查看原文
其他

官宣全文| 交互式机器翻译公司Lilt宣布2500万美元B轮融资

Spence Green 机器翻译观察 2022-04-24

The languages one learns as a child willinfluence nearly every aspect of their life: their community, their access toinformation, and even their career prospects. I observed this most acutelywhile living in the Middle East, where I met bright and ambitious people whowere often cut off from intellectual work because they didn't speak English. Whilethe capacity for language is shared by all of humanity and one of the mostfascinating aspects of human intelligence, language differences can divide ussocially and economically.

儿童时期学习的语言几乎会影响到儿童生活的方方面面: 他们的社区,他们获取信息的途径,甚至他们的职业前景。在中东生活的时候,我对这一点最为感同身受。在那里,我遇到了一些聪明而有抱负的人,他们常常因为不会说英语而无法从事有知识含量的工作。虽然语言能力是全人类共有的,也是人类智力中最吸引人的方面之一,但语言差异会在社会和经济方面使人群产生分化。

My co-founder John and I have spent ourcareers developing artificial intelligence technology related to humanlanguages. The last decade has transformed the field. Computers still can't useand understand language like people do, but automatic translation technology isnow undeniably useful in assisting people to write more quickly and accurately.We founded Lilt inorder to maximize the impact of this technology for information access.

我和我的联合创始人约翰一直致力于开发与人类语言相关的人工智能技术。过去十年,这个领域发生了翻天覆地的变化——计算机仍然不能像人一样使用和理解语言,但是自动翻译技术在帮助人们更快更准确地输入方面无疑是非常有用的。我们创建 Lilt 是为了最大化这项技术对信息获取的影响。

Institutions—businesses, governments, anduniversities—are the stewards of the world's most critical written knowledge.We reasoned that if we could make translation more efficient and affordable forthese organizations, then everyone would have access to more opportunity.Consumers are already exercising choice by activating machine translation intheir information tools (i.e., browsers, phones, email apps, etc.) when theyencounter language barriers. That's great, but it means that organizationsroutinely lose control of linguistic quality, brand identity, and datasecurity.

各种机构——企业、政府和大学是世界上最重要的书面知识的管理者。我们的理由是,如果我们能让这些机构更有效率、更便宜,那么每个人都将获得更多的机会。当消费者遇到语言障碍时,他们已经在自己的信息工具(例如浏览器、电话、电子邮件应用程序等)中运用机器翻译来做出选择。这很好,但这意味着这些机构会失去对语言质量、品牌身份和数据安全的控制。

What if we could enable organizationsto proactively remove that language barrier?

如果我们能够让这些机构主动消除这种语言障碍,会发生什么?

Today I am pleased to announce that Intel Capital has led our $25M Series B financing toaccelerate this mission. Mark Rostick, VP and Senior Managing Director at IntelCapital, has joined our board. Our existing investors—Sequoia Capital, RedpointVentures, Zetta Venture Partners, and XSeed Capital—all participated in theround.

今天,我很高兴地宣布,我们获得了英特尔资本领投的2500万美元融资,以加速这项使命。英特尔资本公司副总裁兼高级总经理马克 · 罗斯蒂克(Mark Rostick)加入了我们的董事会。我们现有的投资者——红杉资本(sequoia Capital)、红点资本(Redpoint Ventures)、 Zetta Venture Partners 和 XSeed Capital ——都参与了这轮投资。

We met Intel Capital through IntelCorporation. Last summer we started working with Intel to re-think their mainenterprise localization workflow. Our technology-enabled services have enabledIntel to increase quality and turnaround time while reducing costs by 40%year-over-year. As the commercial program accelerated, a deeper partnershipdiscussion began. By late last year we decided to cement the partnership withan investment.

我们通过英特尔公司认识了英特尔资本。去年夏天,我们开始与英特尔合作,重新考虑他们的企业本地化的主工作流程。我们的技术支持服务使英特尔能够提高质量和周转时间,同时每年减少40% 的成本。随着商业项目的加速,我们开始了更深层次的合作关系讨论。到去年年底,我们决定通过投资来巩固这种伙伴关系。

While COVID-19 has compelled all of us tore-evaluate long-held beliefs, we are more certain than ever that:

虽然新型冠状病毒肺炎迫使我们所有人重新评估很多长期持有的信念,但我们比以往任何时候都更加坚信:

  • The future of work will be increasingly digital. 
  • 未来的工作将越来越数字化
  • The future of commerce will be increasingly online. 
  • 未来的商务将越来越多地依赖于网络
Businesses that strategically expand theirdigital reach to all stakeholders—customers, employees, partners—are certain tofind new opportunities.
战略性地向所有利益相关者(客户、员工、合作伙伴)扩展数字影响力的企业,无疑会找到新的机会。
We'll invest this new capital to help ourcustomers—and as a consequence their customers—access opportunity. We'll focuson four areas: customer enablement, research and product, our service model,and people.
我们应用这些新投资来帮助我们的客户——以及他们的客户——获得机会。我们将聚焦四个方面: 客户赋能、研究和产品、我们的服务模式和团队。

