Belen Ortiz |
Granada (EFE).- In the complex world of artificial intelligence, natural language processing plays an essential role in the goal of humanizing technology, an area of knowledge in which Spanish, one of the most spoken languages in the world, , seeks to strengthen its position against other languages such as English.
Natural language processing is the set of computing and computing techniques that allow machines to understand and generate human language, whether spoken or written, in any language. Its possible applications go beyond automatic translators or virtual assistants like “Alexa” and cover almost all areas of life.
The professor of the Department of Computer Science and Artificial Intelligence of the University of Granada Eugenio Martínez Cámara and the professor of the Department of Computer Languages and Systems Zoraida Callejas Carrión explain in a joint interview with EFE the scientific, but also social and linguistic challenges , confronted with automatic natural language processing.
And for the Spanish scientific community, one of them, they point out, is that “the development of this type of technology in Spanish and in the co-official languages of Spain is happening at the same time as other languages”, a further way of guaranteeing equal access to services.
“If we only depend on the big tech companies, which will develop the systems in the languages in which they consider they will have more users (like English), we will miss the boat,” according to Callejas, whose opinion is seconded by Martínez Cámara, who advocates that Spain “benefit from the enormous wealth of its language” and bet on investing in the processing of the language in Spanish.
Specifically, according to the Cervantes Institute, Spanish is the world’s second native language in number of speakers – only behind Mandarin Chinese – and third in overall count.
And with this objective, the University of Granada, candidate city to host the State Agency for the Supervision of Artificial Intelligence, accumulates a vast experience of more than 20 years of experience in everything related to linguistic technologies and dialogue systems, emphasize the professionals.
Humanize language and avoid prejudice
In order to humanize and make robotic language more natural, explains Martínez Cámara, it is necessary to meet other challenges such as linguistic and phonological ones, so that the machine is able to capture the different accents of the same language, as well as nuances such as irony or sarcasm, multi-expressions, fixed sentences, complex expressions, anaphora, synonymy…
“Something that we humans solve naturally but for which a machine has no capacity, and which in the field of computer assistants is also very important”, explains the expert.
Added to this are other issues such as avoiding bias, “a challenge specific to Artificial Intelligence in general but also to language processing in particular”.
For example, the ability to identify controversial aspects such as whether or not a message has sexist content, which natural language processing must also pay attention to.
New profiles and jobs related to language processing
Dealing with all this complexity of variants requires knowledge not only of computer science, but also of linguistics, underlines Callejas, which has generated the appearance of new profiles and professions such as that of the computational linguist, whose job is to help professionals in computer engineering to process and understand language.
“Language engineers are also emerging, who are computer engineers specializing in language processing,” adds Martínez Cámara, who recalls that language processing does not only refer to spoken language, but is present in a wide variety of languages. applications and devices such as search engines. or machine translation systems.
“You have to remember that the first application that brought us closer to this language processing was the spam filters that we had many years ago,” he says.
From identifying a false complaint to simplifying a text
The possible applications of the processing are not limited to conversational assistants like “Alexa”, but cover many facets: from helping the police to identify a false statement to simplifying a text to make it more readable, passing by researching legal arguments in a sentence, which would facilitate the lawyer’s work, or identifying arguments in a political speech, which would help journalistic work.
For professionals in this area of knowledge, automatic natural language processing is “the great artificial intelligence, because it is about processing human knowledge that is encoded in text, understanding it and communicating it”, conclude -they.