The Most Important Trends in Translation Technology for 2018

Much has been written recently on Technology, Artificial Intelligence (AI) and Machine Translation (MT), but very little of it comes in a clear, concise format with powerful insights delivered succinctly. This is what the memoQ Trend Report 2018 seeks to accomplish with its ‘to the point’ format of short commentaries filled with bold quotes and predictions from each expert contributor.

Seven key trends are discussed in this report. Beyond AI and MT, the report also explores the increasing complexities of CAT tool features, the impact of new media to language services, how the industry will change with cloud-based services, connectivity and big data, and much more. Here’s a quick look at selected quotes and insights from the report.

Machine Translation: Good, but not the Best

“With quantum computing, machine translation and AI would gain a fresh approach with significantly more processing power. The expected results of using quantum computing in the language industry would be to get even more accurate and natural-sounding language.” Jure Dernovšek, Solution Engineer at memoQ, shared in the 2018 Trend Report.

Advertisement

Current advancements in AI are largely driven by Deep Learning algorithms, which is a subset of neural network mathematics. They have existed since 1965 but were not used for practical applications because computing power wasn’t fast enough yet. Now it is, so machine learning has improved significantly in pattern recognition and prediction.

Nonetheless, for it to rival the capabilities of a human brain, two things are needed: a full understanding of the human brain, and the computing power to replicate it. Neural scientists will tell you that mankind is far from the first, whereas the second is in its early stages in the form of quantum computing.

“MT works stunningly well in repetitive, machine-like translation projects where little cultural context and understanding is needed”

Zsolt Varga, Product Owner at memoQ, commented, “MT works stunningly well in repetitive, machine-like translation projects where little cultural context and understanding is needed.” So his prediction is “machines will take over translation jobs only in certain areas”.

The trend report also drew observations that the debate will move very quickly in 2018 to the “ethical, moral and economic implications” of AI.

Gábor Ugray, Kilgray Founder & Head of Innovation at memoQ, probably had this in mind when he said, “(In 2018), the limits of AI as we know it will become clear, and businesses will be concerned about liability in the absence of human agents.”

Varga echoed this concern, “Although the linguistic quality of the output of neural machine translation often looks very much human-like, things can go wrong easily if the machine is left in the dark without cultural context, human experience and expertise.”

The end of the feature race for CAT tools?

“The fact is that nowadays even hard liner, technology-minded power users will not be able to keep up with all the services and features of more advanced tools,” observed Łukasz Rejter, Account Executive at memoQ.

Rejter was referring to the fact that Computer Assisted Translation (CAT) tools, as it is, have gotten far too complex for users. He predicted that “2018 will see an emerging trend for the reduction of complexity and an end to the traditional feature race among developers”. He also felt that new features will increasingly operate in a ‘hidden’ manner, so that they “impact results… without making the user learn or operate something new”.

Indeed, this prediction makes sense if we look at how Adobe, a highly successful company that makes specialized software for creative professionals, has evolved its products. Adobe’s products are of necessity, highly sophisticated and full of various features that even their loyal users may take a while to learn whenever a new one is introduced.

To make their products more user friendly, Adobe begun to build ‘intelligent’ user interfaces some years back. Today this approach has been institutionalized across all their products as Adobe Sensei, their brand name for their proprietary AI system.

This AI system is integrated into all their products and drives the smart features in them, working in the background to anticipate the user’s next move and deliver features with minimal user effort. The language technology industry could certainly follow the steps of the Silicon Valley Giant.

Csaba Gutléber, Lead Product Designer at memoQ, also suggested that rationalization can be achieved through a minimalist approach to interface design and also role-based personalization of the user interface by leveraging on machine learning to automate customization of features.

Paying for customized comfort with data

However, using AI to do user customization has its dark side – the necessity to collect vast amounts of user data in order to train and refine a model. Of course, capturing data on feature usage patterns in CAT tools is unlikely to pose any privacy worries for a translator, but this question is certainly provoking controversy in the wider context of what companies like Google, Facebook, Uber or Amazon are doing in terms of data collection of their users’ online activities and whereabouts.

Katalin Hollósi, Professional Services Consultant at memoQ, commented in the report that most of us will grow increasingly comfortable with this personalization of ‘user experience vs privacy’ conundrum.

“I think 2018 will see more of this: more and more solutions will adapt to our preferences and we will be less and less reluctant to make sacrifices in our privacy”

“Though many of us are ready to wear our tin foil hats when it comes to evil Google/Facebook collecting data on our habits, in fact, 99 of 100 of us welcome the personalized and comfy digital spaces. These spaces have become our natural habitat and we want them to adapt to us,” she said.

She predicted, “I think 2018 will see more of this: more and more solutions will adapt to our preferences and we will be less and less reluctant to make sacrifices in our privacy.”

Nonetheless, the issue of how much personal data privacy to sacrifice may be highly subjective since it involves people from all over the world. It also isn’t just about getting personalized digital comfort in return; it is also an issue of cost. In markets such as China and India, many users are happy to give away some personal data or receive advertisements in exchange for free or heavily subsidized products or services, for e.g., mobile apps or games. This is not the case in many western, developed countries.

At a Slator Roundtable held in March 2017 on “Cybersecurity, Privacy and Data Protection in Asia”, legal expert Elizabeth Cole, Partner at Jones Day, commented that “the reality is that, in today’s world, personal data is an asset,” and it is hard for any organization to avoid having to deal with it.

“The year 2018 will see multiple high-profile data breaches and chilling enthusiasm for online AI services”

As a service provider to businesses, LSPs will inevitably have to deal with personal data and should think more about data protection and having an incident response plan, for as Ugray boldly predicts in the Trend Report, “The year 2018 will see multiple high-profile data breaches and chilling enthusiasm for online AI services.”

It remains to be seen if he is right, but regardless, in 2018 no one can deny that mankind has already stepped into the AI age. The memoQ 2018 Trend Report doesn’t propose to address all the issues that may arise or possess definite solutions, but it does believe in the wisdom of the crowd. So the report actively invites readers to participate in the discourse by sharing their thoughts in surveys sprinkled throughout the report, like this one below.

The Trend Report 2018 was published by the memoQ team on 18 January 2018. The following seven trends are discussed in the Report.

1. The Rise of the Machine: AI, ML, MT and More. What do they mean for the language services industry?
2. Convergence & Connectivity. So many systems, services and platforms; will they connect or collide in 2018?
3. User Experience. Is it time for CAT tools to rationalize features and use AI to make them more user friendly?
4. Tiptoeing on the Cloud, Working the Web. Will SaaS and cloud services take over native apps and storage?
5. Business. Terminology management and business intelligence will evolve language services further from just words in, words out!
6. New Media. How will video streaming and augmented reality influence language services?
7. Factory. What’s getting automated? What’s going obsolete?