Meet Dr. Patrik Lambert, Iconic’s newest Machine Translation Scientist
It’s hard to believe we’re already 3 months into 2017. Suffice to say that the year has gotten off to a very fast start, not least in the world of machine translation, where we’re in the midst of a paradigm shift in terms of research activity. As a leading provider of translation technology, the onus is on us to stay ahead of the curve in terms of research and development. The best way to do that is by having the best minds in our company. With that being said, we’re delighted to welcome Patrik Lambert to our growing team. After allowing him to settle in for a few weeks, we spoke with Patrik about his background and how he arrived at Iconic. Check it out below!
Could you introduce yourself for our website readers?
Hi, my name is Patrik Lambert. I’m a specialist in Statistical Machine Translation and Cross-language Natural Language Processing. Following completion of a PhD at UPC University in Barcelona, I worked as a post-doctoral researcher at both Dublin City University (DCU) in Ireland and Le Mans University in France. I then worked as a Marie Curie Fellow at the Barcelona Media Innovation Centre and UPF University in Barcelona, before undertaking a project for an e-commerce company. I was focused mainly on word alignment, domain adaptation for machine translation, and cross-language sentiment analysis.
Can you describe your role at Iconic?
I am involved in building high quality custom machine translation engines adapted to our clients’ domains and linguistic resources. I also develop new features to improve existing engines, ensuring state-of-the-art quality levels. This includes developing neural MT systems and integrating them into Iconic’s proprietary Ensemble architecture, with a focus on MT for the legal, eDiscovery, life sciences, and financial services domains.
“I see Iconic as a company with a strong commitment to true, deep MT expertise.”
What attracted you to Iconic?
I have previously worked as both a translator and machine translation researcher. I felt the only side of MT I had yet to explore was a MT company. I see Iconic as a company with a strong commitment to true, deep MT expertise, with a similarly strong background in MT research. It seemed like the perfect place to work with other MT experts and apply my knowledge to produce new innovations and the best possible translation software, while improving real-world systems.
You have been involved in machine translation research since 2003, what drew you to it and what do you think about the development of the technology over that time?
After graduating from college with an MSc in Physics, and long before I had even thought about studying machine translation, I ended up taking a job at a translation company as I’ve always had a passion for language. This is where I got to work with both languages and computers. The natural combination of the two was, of course, machine translation.
Since 2003, the development of the technology has been so steady year on year that it’s been difficult to appreciate the vast improvements that have been made. But when we compare the state of the technology in 2003 to today, we can see an extremely impressive degree of advancement. What seemed like breakthrough innovations in translation technology back then would be considered amateurish and unacceptable today. We’ve gone from a technology that was theoretically very powerful but with significant barriers to use, to one that is being used in every level of society, from custom enterprise solutions to phone apps that millions of people use each day.
You spent several years working on the CrossLingMind project for automated sentiment analysis. Can you tell us a bit about it?
We developed a sentiment analysis classifier to predict the polarity (positive, negative or neutral) of opinion segments within text. At the time, we wanted to apply this to Spanish content, but we unfortunately only had English data with which to train the sentiment analysis classifier. In the end, we decided to translate the annotated examples from English into Spanish, which we used to train a Spanish sentiment analysis classifier. This is a great example of the usefulness of MT for cross-language natural language processing. An interesting conclusion was the better the quality of the MT, the better the accuracy of the sentiment analysis classifier in the target language. Ultimately, MT played an important role in the project.
“Neural isn’t the be all and end all of MT”
There has been a lot of hype surrounding neural MT and the possibilities it presents, do you think it can live up to that hype?
Neural isn’t the be all and end all of MT – I don’t think it is going to replace all existing systems and methods in the short- to medium-term. However, in certain situations, where there is a large amount of data available I believe it will live up to the hype, delivering more natural sounding translations with greater fluency. When that body of data is available it’ll also make tackling more complex languages and content types much easier.
You’ve lived here in Ireland before, how does it feel to come back?
Although I work remotely from Barcelona, I am always happy to be back in Ireland. Since the excellent Erasmus year I spent in Dublin, this city has always been a magical place for me.
About Iconic Translation Machines
Iconic Translation Machines is a leading machine translation software and solutions provider who specialises in custom solutions tailored with subject matter expertise for specific industry sectors including legal, life sciences and financial services. Iconic is the MT partner of choice for some of the world’s largest translation companies, information providers, and government and enterprise organisations, helping them to translate more content, more accurately and in less time, resulting in significant cost savings and increased revenue.