Connect with us


Lilt is building a machine translation business with humans at the core – TechCrunch



Lilt is building a machine translation business with humans at the core – TechCrunch

The ability to quickly and automatically translate anything you see using a web service is a powerful one, yet few expect much from it other than a tolerable version of a foreign article, menu, or street sign. Shouldn’t this amazing tool be put to better use? It can be, and a company called Lilt is quietly doing so — but crucially, it isn’t even trying to leave the human element behind.

By combining the expertise of human translators with the speed and versatility of automated ones, you get the best of both worlds — and potentially a major business opportunity.

The problem with machine translation, when you really get down to it, is that it’s bad. Sure, it won’t mistake “tomato” for “potato,” but it can’t be trusted to do anything beyond accurately translate the literal meaning of a series of words. In many cases that’s all you need — for instance, on a menu — but for a huge amount of content it simply isn’t good enough.

This is much more than a convenience problem; for many language provides serious professional and personal barriers.

“Information on a huge number of topics is only available in English,” said Lilt co-founder and CEO Spence Green; he encountered this while doing graduate work in the Middle East, simultaneously learning Arabic and the limitations placed on those who didn’t speak English.

Much of this information is not amenable to machine translation, he explained. Imagine if you were expected to operate heavy machinery using instructions run through Google Translate, or perform work in a country where immigration law is not available in your language.

“Books, legal information, voting materials… when quality is required, you need a human in the loop,” he said.

Working on translation projects there and later at Google, where he interned in 2011, Green found himself concerned with how machine translation could improve access to information without degrading it — as most of the systems do.

His realization, which he pursued with co-founder John DeNero, was that machine learning systems worked well not simply as a tool for translation, but as tool for translators. Working in concert with a translation system makes them faster and better at their work, lightening the cognitive load.

The basic idea of Lilt’s tool is that the system provides translations for the next sentence or paragraph, as a reference for structure, tense, idiom, and so on that the translator can consult and, at least potentially, work faster and better. Lilt claims a 5x increase in words per hour translated, and says the results are as good or better than a strictly human translation.

“We published papers — we knew the technology worked. We’d worked with translators and had done some large-scale experiments,” Green said, but the question was how to proceed.

Talk to a big company and get them interested? “We went through this process of realizing that the big companies are really focused on the consumer applications — not anywhere there’s a quality threshold, which is really the entire translation industry,” Green said.

Stay in academic research, get a grant and open-source it? “The money kind of dried up,” Green explained: money was lavishly allocated after 9/11 with the idea of improving intelligence and communication, but a decade later the sense of urgency had departed, and with it much of the grant cash.

Start a company? “We knew the technology was inevitable,” he said. “The question was who would bring it to market.” So they decided it would be them.

Interestingly, a major change in language translation took place around the time they were really getting to work on it. Statistical neural network systems gave way to attention-based ones; these have a natural sort of affinity to efficiently and effectively parsing things like sentences, where each word exists not like a pixel in an image, but is dependent on the words nearby it in a structured way. They basically had to reinvent their core translation system, but it was ultimately for the better.

“These systems have much better fluency — they’re just a better model of language. Second, they learn much faster; you need fewer updates to adapt to a domain,” Green said. That is to say, as far as domains, that the system can quickly accommodate jargon and special rules found in, say, technical writing or real estate law.

Of course, you can’t just sprint into the midst of the translation business, which spans publishing, real-time stuff, technical documents, and a dozen other verticals, and say “here, use AI!”

“There’s enormous structural resistance in the industry to automating in any real way,” Green said. There was no way a major publishing house was going to change the way it worked.

“We tried several business models before we found one that works. There really hasn’t been a company that has decided ‘Okay, this human-in-the-loop method is the fundamental way to solve this problem, let’s just build a company around that.’ So we’re vertically integrated, we work with big enterprises and governments, and we just own the entire translation workflow for them.”

A faster method that doesn’t adversely affect translation quality is basically an efficiency multiplier — catnip for organizations that have a lot of content that needs accurate translation but needs to get the most for their money.

Think about it like this: if you’re a company that puts out products in 20 countries that speak as many languages, translation of packaging, advertising, documentation, and so on is a task that’s essentially never done. The faster and cheaper you can get it done, the better, and if you have a single company that can handle it all, that’s just a cherry on top.

“We work with Zendesk, Snap, Sprinklr… we just take over the whole localization workflow for them. That helps with international go to market.” said Green. If a company’s translation budget and process before using Lilt limited it to targeting 5 or 6 new markets in a given period, that could double or triple for the same price and staff, depending on efficiency gains.

