With AI often thrown around as a buzzword in business circles, people often forget that machine learning is a means to an end, rather than an end in itself. For most companies, building an AI is not your true goal. Instead, AI implementation can provide you with the tools to meet your goals, be it better customer service through an intuitive chatbot or streamlining video production through synthetic voiceovers.

随着AI在商业界经常被当作流行语，人们常常忘记了机器学习是达到目的的手段，而不是目的本身。 对于大多数公司而言，构建AI并不是您的真正目标。 相反，AI实施可以为您提供实现目标的工具，无论是通过直观的聊天机器人提供更好的客户服务，还是通过合成配音简化视频制作。

To help shed light on some real-world applications of machine learning, this article introduces five innovative AI software that you should keep on eye on throughout 2020.

为了帮助您了解机器学习的一些实际应用，本文介绍了五种创新的AI软件，您应该在整个2020年密切关注这些软件。

1. Scanta (1. Scanta)

Scanta is an AI startup with a very interesting history. The company started off creating augmented reality games and AR software for social media. You can get a glimpse of some of their technology in this episode of Expedition Unknown on the Discovery Channel:

Scanta是一家具有非常有趣历史的AI初创公司。 该公司开始为社交媒体创建增强现实游戏和AR软件。 您可以在探索频道的“ Expedition Unknown”这一集中窥见他们的一些技术：

Despite their interesting projects in AR, the company has made a huge pivot into a seemingly untapped sector of the AI industry: advanced security for chatbot virtual assistants. Founded by Chaitanya Hiremath in 2016, the company operates out of San Francisco, with additional offices in India.

尽管他们在AR方面的项目很有趣，但该公司还是将其打入了AI行业看似尚未开发的领域：聊天机器人虚拟助手的高级安全性。 该公司由Chaitanya Hiremath于2016年创立，在旧金山以外的地区运营，并在印度设有其他办事处。

Scanta的Chatbot安全软件：VA Shield (Scanta’s Chatbot Security Software: VA Shield)

Today, Scanta’s main service is the protection of virtual assistant chatbots against machine learning attacks. Their solution, known as VA Shield, “analyzes requests, responses and conversations to and from the system to provide a new layer of supervision.”

如今，Scanta的主要服务是保护虚拟助手聊天机器人免受机器学习攻击。 他们的解决方案称为VA Shield ，“可以分析与系统之间的请求，响应和对话，以提供新的监管层。”

Scanta’s security services differ from the safety provided by conventional IT security teams. Often, teams are either unprepared to deal with machine learning attacks or are unaware that such vulnerabilities exist and therefore aren’t tracking them at all. However, vulnerabilities in chatbots do exist. One famous example is the lawsuit by Delta Airlines in 2019. Delta sued its chatbot developer for a large data breach due to vulnerabilities in their chatbot system which resulted in a leak of confidential user data and credit card information.

Scanta的安全服务不同于常规IT安全团队提供的安全。 通常，团队要么没有做好应对机器学习攻击的准备，要么就没有意识到这种漏洞的存在，因此根本无法跟踪它们。 但是，聊天机器人中确实存在漏洞。 一个著名的例子是达美航空在2019年提起的诉讼 。 由于其聊天机器人系统中的漏洞，Delta起诉其聊天机器人开发人员存在大量数据泄露，从而导致机密用户数据和信用卡信息泄露。

As more and more companies begin adopting chatbots as part of their customer service infrastructure, it is possible that these attacks will become more commonplace. Scanta is positioning itself to become an industry leader in Chatbot security services, but also plans to extend its reach to provide security for other machine learning technologies.

随着越来越多的公司开始采用聊天机器人作为其客户服务基础架构的一部分，这些攻击有可能会变得越来越普遍。 Scanta将自己定位为成为Chatbot安全服务的行业领导者，但还计划扩大其覆盖范围，以为其他机器学习技术提供安全性。

2.描述 (2. Descript)

Descript is a software company that develops products for content creators. Founded in 2017 by Andrew Mason (co-founder of Groupon), the company operates out of San Francisco.