Customer Enablement
客户赋能
Today Lilt has two customer offerings:
Lilt目前有两种客户服务:
1.Technology-enabled     services:  Text     localization services and software for businesses. 
1. 技术赋能的服务:为企业提供文本本地化服务和软件
2. Language     infrastructure software: A     complete language software stack, primarily for on-premise government     applications. 
2. 语言基础设施软件:一个完整的语言软件栈,主要用于政府应用。
These offerings differ only by the managerof the professional translators. Businesses typically do not employ full-timetranslators, so Lilt sources and retains them; security-conscious governmentstypically do. Our intelligent software makes both workloads radically moreefficient.
这些服务的不同之处仅仅在于专业翻译的管理者。企业通常不会雇佣全职翻译,因此Lilt会提供译员支持,而对安全较敏感的政府通常会雇佣全职译员。我们的智能软件可以从根本上提高这两种工作负载的效率。
For the technology-enabled servicesoffering, we've learned that for the majority of businesses:
对于技术赋能的服务,我们已经了解到,对于大多数企业来说:
1.Translation is not a core competency, and 
1.  翻译不是核心竞争力
2. They are encumbered with legacy systems and thinking. 
2.  他们被遗留的系统和思维所束缚。
There's a decades-old playbook: run an RFPfor a translation management system (TMS), run an RFP for a service provider(or three), dip your toe in machine translation, but don't let it near the"serious" content. This is the oatmeal of localization: it preventsstarvation and it might even reduce the risk of heart disease, but it won't winany awards for creativity. Our task is to partner with and empower ourpioneering customers to realize transformational business outcomes: shred theexisting playbook and re-think language operations to expand global reach.
有一个几十年的老套路:运行一个翻译管理系统(TMS)的RFP,运行一个服务提供商(或三个)的RFP,在机器翻译中沾沾自喜,但不要让它靠近 "严肃 "的内容。这就是本地化的“燕麦片”:它可以防止饥饿,甚至可能会降低心脏病的风险,但它不会因任何创意而获得回报。我们的任务是与我们的先锋性客户合作,并为其赋能,以实现转型的商业成果:颠覆现有的玩法,重新定义语言运营,扩大全球影响力。
For the language infrastructure offering,we've learned that the volume of information that national securityorganizations must analyze far exceeds their human capacity. They must closethe gap with technology. Our tasks are to integrate our software infrastructureinto mission-critical pipelines, and to equip analysts withproductivity-enhancing technology in the most sensitive applications.
对于语言基础设施的提供,我们已经了解到,国家安全组织必须分析的信息量远远超过了他们的人力能力。他们必须用技术来弥补这个差距。我们的任务是将我们的软件基础设施整合到任务关键的渠道中,并在最敏感的应用中为分析人员配备提高生产力的技术。
Research and Product
研究和产品
Lilt believes that next-generationlocalization requires purpose-built machine learning systems. Simple enginecustomization isn't enough. There are a range of interesting, open researchproblems that we are expanding our team to investigate:
Lilt认为,下一代本地化需要专门的机器学习系统,简单的引擎定制化是不够的。我们正在扩大我们的团队来研究一系列有趣的、开放的研究问题:
  • Online domain adaptation 交互式在线引擎
  • Mixed-initiative systems and collaborative online work 混合机翻系统和协作式在线翻译
  • Grammatical error correction 语法错误自动纠正
  • Application     of classic NLP algorithms   to localization-specific tasks     such as tag and terminology handling 应用经典 NLP 算法解决特定本地化的任务,比如标签和术语处理
Because our systems are purpose-built fromscratch, our customers benefit from the latest breakthroughs in AI researchthat apply to localization. Moreover, governments like that our integratedinfrastructure can deploy in on-premise environments without any dependencieson third-party cloud systems.
因为我们的系统是从零开始建立的,我们的客户受益于面向本地化的人工智能研究最新突破。此外,政府希望我们的集成基础设施可以部署在本地环境中,而不依赖于任何第三方云系统。
Service Model
服务模式
We view business as partnership. Globalenterprises demand round-the-clock, follow-the-sun supply-chain partners. Toincrease supply-chain flexibility we're investing in production:
我们视业务为伙伴关系。全球企业需要全天候、跟随太阳的供应链合作伙伴。为了增加供应链的灵活性,我们正在投资生产:
我们把伙伴关系看作业务。全球企业需要全天候的供应链合作伙伴。为了提高供应链的灵活性,我们正在开展:
  • Broader timezone coverage: We've     recently opened new offices in Indianapolis, Indiana and Dublin,     Ireland. 
  • 更广泛的时区覆盖范围:我们最近在印第安纳波利斯、印第安纳州和爱尔兰的都柏林开设了新的办事处。
  • Additional linguistic services: We're introducing both text pre- and post-processing     services. 
  • 额外的语言服务:     我们正在引入文本预处理和后处理服务
  • Translator community: We'll     expand training and support for our community of world-class freelance     translators (aka, the "Liltlancers").
  •  翻译社区: 我们将扩大培训和支持我们的社区的世界级自由译者(又名“Liltlancers”)
People
团队
Language translation in both the public andprivate sectors is a complex, operationally intense process. The manual methodsof yesterday can't scale with the volume of digital information produced in theworld today. Today, businesses often compromise their non-English customerexperiences due to scale and cost.
无论是公共部门还是私营部门,语言翻译都是一个复杂的、操作性强的过程。过去的手工方法无法与当今世界产生的数字信息量相提并论。如今,由于规模和成本的原因,企业往往会牺牲非英语的客户体验。
We're writing the playbook for thevertically-integrated, technology-enabled supplier that can give our customersa better option. We need tough, humble, quick, and meticulous people to writethe playbook from initial outreach to full-scale production.
我们要写的是垂直整合、科技赋能的供应商的脚本,可以给客户一个更好的选择。我们需要的是坚韧、谦逊、快速、细致的人,来编写从最初的推广到全面生产的脚本。
To attract curious, growth-oriented people,we must provide hard problems, intellectual freedom, and avenues for personaland professional development. We're investing in onboarding, learning, anddevelopment programs not only for our full-time staff but also for theLiltlancers.
为了吸引好奇的、以成长为导向的人们,我们必须提供困难的问题、思想自由以及个人和专业发展的途径。我们正在投资新员工培训、学习和发展项目,不仅是为了我们的全职员工,也是为了社区译员。
We're also investing in a strong leadershipteam from both the language and technology industries. Senior people fromSalesforce, Lionbridge, Welocalize, and SDL have recently joined our team.
我们还将建设一个来自语言和技术行业的强有力的领导团队。Salesforce、Lionbridge、 Welocalize 和 SDL 的一些资深人士最近加入了我们的团队。
In 2011 when John and I first startedworking on machine translation together at Google, we were amazed to learn thatGoogle’s localization team didn't use Google Translate. A decade later weremain amazed that the majority of companies and governments have not embracedmachine learning to expand their global digital reach. The information-accessbaseline is still low. We started Lilt so that organizations can aim high.
2011年,当约翰和我第一次一起在谷歌开始研究机器翻译时,我们惊讶地发现谷歌的本地化团队并没有使用谷歌翻译。十年后,我们仍然很诧异,大多数公司和政府没有采用机器学习来扩展他们的全球数字影响力。信息访问的基线仍然很低。我们创办Lilt,就是为了让这些组织能把目标定得更高。
-END-