Right now the working on acquiring customers, naturally. “In Q4 last year we built our first sales team,” Green admitted. But initial work with governments especially has been heartening, since they have “more idiosyncratic language needs” and a large volume of text. The 29 languages Lilt supports right now will be 43 by the end of the year. A proofreading feature is in the works to improve the efficiency of editors as well as translators.

They’re also working hard on connecting with academics and building the translation community around Lilt. Academics are both a crucial source of translators and language experts and a major market. A huge majority of scientific literature is only published in English because it would be onerous to translate this highly technical text for others.

Green’s pet peeve seems to be that brilliant researchers are being put to work on boring consumer stuff: “Tech companies are kind of sucking up all the talent and putting them on Assistant or Alexa or something.” It’s a common refrain in frontier tech like AI and robotics.

Finally, Green said, “it’s my great hope that we can close this circle and get into book translation as we go on. It’s less lucrative work but it’s the third part of the vision. If we’re able to, it’s a choice where we’ll feel like we’ve done something meaningful.”

Although it may start out as support documents for apps and random government contracts, the types of content and markets amenable to Lilt’s type of human-in-the-loop process seem likely to only increase. And a future where AI and people work in cooperation is certainly more reassuring than one where humans are replaced. With translation at least, the human touch is nowhere near ready to be excluded.

Continue Reading
Click to comment

Leave a Reply

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


Respawn will premiere its ‘Star Wars’ game on April 13th




After years of work, Respawn is nearly ready to show what its Star Wars game is all about. Lucasfilm has announced that EA and Respawn will formally reveal Star Wars Jedi: Fallen Order at a Celebration Chicago panel on April 13th. The two are unsurprisingly shy about details, but you’ll meet a Padawan who survived Order 66 (the command to exterminate the Jedi) and experience what it’s like to live in an era where there are seemingly no Jedi left. You can expect “never-before-released” details of the game, Lucasfilm said, which isn’t hard when the game is largely a secret.

Continue Reading


Spotify launches in India – TechCrunch




Spotify launches in India – TechCrunch

The Daily Crunch is TechCrunch’s roundup of our biggest and most important stories. If you’d like to get this delivered to your inbox every day at around 9am Pacific, you can subscribe here.

1. Spotify launches its streaming service in India

Just for India, Spotify users who do not pay for a subscription can play any song on demand on mobile. There are also playlists for India and a “Starring…” feature that includes music from Bollywood movies.

“Not only will Spotify bring Indian artists to the world, we’ll also bring the world’s music to fans across India,” said Spotify CEO Daniel Ek.

2. FTC creates antitrust task force to monitor tech industry

This isn’t necessarily a precursor to some big action like breaking up a big company or imposing rules or anything like that. It seems more like a recognition that the FTC needs to be ready to move quickly and decisively in tech matters.

3. This is the Stanford thesis presentation that launched Juul

Against a backdrop of public backlash and looming federal regulations, the world’s biggest e-cigarette manufacturer has released video of the original thesis presentation that launched the company.

4. We’re ready for foldable phones, but are they ready for us?

After years of prototypes, the age of foldables has finally arrived.

5. D-Wave announces its next-gen quantum computing platform

With the latest improvements, developers can use the machine to solve larger problems with fewer physical qubits — or larger problems in general.

6. How Amazon took 50 percent of the e-commerce market and what it means for the rest of us

Some thoughts from the former SVP of Walmart’s global e-commerce supply chain.

7. Steam fights for future of game stores and streaming

Cracks are starting to appear in Steam’s armor, threatening to make it the digital equivalent of GameStop — a once unassailable retail giant whose future became questionable when it didn’t successfully change with the times. (Extra Crunch subscription required.)

Continue Reading


FTC ruling sees (TikTok) fined $5.7M for violating children’s privacy law, app updated with age gate – TechCrunch




FTC ruling sees (TikTok) fined $5.7M for violating children’s privacy law, app updated with age gate – TechCrunch

A significant FTC ruling issued today will see video app TikTok fined $5.7 million for violating U.S. children’s privacy laws, and will impact how the app works for kids under the age of 13. In an app update being released today, all users will need to verify their age, and the under 13-year-olds will then be directed to a separate, more restricted in-app experience that protects their personal information and prevents them from publishing videos to TikTok .

In a bit of bad timing for the popular video app, the ruling comes on the same day that TikTok began promoting its new safety series designed to help keep its community informed of its privacy and safety tools.