Descript是一家为内容创作者开发产品的软件公司。 该公司由Groupon的联合创始人安德鲁·梅森(Andrew Mason)于2017年创立，在旧金山以外的地区运营。

Descript的Synthetic Voice AI软件 (Descript’s Synthetic Voice AI Software)

Descript’s main product is their video and audio editing software made for podcasters and video content creators. However, in 2019 the company acquired Lyrebird, an AI startup from Montreal, Canada. Lyrebird now operates as Descript’s AI research team, which is working on automated speech-to-text and synthetic voice technology.

Descript的主要产品是为播客和视频内容创建者制作的视频和音频编辑软件。 然而，在2019年，该公司从加拿大蒙特利尔收购了AI初创公司Lyrebird 。 Lyrebird现在作为Descript的AI研究团队运作，该团队致力于自动化语音到文本和合成语音技术。

Synthetic voice technology is a niche sector of the AI industry. However, synthetic voices show a huge potential to improve video game development and film-making. Lyrebird is one of the first companies to dive into synthetic voice development and have thus garnered attention from media outlets such as Wired and Techcrunch.

合成语音技术是AI行业的一个细分领域。 但是，合成声音在改善视频游戏开发和电影制作方面显示出巨大的潜力。 Lyrebird是最早从事合成语音开发的公司之一，因此赢得了Wired和Techcrunch等媒体的关注。

3.复制品 (3. Replica)

Replica is an AI startup that also develops synthetic voice technology. Founded in December 2017 by Shreyas Nivas, Riccardo Grinover, and Keni Mardira, the company operates out of Australia and the United States. In a 2019 interview, CEO Shreyas Nivas said that they were building “a marketplace for the world’s voices” where voice actors and regular people could license their voices to be used in video games, commercials, television programs, and any other form of media that requires voiceovers.

Replica是一家AI创业公司，还开发了合成语音技术。 该公司由Shreyas Nivas，Riccardo Grinover和Keni Mardira于2017年12月成立，业务遍及澳大利亚和美国。 在2019年的一次采访中 ，首席执行官Shreyas Nivas表示，他们正在建立“世界声音市场”，在这里声音演员和普通人可以将其声音许可用于视频游戏，商业广告，电视节目以及任何其他形式的媒体。需要画外音。

Replica Studios合成语音软件 (Replica Studios Synthetic Voice Software)

Replica Studios is an industry-leading platform that allows game developers, video content creators, and the general public to create and train their own synthetic voices.

Replica Studios是行业领先的平台，允许游戏开发人员，视频内容创建者和一般公众创建和训练自己的合成声音。

From video game development to narration of television programs, there are many applications for synthetic voice technology. One of the most interesting and beneficial use cases may be the creation of a synthetic voice for people with health conditions, such as ALS, which cause them to lose the ability to speak. By recording their voice before they lose their speaking ability, we can create a synthetic copy of their voice to be used in speech aid devices.

从视频游戏开发到电视节目旁白，合成语音技术有许多应用。 最有趣和有益的用例之一可能是为具有健康状况的人(例如ALS)创建合成语音，这会导致他们丧失说话能力。 通过在他们失去说话能力之前记录他们的声音，我们可以创建他们的声音的合成副本以用于语音辅助设备。

With huge improvements in the Replica Studio platform and the release of the Replica speech generation API, this is definitely one company to keep your eye on.

随着Replica Studio平台的巨大改进以及Replica 语音生成API的发布 ，这绝对是一家值得关注的公司。

4. Clearview (4. Clearview)

One of the most infamous names in the AI industry today, Clearview is a company that provides a reverse face image search solution for law enforcement. Using a state-of-the-art facial recognition algorithm, Clearview scans the face image of a target subject and then scours the internet for all publicly available images that may match the facial features present in the target image. Founded in 2017 by Hoan Ton-That and Richard Schwartz, the company operates out of New York City.

Clearview是当今AI行业最臭名昭著的名字之一，是一家为执法部门提供反面图像搜索解决方案的公司。 使用最先进的面部识别算法，Clearview扫描目标对象的面部图像，然后在互联网上搜寻可能与目标图像中存在的面部特征相匹配的所有公共可用图像。 该公司由Hoan Ton-That和Richard Schwartz于2017年创立，在纽约市以外的地区运营。

Clearview’s Facial Recognition AI Software

Clearview的面部识别AI软件

Clearview houses a large facial image database comprised of images that are publicly available on the web through social media, blogs, and other websites. The ultimate goal of Clearview is to provide law enforcement with powerful tools to catch criminals. However, many people have concerns about how the existence of such technology will affect privacy laws. In fact, the New York Times did an exposé about Clearview and claimed that the company might “end privacy as we know it.”

