San Francisco, CA, September 1, 2022 – Today, Intento, the leading AI integration platform and source of expertise for the AI and localization industries, shares its much-anticipated State of Machine Translation Report for the sixth consecutive year.
The report gives those working in and around the MT landscape an all-encompassing understanding of the current vendors through in-depth analysis and the best strategies to leverage their offerings successfully. The 2022 edition is available for download here.
Among the 31 MT engines evaluated this year, Intento experts identified 16 leaders for 11 language pairs across 9 domains in the report. Other key takeaways include:
- Four more vendors have offered pre-trained MT models since the 2021 MT Report, and several open-source pre-trained MT engines are available. We have evaluated 31 MT engines — among them is NLLB by Meta AI which has just been made public.
- Many engines perform best with English to Spanish and Chinese. Legal, Financial, IT, and Healthcare require a careful choice of MT vendors, as relatively few perform at the top level.
- Despite having several comparable MT engines per language pair, Entertainment and Colloquial show relatively low scores, which may indicate the importance of customization in these domains.
The state of the MT landscape is in constant flux, and Intento’s extensive report also highlights the changes the discussed providers have undergone since the 2021 evaluation. Readers will be given access to a transparent pricing comparison, along with a breakdown of the “best MT per sector”.
Intento’s insights are grounded in data, using COMET to determine how well certain engines process different content samples across several domains. COMET was chosen from among 6 metrics for a better correlation with human translation, as COMET showed the best correlation with human judgment for most language pairs.
Companies of any size and industry can benefit from these insights, no matter their desired content type or business applications.
“We published our first MT report back in 2017 when all you could quickly get were stock models from nine MT providers. Five years later, we see six times more available MT platforms, 10x more language pairs, about the same x200 price difference, and new automated customization possibilities such as domain adaptation and custom glossaries”, says Konstantin Savenkov, Intento CEO and co-founder. “We believe it’s just the beginning of the upcoming Cambrian explosion in Machine Translation technology, fuelled by the need for global businesses to work in a distributed fashion, being inclusive to all their partners, employees, and communities with their native languages.”
This year’s report was made in collaboration with e2f, the go-to communication platform for data collection services, AI data annotation, content and audio data translation, and linguistic data sets. “We’re pleased to extend our expertise to drive exceptional MT evaluation results in the latest Intento report. e2f continues to be a trusted partner throughout the industry because our datasets drive top-quality AI for our customers,” says Michel Lopez, CEO of e2f. He continues, ’’The caliber of our lingosets (golden datasets) is backed by a talented force of translation and data scientists that have enabled e2f to partner with some of the top players in the localization and AI landscape. As a result, enterprises are attaining a new standard of excellence for conversational AI.”
Feel free to download the full PDF here.
With Intento as their partner, global companies can choose, deploy, and improve the best-fit MT and AI services, including automated content creation, transformation, and localization, to produce 20x more multilingual content without increasing their budget.
Since 2017, Intento has offered its patented, ISO-certified AI integration platform to international businesses for use across all industries and functions. Marketing, localization, software development, community, customer service, and office productivity are some sectors where Intento solutions can make a big difference thanks to its many integrations readily available out-of-the-box.
Established in 2004, e2f helps people and machines understand each other fluently, regardless of language, content, and culture. e2f solutions empower Fortune 50 brands to monitor, objectively assess, and improve communications on a global scale.
e2f delivers world-class translation and training data with its proprietary technology stack for translation, quality review, and AI services. e2f offers a global resource pool of skilled professionals in virtually all countries and languages.