In-Q-Tel partners with AI-human translation company Lilt

December 20, 2018


facebook icon facebook icon

The CIA’s investment arm In-Q-Tel partners with AI-human translation company Lilt to offer US government agencies a high-quality solution for enterprise translation.

Read More: CIA venture arm In-Q-Tel invests in VR, AR intelligence, collaboration platform

Founded on the principle that “a person’s native language shouldn’t limit their ability to learn, grow, and support themselves,” Lilt aims to enable businesses — and now the US government — to offer the same experience to all customers and employees irrespective of language.

“Lilt takes a novel approach to translation by combining human skills and machine learning so that text can be translated faster and with superior quality”

Lilt’s machine translation (MT) engine leverages a human-in-the-loop approach and machine learning to continuously learn in real-time from human feedback—this enables enterprises to significantly increase translation speed and quality without spiking costs.

“Lilt takes a novel approach to translation by combining human skills and machine learning so that text can be translated faster and with superior quality,” said A.J. Bertone, Principal, Investments at IQT.

“With this partnership, we look forward to providing government agencies with a world-class translation capability that will improve the flow of information across languages and amplify human translation resources,” he added.

According to a recent article in Quartz, “The most effective translators, at least for complex bodies of text, are neither human nor machine. They are, in fact, centaurs—humans augmented with the support of technology.”

Augmenting the work of human translators with Lilt’s adaptive MT solution allows all types of businesses to localize materials into foreign languages and accomplish strategic objectives.

The company’s carefully vetted translators contribute to the system’s continuous improvement and are chosen for their domain expertise and ability to localize with sensitivity to cultural nuance. Using Lilt’s MT engine, they are able to translate at nearly five times their normal speed while maintaining the same level of quality and accuracy.

In-Q-Tel (IQT) is the independent, not-for-profit strategic investor that identifies and accelerates the development and delivery of innovative technology solutions to support the missions of US government agencies.

intelligence community ai-human translation lilt

Spence Green

“IQT’s support enables Lilt to better serve customers with the highest quality and latency requirements,” said Spence Green, Co-Founder and CEO of Lilt.

“Our mission is to make the world’s information accessible to everyone, regardless of where they were born or what language they speak. Empowering our public institutions to send and receive mission-critical information across languages is a fundamental component of that mission,” he added.

Green met future co-founder John DeNero while working on Google Translate in 2011.

Prior to heading Lilt, Green held many roles in varying organizations from interning at Google and Johns Hopkins University to conducting research at Stanford and managing projects at Northrop Grumman.

In August Northrop Grumman was chosen as one of the performers for DARPA’s Electronics Resurgence Initiative (ERI) program.

Read More: DARPA seeks tech to make sure hypersonic craft don’t burn up at 5X speed of sound

The five-year research initiative led by DARPA’s Microsystems Technology Office is potentially worth upward of $1.5 billion over five years, shared among all participants.

In October 2011, DARPA — the Defense Advanced Research Projects Agency — launched the Broad Operational Language Translation (BOLT) program to attempt to create new techniques for automated translation and linguistic analysis that can be applied to the informal genres of text and speech common in online and in-person communication.

BOLT is aimed at enabling communication with non-English-speaking populations and identifying important information in foreign-language sources by:

  1. Allowing English-speakers to understand foreign-language sources of all genres, including chat, messaging and informal conversation
  2. Providing English-speakers the ability to quickly identify targeted information in foreign-language sources using natural-language queries
  3. Enabling multi-turn communication in text and speech with non-English speakers. If successful, BOLT will deliver all capabilities free from domain or genre limitations.

Lilt is a platform that equips businesses to optimize among speed, quality, and cost for large-scale localization projects. The core technology is an interactive, adaptive machine translation systems that learns in real-time from human feedback and/or existing translation memory data. Adaptation allows the system to progressively provide better suggestions to human translators and higher quality for fully automatic translation.

Modern companies like Zendesk, Snap, Hudson Bay Company, and Sprinklr use Lilt’s technology to reach emerging markets all over the world. Headquartered in San Francisco, and with a growing office in Berlin, Lilt is backed by In-Q-Tel, Sequoia Capital, Redpoint Ventures, XSeed Capital and Zetta Venture Partners.

In October Lilt raised $9.5 million in new funding led by Sequoia Capital along with existing investors Redpoint Ventures, Zetta Venture Partners, and XSeed Capital.

Enterprise localization is a complex production process. Lilt’s choice to integrate vertically means that it must internalize the complexity of the whole process: from the development of new algorithms to sophisticated customer support and workflow integration.

This requires operational discipline, a service disposition toward its customers, and coordinated investments. The first two requirements will be the responsibility of the people its hires and the culture it builds. The investments in people, data, algorithms, and workflow are enabled by both the growth of its business and by fresh funding.


facebook icon facebook icon

Sociable's Podcast