The Federal Trade Commission had begun looking into TikTok back when it was known as, and the ruling itself is a settlement with

The industry self-regulatory group Children’s Advertising Review Unit (CARU) had last spring referred to the FTC for violating U.S. children’s privacy law by collecting personal information for users under the age of 13 without parental consent. (The complaint, filed by the Department of Justice on behalf of the Commission, is here.), technically, no longer exists. It was acquired by Chinese firm ByteDance in 2017. The app was then shut down mid-2018 while its user base was merged into TikTok.

But its regulatory issues followed it to its new home.

According to the U.S. children’s privacy law COPPA, operators of apps and websites aimed at young users under the age of 13 can’t collect personal data like email addresses, IP addresses, geolocation information or other identifiers without parental consent.

But the app required users to provide an email address, phone number, username, first and last name, a short biography and a profile picture, the FTC claims. The also app allowed users to interact with others by commenting on their videos and sending direct messages. In addition, user accounts were public by default, which meant that a child’s profile bio, username, picture and videos could be seen by other users, the FTC explained today in its press release.

It also noted that there were reports of adults trying to contact children in, and until October 2016 there was a feature that let others view nearby users within a 50-mile radius.

“The operators of—now known as TikTok—knew many children were using the app but they still failed to seek parental consent before collecting names, email addresses, and other personal information from users under the age of 13,” said FTC Chairman Joe Simons, in a statement. “This record penalty should be a reminder to all online services and websites that target children: We take enforcement of COPPA very seriously, and we will not tolerate companies that flagrantly ignore the law.”

COPPA law, of course, becomes a bit complex to implement for apps like TikTok that sit in a gray area between being oriented toward adults and being aimed at kids. Specifically, apps preferred by tweens and teens — like Snapchat, Instagram, YouTube and TikTok — are often clamored for by younger, under-13 kids, and parents often comply.

But some parents are caught off guard by these apps. The FTC says had fielded “thousands of complaints” from parents because their children under the age of 13 had created accounts.

In addition to the $5.7 million fine, the FTC settlement with includes an agreement that will impact how the TikTok app operates.

It says TikTok is now considered a “mixed audience” app, which means there needs to be an age gate implemented on the app. Instead of locking out under-13 users from the TikTok service, younger users will be directed to a different in-app experience that restricts TikTok from collecting the personal information prohibited by COPPA.

TikTok is also complying with the ruling by making significant changes to its app. It will now restrict under-13 kids from being able to film and publish their videos to the TikTok app. It will also take down all videos from kids under 13.

Instead, the under-13 crowd will only be able to like content and follow users. They will be able to create and save videos to their device — but not to the public TikTok network. Nor can they share videos on the app with their friends if they use TikTok via a private account.

As TikTok already has a large number of younger kids on its app, it will push an app update today that displays the new age gate to both new and existing users alike. Kids will then need to verify their birthday in order to be directed to the appropriate experience.

This is not likely going to have an impact on how many kids use TikTok, however. Kids today already know to lie to age pop-ups so they can enter a restricted app. That’s how they set up accounts on Facebook, Instagram, Snapchat and elsewhere.

However, the move at least puts TikTok on a level playing field with other “mixed audience” apps instead of allowing it to pretend U.S. children’s privacy laws do not exist.

TikTok reportedly has been installed a billion times worldwide, according to recent data from Sensor Tower. The company doesn’t publicly disclose its figures, but the FTC says since 2014, more than 200 million users had downloaded the app worldwide, with 65 million accounts registered in the United States.

The Commission vote to authorize the staff to refer the complaint to the Department of Justice and to approve the proposed consent decree was 5-0. Commissioner Rohit Chopra and Commissioner Rebecca Kelly Slaughter issued a separate statement, shared below:

The Federal Trade Commission’s action to crack down on the privacy practices of, now known as TikTok, is a major milestone for our Children’s Online Privacy Protection Act (COPPA) enforcement program. Agency staff uncovered disturbing practices, including collecting and exposing the location and other sensitive data of young children. In our view, these practices reflected the company’s willingness to pursue growth even at the expense of endangering children. The agency secured a record-setting civil penalty and deletion of ill-gotten data, as well as other remedies to stop this egregious conduct. This is a big win in the fight to protect children’s privacy.

This investigation began before the current Commission was in place. FTC investigations typically focus on individual accountability only in certain circumstances—and the effect has been that individuals at large companies have often avoided scrutiny. We should move away from this approach. Executives of big companies who call the shots as companies break the law should be held accountable.

When any company appears to have a made a business decision to violate or disregard the law, the Commission should identify and investigate those individuals who made or ratified that decision and evaluate whether to charge them. As we continue to pursue violations of law, we should prioritize uncovering the role of corporate officers and directors and hold accountable everyone who broke the law.

Continue Reading


Recent Posts

Like Us On Facebook