to filter, researcher Francois Yvon told AFP.

Advertisement
free widgets for website

The tool would help English-speaking moderators, for example, to identify such content in many other languages.

See also  WhatsApp Can Now Be Used on Secondary Devices Without Having to Keep Your Phone Online: How to Enable

Meta’s promotional videos, however, focus on the liberating aspects of the technology — amateur chefs having recipes from far and wide appearing at their fingertips.

But both companies are also at the forefront of AI research, and both accompanied their announcements with academic papers that highlight their ambitions.

The Google paper, titled Building Machine Translation Systems for the Next Thousand Languages, makes clear that the firm is not satisfied with the 133 languages it already features on Google Translate.

However, as the cliche goes, quantity does not always mean quality.

Advertisement
free widgets for website

European primacy

“We should not imagine that the 200×200 language pairs will be at the same level of quality,” said Yvon of Facebook’s model.

European languages, for example, would probably always have an advantage simply because there are more reliable sources.

As regular users of tools such as Google Translate and other automatic programmes will attest, the text produced can be robotic and mistakes are not uncommon.

While this may not be a problem for day-to-day use lie restaurant menus, it does limit the utility of those tools.

“When you’re working on the translation of an assembly manual for a fighter jet, you can’t afford a single mistake,” said Vincent Godard, who runs French tech firm Systran.

Advertisement
free widgets for website

And the ultimate nut to crack is inventing a tool that can seamlessly translate the spoken word.

“We’re not there yet, but we’re working on it,” said Antoine Bordes, who runs Fair, Meta’s AI research lab.

He said Meta’s speech translation project works on far fewer languages at the moment.

See also  Delhi High Court Rejects Appeals by WhatsApp, Facebook in CCI Probe Hearing: Report

“But the interest will be in connecting the two projects, so that one day we will be able to speak in 200 languages while retaining intonations, emotions, accents,” he said.


Advertisement
free widgets for website

OTHER

WhatsApp Reveals Critical Vulnerabilities in Older App Versions That Let Attacker Exploit Phones via Video Call

Published

on

By

whatsapp reveals-critical-vulnerabilities-in-older-app-versions-that-let-attacker-exploit-phones-via-video-call

WhatsApp, Meta’s instant messaging and calling service, has published details of a ‘critical’ vulnerability that has been patched in a newer version of the app but might still affect older installed versions that have not been updated.

The details regarding the vulnerability were revealed in a September update of WhatsApp‘s page on security advisories affecting the app and came to light on September 23.

WhatsApp, in the update, shared a detailed issue related to vulnerability CVE-2022-36934, according to which “an integer overflow in WhatsApp for Android prior to v2.22.16.12, Business for Android prior to v2.22.16.12, iOS prior to v2.22.16.12, Business for iOS prior to v2.22.16.12 could result in remote code execution in an established video call.”

According to the details, the bug would let an attacker exploit integer overflow, after which they can get access to execute their own code on a victim’s smartphone through a specially crafted video call.

This vulnerability has been given a severity score of 9.8 out of 10 on the CVE scale.

Advertisement
free widgets for website

In the same security advisory update, WhatsApp also explained another vulnerability, CVE-2022-27492. According to the social media company, “an integer underflow in WhatsApp for Android prior to v2.22.16.2, WhatsApp for iOS v2.22.15.9 could have caused remote code execution when receiving a crafted video file.”

This said, the bug would let attackers execute the code on the victim’s smartphone using a malicious video file. The vulnerability was scored 7.8 out of 10.

In an India-related development for the social media platform, the head of WhatsApp’s India payment business, Manesh Mahatme, has quit after more than a year with the Meta Platforms-owned company to join Amazon India, a source told Reuters on Thursday.

See also  Delhi High Court Rejects Appeals by WhatsApp, Facebook in CCI Probe Hearing: Report

Mahatme’s exit comes at a critical time for WhatsApp, which is seeking to ramp up its payments service in a highly competitive market and lock horns with more established players such as Alphabet’s Google Pay, Ant Group-backed Paytm and Walmart’s PhonePe.

During his stint at WhatsApp Pay, the company won regulatory approval to more than double its payments offering to 100 million users in India, its biggest market with more than half a billion users overall.

Advertisement
free widgets for website

Affiliate links may be automatically generated – see our ethics statement for details.

Continue Reading

OTHER

Elon Musk Seeks to End Pre-Approval of His Tweets, Calls SEC Mandate “Government-Imposed Muzzle”

Published

on

By

elon-musk-seeks-to-end-pre-approval-of-his-tweets,-calls-sec-mandate-“government-imposed-muzzle”

Elon Musk’s lawyers urged a federal appeals court to throw out a provision in his 2018 consent decree with the US Securities and Exchange Commission (SEC) requiring a Tesla lawyer to vet some of his posts on Twitter.

In a brief filed late on Tuesday with the 2nd US Circuit Court of Appeals in Manhattan, lawyers for Musk called the pre-approval mandate a “government-imposed muzzle” that inhibited and chilled his lawful speech on a broad range of topics.

They also said the requirement violated the US Constitution, and undermined public policy by running “contrary to the American principles of free speech and open debate.”

The SEC did not immediately respond to a request for comment outside market hours. It is expected to file its own brief with the appeals court.

