Computational Linguistics
Milan , December 10, 2014 – 18:43
The algorithm, strictly Italian , analyzes in real time millions of tweets and posts. Breaks them down, word after word, compares them, and start a search semantics of the text by drawing on artificial intelligence techniques. And in the end tells the “sentiment” of Twitter, Facebook and other social networks, or preferences that users of cyberspace have on a subject, on a product, on a party, on an idea.
The new impulse of computational linguistics
is one of the software (though it may seem simplistic to call it that), just awarded internationally, presented in CLiC-it, the first Italian conference computational linguistics that is taking place at the University of Pisa, organized by companies and associations of linguistics, artificial intelligence, linguistics and science of voice. An event during which was sanctioned the creation of a permanent forum that, by combining different knowledge, wants to become an international reference point of computational linguistics. “A multi-disciplinary matter to which the web has given a new impetus – explains Professor Roberto Basili – and combining humanists, technicians and scientists.”
The push given by Chomsky
And it was in Pisa that it all started in the seventies on the input of a grade American researcher, Noam Chomsky, the father of generative linguistics. Roberto Basili, Professor of Artificial Intelligence University of Rome Tor Vergata, he worked to organize the event along with two other experts, Bernardo Magnini, the Fondazione Bruno Kessler and Alexander Ilex, a professor at the University of Pisa.
Distinguishes between irony and true opinion
Pisa-Rome-Trent is considered by experts to the dorsal column of computational linguistics Italian line projects worldwide. Just as the expert system capable of analyzing the “sentiment” of the Internet reading messages automatically. “And above all, discriminating between ironic and contradictory statements – says Basili -. It is a semantic analysis is not trivial, but complex algorithm that takes place in real time and with great precision. What is? To understand, for example, as the people of the Network judge a product just launched, which has a political opinion, what is your opinion on parties, association and projects. “
The super poll neutral
In short, a super poll. With at least three advantages: the speed, the number of respondents (in some cases we speak of hundreds of thousands or even millions of messages) and neutrality, because while an interviewer may unconsciously influence the interviewee, the algorithm interprets opinions free.
Aid also philologists and historians
” The sentiment analysis is expanding – explains Professor Alessandro Lenci – has a great application in identifying the positive or negative polarity and the type of emotion (joy, anger, disgust, etc. .) of the tweet or post, but at the same time can help, thanks to digital humanities , scholars, archaeologists, philologists and historians, increasingly oriented to use computational linguistics to study and explore the texts’ . And it is no coincidence that, during the conference in Pisa, were presented research very heterogeneous: the automatic translation of the Talmud, the study of the simplification of the texts for teaching, from Latin epigraphs, computational analysis of the memories of the First World War.
December 10, 2014 | 18:43
© ALL RIGHTS RESERVED
No comments:
Post a Comment