Interview with machine translation researcher Celia Rico

Synopsis (Description): Celia Rico is a doctor of English Philology from the University of Alicante, she has a master’s degree in Machine Translation (UMIST) from the University of Manchester, and a master's degree in Business Administration (IEDE) from the European University of Madrid. She is currently a professor at the Complutense University of Madrid. Her area of expertise is translation technology, and more specifically, machine translation and post-editing.

Interview with machine translation researcher Celia Rico

Celia Rico: "Machine translation should not be perceived as a threat to the translator, but as the possibility of choosing the adequate tool for each type of text, while always relying on the translator."

Celia Rico is a doctor of English Philology from the University of Alicante, she has a master’s degree in Machine Translation (UMIST) from the University of Manchester, and a master's degree in Business Administration (IEDE) from the European University of Madrid. She is currently a professor at the Complutense University of Madrid. Her area of expertise is translation technology, and more specifically, machine translation and post-editing.

Celia Rico's approach to the world of translation came about thanks to machine translation. Her last year as a student of English Philology was a key moment in her career, as that was when she was introduced to this technology. Years later and thanks to her work, she has become one of the field's most knowledgeable scientists in Spain.

Machine translation technology: Spain's position

"What is happening in Spain in machine translation is very much in line with what is happening in Europe."

Celia Rico, as the expert researcher she is, acknowledges having analyzed some data prior to this interview for the Pangeanic "Pangea AI & Languages" podcast in order to support her answers with her characteristic rigor.

When describing Spain's position in the field of machine translation, Celia Rico referred to the latest ELIS 2021 report on the European translation industry: "I found that 58% of companies surveyed are already using machine translation, and 20% plan on doing so soon." The survey also indicates that in reference to the academic environment, "use is at a huge 74%."

In reference to Spain, the latest report is from 2016, and was conducted by the Autonomous University of Barcelona. It states that "47% of companies use machine translation." For Celia Rico, considering that a few years have already gone by since that report, "there is a lot of interest in machine translation among companies, and usage is high." This feeling is supported by the many specific machine translation courses offered by the various translators' associations. Even "freelance translators are very interested in machine translation and post-editing."

On research in this field, she comments that "a ceiling has been reached with respect to neural engines." For her, the next step is "the interaction between machines and humans; there is still a long way to go." She also mentions "research into single engines (multilingual engines) that can translate from and to different language pairs."

How will traditional translators and machine translation coexist?

There is a lot of controversy around the "difficult" coexistence between traditional translators and machine translation, which Celia Rico acknowledges is more of a "social networks issue." She states that "it is not so much the use of machine translation and its results, but how it is used."

The Spanish Association of Translators, Proofreaders and Interpreters has produced a guide explaining what post-editing is, in which it also makes a recommendation on how to set rates and, above all, focuses on the importance of "balancing companies' needs when translating," as it recognizes the costs involved in translation. Celia Rico confirms that there is already "a large post-editing market and pool of translators who are choosing this type of work."

Deep Adaptive, a tool developed by Pangeanic, allows machines to perform a much deeper analysis, relating the source and target languages in an almost human-like way. This engine allows users to adapt the models with their own data from their field. The model can "learn a company's stylistic usages and patterns." The engines learn through the data provided by humans, creating a unique learning loop for each client. This can be used by law firms, for example. If they have previous legal rulings, they can take advantage of them to train the machine translation engine in question, generating improved translation results.

Specific machine translation training for traditional translators

"It is essential that future translators be familiar with machine translation."

For Celia Rico, specific machine translation training in universities is still being developed. "Translation students currently only receive information tangentially," and this is because curricula take a few years to change. "They are not as agile as we would like due to legislation, and this prevents substantial changes from being introduced." However, the academic world is overcoming this obstacle by "holding seminars, lectures, and courses that promote the use of machine translation among future translators."

The advantages of post-editing

Speaking about the advantages of machine translation post-editing over purely human translation, Celia Rico explains the following: "I question whether post-editing is faster, as it depends on the translation engine, the type of text, and the engine's training, as not all engines are the same." In this sense, she confirms that if the engine is generic, "it will be great for generic texts, but not for specialized ones."

With this in mind, she bases her work on the various research studies carried out to date, which reflect "the greater cognitive effort involved in a post-editing job, since it encompasses everything from the translator's effort put into revising, for example, to the naturalness of the language."

The Poseditrad project and its educational results

Celia Rico spoke about the Poseditrad project, a Universitat de València initiative coordinated by Diana González Pastor, which incorporates the use of machine translation and post-editing in the Degree in Translation and Interlinguistic Mediation.

In this project, a series of surveys were conducted among students and professors, training courses were developed, and teaching material was produced that "can be used freely" by those interested in the subject. The project also highlights all the professors' concerns about how to introduce machine translation into our future translators' university programs.