Musk wants to overturn part of an April 27 decision by US District Judge Lewis Liman that rejected his bid to throw out the consent decree altogether.

Advertisement
free widgets for website

Liman said Musk’s arguments amounted to a “bemoaning” of requirements he no longer wanted to adhere to now that “his company has become, in his estimation, all but invincible.”

Musk, 51, is worth $259.8 billion (roughly Rs. 21,25,878 crore), nearly twice as much as anyone else, Forbes magazine said on Wednesday.

The decree resolved a lawsuit accusing Musk of defrauding investors with an August 7, 2018 tweet that he had “funding secured” to take his electric car company private, though a buyout was not close. Musk has said the tweet was truthful.

In settling, Musk agreed to let a Tesla lawyer screen tweets that might contain material information about the company.

See also  Google, Apple Urged to Remove TikTok From App Store, Play Store by FCC Commissioner

He and Tesla each also paid $20 million (roughly Rs. 163 crore) in civil fines, and Musk gave up his role as Tesla chairman.

Advertisement
free widgets for website

But the SEC later opened a probe and subpoenaed documents concerning Musk’s and Tesla’s compliance, after Musk asked his followers in a November 6, 2021 tweet whether he should sell 10 percent of his Tesla stake to cover tax bills on stock options.

In Tuesday’s filing, Musk’s lawyers said it was time to rein in the SEC.

“Under the shadow of the consent decree, the SEC has increasingly surveilled, policed, and attempted to curb Mr. Musk’s protected speech that does not touch upon the federal securities laws,” the lawyers wrote. “Any objective served by the pre-approval provision has been served.”

Musk is separately trying to abandon his April agreement to buy Twitter for $44 billion (roughly Rs. 3,37,465 crore), saying the company misled him by downplaying the number of fake accounts.

Twitter has sued Musk to force him to complete the merger at the agreed-upon price, which is 23 percent higher than where its shares closed on Tuesday. An October 17 nonjury trial is scheduled in Delaware Chancery Court.

Advertisement
free widgets for website

The case is Musk v SEC, 2nd US Circuit Court of Appeals, No. 22-1291.


Affiliate links may be automatically generated – see our ethics statement for details.

Continue Reading

OTHER

Meta Disrupts Chinese Propaganda Operation Across Facebook, Instagram Ahead of US Midterm Elections

Published

on

By

meta-disrupts-chinese-propaganda-operation-across-facebook,-instagram-ahead-of-us-midterm-elections

Meta Platforms said on Tuesday it disrupted the first known China-based influence operation focused on targeting users in the United States with political content ahead of the midterm elections in November.

The network maintained fake accounts across Meta’s social media platforms Facebook and Instagram, as well as competitor service Twitter, but was small and did not attract much of a following, Meta said in a report summarising its findings.

Still, the report noted, the discovery was significant because it suggested a shift toward more direct interference in US domestic politics compared with previous known Chinese propaganda efforts.

“The Chinese operations we’ve taken down before talked primarily about America to the world, primarily in South Asia, not to Americans about themselves,” Meta global threat intelligence lead Ben Nimmo told a press briefing.

“Essentially the message was ‘America bad, China good,’” he said of those operations, while the new operation pushed messages aimed at Americans on both sides of divisive issues like abortion and gun rights.

Advertisement
free widgets for website

Another Meta executive at the briefing said the company did not have enough evidence to say who in China was behind the activity.

Asked about Meta’s findings at a news conference, US Attorney General Merrick Garland said his office was “very concerned” about intelligence reports of election interference by foreign governments “starting back some time ago and continuing all the way into the present.”

A Twitter spokesperson said the company was aware of the information in Meta’s report and also took down the accounts.

According to Meta’s report, the Chinese fake accounts posed as liberal and conservative Americans in different states. They posted political memes and lurked in the comments of public figures’ posts since November 2021.

See also  US Senators Ask FTC to Probe TikTok for ‘Repeated Misrepresentations’ Over US Data Access

A sample screenshot showed one account commenting on a Facebook post by Republican Senator Marco Rubio, asking him to stop gun violence and using the hashtag #RubioChildrenKiller.

Advertisement
free widgets for website

The same network also set up fake accounts that posed as people in the Czech Republic criticizing the Czech government over its approach to China, according to the report.

Meta also said it had intercepted the largest and most complex Russian-based operation since the war in Ukraine began, describing it as a sprawling network of more than 60 websites impersonating legitimate news organisations, along with about 4,000 social media accounts and petitions on sites like US-based campaign group Avaaz.

That operation primarily targeted users in Germany, as well as France, Italy, Ukraine and the United Kingdom, and spent more than $100,000 (roughly Rs. 81.8 lakh) on advertisements promoting pro-Russian messages.

On a few occasions, Russian embassies in Europe and Asia amplified the content.

The Russian embassy in Washington said Meta’s move follows “the instructions of the US authorities” and is a violation of freedom of speech.

Advertisement
free widgets for website

“This suggests that American tech giants, who own the most popular Internet resources, have become servants of the US administration’s policy of suppressing dissent,” the embassy said on its Telegram channel.

© Thomson Reuters 2022


Affiliate links may be automatically generated – see our ethics statement for details.

Continue Reading

Trending