CiTIUS language technologies strengthen their international presence with the licensing of the grammar checker “Avalingua” to a University of Oxford spinoff

The transfer of the technology developed by Pablo Gamallo’s group adds to the researcher’s appointment as co-chair of the International Conference on Computational Processing of Portuguese (PROPOR 2022), the main scientific event in the field of natural language processing in the Portuguese-speaking world, which for the first time will include Galician as a working language.

The British company ProWritingAid has just acquired the rights to exploit the Spanish‑language version of Avalingua, a system for automatically correcting and evaluating the linguistic quality of texts, which uses natural language processing techniques based on artificial intelligence to detect and classify errors and deviations at different linguistic levels (spelling, lexis, morphology, syntax and style). The system was developed at CiTIUS (Singular Research Centre in Intelligent Technologies of the University of Santiago de Compostela) by a team of experts in the design and implementation of language technologies: researchers Isaac González, Pablo Gamallo, Marcos Garcia and Daniel Bardanca.

Avalingua, a “diamond in the rough”

The Avalingua project was launched almost a decade ago, with the development of a prototype for Galician created within the former CiTIUS spin‑off Cilenis, which in 2014 received an “Honorable Mention” in the final of Building Global Innovators, an initiative of the prestigious US research centre Massachusetts Institute of Technology (MIT) aimed at promoting innovative projects.

However, despite the various forms of recognition and the platform’s potential, it was only about a year ago that the British company ProWritingAid began to show interest in the tool, particularly in its potential application to the Spanish language. The company, a University of Oxford spinoff that currently markets a powerful grammar checker and style editor for English, saw in Avalingua an opportunity to break into the Spanish‑speaking market. Thus, after a series of conversations and successive agreements, CiTIUS researchers began working on adapting Avalingua to Spanish, supported by funding from the Emprendia Transfer Accelerator (USC), which also managed the technology transfer licence with the company. Final licensing to the UK startup was approved by the USC Governing Council on 25 February 2021.

First Galician‑European presidency of a transoceanic conference

At the same time, CiTIUS researchers Pablo Gamallo and Marcos Garcia have just been appointed co‑chair and workshops and tutorials chair, respectively, of the 15th International Conference on Computational Processing of Portuguese (PROPOR), a highly prestigious scientific conference in the field of computational processing of Portuguese. The event will be held next year at the University of Fortaleza (Brazil) and, for the first time, will include Galician as one of its official working languages.

“This is a magnificent opportunity to promote the publication of new scientific work related to the Galician language,” says Gamallo, who until now has taken on various roles in organising this biennial meeting, in which he has been involved since 2016. From their positions at the conference, Gamallo and Garcia will thus encourage the submission of work related to Galician from research groups in this field of knowledge, which since last 17 May have been invited to share their contributions on both sides of the Atlantic.

The PROPOR conference, which maintains the tradition of a dual American‑European presidency, will be held in 2022 in the Brazilian city of Fortaleza, focusing on theoretical and technological issues related to Portuguese and, on this occasion, also to Galician. “It is a challenge to motivate people who work with our language, the perfect opportunity to further boost Galician‑Portuguese collaboration,” says Pablo Gamallo, who hopes that this new channel of scientific exchange will encourage more and more people to get involved in this area. “There are already quite a few of us,” he recalls, “but the idea is that in this field of language processing there will be more and more of us.”