fbpx

LocWorld40 Portugal – Conference Report

Innovation is what the people want, but of late no one has been able to deliver anything truly groundbreaking.

Last week, Estoril in Portugal was the place to be for language industry professionals. While the neighboring town of Cascais was hosting the 28th annual Harley-Davidson meetup to the tune of motorcycles humming and Led Zeppelin braying through the loudspeakers (and yes, Nimdzi team members rocked out there too), all eyes were turned to the 40th edition of the Localization World conference. As usual, the venue was impeccable, the organizers outdid themselves with the pre-conference and after-work events (shoutout to the sponsors,) and Nimdzi joined a slew of first-time exhibitors.

#LocWorld40:

500+ attendees

50 talks, panels & presentations

Three very busy days!

Overture

The opening keynote titled “Go Digital, Be Human,” delivered by Isabel Aguilera, former CEO of Google and General Electric for Spain and Portugal,  put the onus squarely on our shoulders. In an increasingly digitized world, where we’re arguably in part digital beings ourselves, it is our basic human faculties which will help businesses and professionals to carry the day. Intuition, empathy, ethics, experience, compassion are the difference-makers, Isabel opined. The first key to success, according to Isabel, is spending quality time with your clients. Only then can you learn the answers to what they want.

“You need to be good enough to be accepted by the best of your time.”

Isabel’s second tip was to surround yourself with talent. And how do you attract talent? Why, by continuously raising the bar, throwing down the gauntlet in challenge. Give them the right tools so they want to pick it up and recognize their merit. Do not wait and adapt to change. Instead, lead the change.

It’s not surprising the human element is something thought leaders continue to insist upon, no doubt mirroring the wider industry, placing people in the center of the discussion. Some would prefer the talks to focus more on technology, but the human element remains intertwined with any technology advancement we achieve. LocWorld is a lot of things, but it is not a technology-centric conference.

LocWorld is a lot of different things to a lot of different people – but it’s an event for the people first and foremost.

Let’s take a tour of the debated topics:

Predictive analytics for MT

A number of presentations and talks, as well as the discussions during the coffee breaks, centered around machine translation (MT). By now, MT has moved from early adoption to mass adoption. While newcomers to MT try to figure out the basics like pricing and productivity, the savvier are moving into refinement and optimization. One of the key elements to make MT more useful is automated quality estimation – predictive analytics. It shows whether MT is good before a human sets eyes on the output.

Memsource pioneered MT predictive analytics last year in their MTQE feature. Their approach is to compare new MT to human translations made by all of Memsource users – a database with tens of billions of words. Other similar companies do not necessarily have a database of the same size, so they are trying new approaches, for example, based on confidence scores in NMT toolkits. ModelFront, a new startup on the scene, was present at LocWorld too. Companies with AI teams such as TransPerfect, Intento, and Unbabel are working on separate solutions.

LSPs will use these tools to send MT output segments with high scores directly into publication and have translators concentrate on segments with low scores, resulting in faster and less expensive post-editing, and also in changes to pricing.

AI: to build or to buy?

Artificial intelligence (AI) continues to be, unsurprisingly, on people’s lips too – it’s not a question of if it will be affecting the language industry but when. The “Future of Localization Jobs in a Global AI World” talk delivered by Agustin da Fieno Delucchi of Microsoft, addressed ethical considerations and how governments are moving to build a legal framework for the use of AI. There is yet another dilemma looming on the horizon: should you buy or build, once you identify what can be automated? Vital questions, which both localization buyers and service providers alike will need to ponder.

Technology giants preach AI services to LSPs

Technology giants like Amazon, Microsoft, and Yandex sent their representatives to LocWorld to promote cloud AI solutions. They see LSPs as partners that use their tools to build linguistic AI engines for clients. This includes transcription, sentiment analysis, categorization, summarization, and annotation. You can build monitoring and information retrieval systems, chatbots, prediction systems.

The big companies are the experts in drilling, and the clients have the oil reservoirs, but there has to be a layer of builders in the middle to connect the two via the services component. It falls to professional services companies to identify use cases, train models, and take the fall if the solutions fail (and this happens often). So, could LSPs take the role of these mechanics?  We spoke to a few CEOs, and the feelings are mixed.

  • The projects to train models are small: maybe USD 50-250 apiece.
  • The projects are technically complex and require training and know-how, even though the main component is outsourcing — something LSP do pretty well.
  • Niche, industry-specific AI companies that are specialists exactly in this exist, and they can often do a better job than the LSPs.

It is a difficult task, one that requires investment. At the same time, it does not result in a big repeatable business like translation and interpreting. Still, if you don’t do it, you might lose access to AI-enabled projects. Bigger companies need to think long and hard about how to approach this: go in cautiously via partnerships, all-out for the optics, or stay away and focus on their key business.

Goodbye, source text!

The Process Innovation Challenge hosted by Dave Ruane of LanguageWire warrants special mention. It has become a fixture during these past few years at LocWorld and presents an opportunity for bright minds with bright ideas to present their innovative tools and processes. Afterward, the audience gets to vote and designate the winner. This time, it was Giulia Tarditi of Monese who took the overall spoils, presenting a fascinating approach to translation. Her innovation, dubbed “Goodbye, Source Text! Discover a Source-free Translation Scenario That Helps Increase Your Conversion Rates” does exactly as its title suggests – doing away with source text, and instead focusing on describing call-to-action buttons in the Monese app. This has worked wonders for Giulia and her team. On one hand, they unshackled the creativity of the translators, who are often too influenced by the source text. On the other hand, it boosted conversion rates and referrals, offering a glimpse at the possibilities for processes that everyone had thought long set in stone.

Has innovation become the Godot of the language industry?

Perhaps you’re familiar with Samuel Beckett’s play, “Waiting for Godot.” The characters engage in conversation while waiting for Godot, who never actually arrives. The LocWorld40 conference certainly felt that way too at times, with its own Godot being eagerly awaited, and not quite deigning to show up. In our particular case, Godot seems to be innovation.

The panel titled “Is There an Innovators Dilemma? Taking the Pulse of Real Innovation in the Globalization Industry” seemed to trend in one direction – there hasn’t been a transformational innovation entering the language services market for quite some time. Everything happening at the moment seems to be only iteration on existing tools, streamlining the process and building better integrations. That being said, there has been a clear consensus among the speakers at the panel: the incentive to innovate should not be coming only from the buyer-side. LSPs should not be afraid to take risks and invest the time and effort to innovate. As Nimdzi’s very own Renato Beninatto likes to say, disruptive innovation comes from outside of the industry. In that sense, we may not know what it will be until it hits us. What the next big innovation will be, only time will tell.

Pictures courtesy of the LocWorld Conference Facebook page.

 

Stay up to date as Nimdzi publishes new insights.
We will keep you posted as each new report is published so that you are sure not to miss anything.

Related posts