Clearview拥有一个大型的面部图像数据库，其中包含可通过社交媒体，博客和其他网站在网络上公开获取的图像。 Clearview的最终目标是为执法人员提供强大的工具来抓捕罪犯。 但是，许多人担心这种技术的存在将如何影响隐私法。 实际上，《纽约时报》曾就Clearview进行过曝光，并声称该公司可能“终止我们所知的隐私”。

Luckily, laws in the GDPR give individuals control of their own data, allowing users to request their profile from Clearview, should they wish to see it. Hopefully, more countries will follow suit and enact private data laws similar to the GDPR.

幸运的是，GDPR中的法律使个人可以控制自己的数据， 允许用户在希望查看的情况下向Clearview 请求其个人资料 。 希望更多的国家能效仿并颁布类似于GDPR的私人数据法。

Clearview is likely going to be at the forefront of the debate around the ethics of facial recognition for the foreseeable future. Any resulting regulations or lack of regulations against the company will set a precedent for other facial recognition developers and startups in the future.

在可预见的未来，Clearview可能会成为围绕面部识别伦理的辩论的最前沿。 任何由此产生的法规或缺乏针对该公司的法规都将在将来为其他面部识别开发商和初创公司树立先例。

5. Lionbridge AI (5. Lionbridge AI)

Lionbridge is a global AI training data provider and data collection company. The company leverages over 50 offices worldwide and a community of over 1 million contributors to create training data at scale. Founded in 1996, Lionbridge began as a language services provider and bolstered its expansion into the machine learning industry by acquiring Gengo AI in January of 2019.

Lionbridge是一家全球AI培训数据提供商和数据收集公司。 该公司在全球拥有50多个办事处，并拥有超过100万贡献者的社区，可以大规模创建培训数据。 Lionbridge成立于1996年，最初是一家语言服务提供商，并于2019年1月收购了Gengo AI，从而促进了向机器学习行业的扩展。

Lionbridge AI’s Data Annotation Software

Lionbridge AI的数据注释软件

The company recently announced the release of the Lionbridge AI Platform, standalone software for image, video, audio and text annotation. Using the platform, data science teams can upload their data, invite other team members, and annotate their datasets together through a collaborative effort. Teams can also track progress and output from individual contributors.

该公司最近宣布发布Lionbridge AI平台 ，这是用于图像，视频，音频和文本注释的独立软件。 使用该平台，数据科学团队可以通过协作来上传数据，邀请其他团队成员并一起注释其数据集。 团队还可以跟踪各个贡献者的进度和输出。

Hailed by Forbes as one of America’s largest employers, such an influential company expanding its reach in the AI market is something to take note of. If the company’s expansion into the AI industry is able to replicate their success in translation and localization, their data annotation software could help data science teams both large and small get access to high-quality training data.

被《福布斯》(Forbes)誉为美国最大的雇主之一 ，这样一家颇具影响力的公司扩大了在AI市场的影响力。 如果该公司向AI行业的扩张能够复制他们在翻译和本地化方面的成功，那么他们的数据注释软件可以帮助大型和小型的数据科学团队访问高质量的培训数据。

Above were just five companies creating AI-powered software, as well as tools to help data science teams build better AI. With new AI startups entering the industry every year, there is a multitude of companies using machine learning to create innovative technologies.

以上只有五家公司创建了AI驱动的软件，以及帮助数据科学团队构建更好AI的工具。 每年都有新的AI初创公司进入该行业，有许多公司使用机器学习来创建创新技术。

To keep up with all the latest in machine learning, please follow me on Medium.

为了跟上机器学习的所有最新动态，请在Medium上关注我 。

Original article reposted with permission.

原始文章经许可重新发布。

Gain Access to Expert View — Subscribe to DDI Intel

获得访问专家视图的权限- 订阅DDI Intel