译文:公众号@机器翻译观察

原文:Lilt官网,点击阅读原文可查看原文

【往期精选】
机器翻译产品全景图
2019西湖机器翻译论坛的几点杂感
2020语言服务展望之七|李光华:神经网络翻译时代,我们需要一款什么样的笔译工具?
2020语言服务展望之一|蒙永业:2020中国语言服务行业发展趋势
2020语言服务展望之二|吴永波:5G元年开启语言数字世界新十年
2020语言服务展望之三|王民杰:2020年同声翻译市场展望
2020语言服务展望之四|崔启亮:中国语言服务40年回顾与展望
2020语言服务展望之五|韦忠和:全球语言服务行业2019回顾
2020语言服务展望之六|韦忠和:全球语言服务行业2020展望
2019全球语言技术谱系图
2019中国语言服务业发展报告摘录
机器翻译什么时候跟人工翻译一样好
机器翻译全球市场观察|Unbabel获6000万美元C轮融资
机器翻译全球市场观察|语言技术全球版图
机器翻译全球市场观察|语料平台Flitto翻易通近日上市
机器翻译全球市场观察|日本Rozetta公司加码机翻投入
机器翻译全球市场观察|Videolocalize发布多语配音平台
机器翻译全球市场观察|金融机器翻译公司Lingua Custodia
机器翻译全球市场观察|德国Lengoo获投660万美元A轮
2019机器翻译“胡说”系列之八|周兴华
2019机器翻译“胡说”系列之七|蔺伟
2019机器翻译“胡说”系列之六|江伟
2019机器翻译“胡说”系列之五|邹剑宇
2019机器翻译“胡说”系列之四|袁行远
2019机器翻译“胡说”系列之三|李光华
2018机器翻译“胡说”系列之二|娄东来、黄国平、魏勇鹏
2018机器翻译“胡说”系列之一|朱靖波、崔启亮、韦忠和
2019全球语言服务业的一组关键数据
2019全球语言服务供应商名录

您可能也对以下帖子感兴趣

文章有问题?点此查看未经处理的缓存