Paolo Rosso: It’s About Sharing Useful Things We Do

 

In June we had the honor of being visited by Paolo Rosso, Associate Professor of Technical University of Valencia, Spain. Earlier this year we were lucky to attend the EACL conference run by the European Chapter of the Association for Computational Linguistics, which is Europe’s largest association for computational linguists (http://www.eacl.org/page.php?id=index). Paolo was one of the people responsible for organizing this very important conference and that was how we got a chance to meet this distinguished professor in person in the wonderful Mediterranean Spanish city. Following the event, we were extremely pleased and delighted to discuss in more detail our current projects and share insights about how we could possibly work on some issues together, which was definitely a great honour for us, as at our Lab we are trying to collaborate with as many people as we can both home and internationally.

 

 

We were really happy to learn that Paolo had got so interested in our current projects as authorship profiling is one of his major research interests at this point, that he suggested he visited us here in Voronezh to see our Lab, meet the rest of its members for a few discussions related to our studies and give a series of lectures to students of our University. Paolo certainly has a lot of things to share with global audiences as he has been in this very popular field for over twenty years and has been Invited Lecturer at a few universities of Europe (including Russia), Latin America, Asia and Africa. Computer science was booming at the time when Paolo decided to pursue a Bachelor degree at the University of Pisa, Italy. The fact that the classrooms, as he remembers, were overcrowded proves how many people wanted to do quite the same. There was no doubt about carrying on working in this field as Rosso got a scholarship to do a PhD in Computer Science at Trinity College University of Dublin, Ireland. His studies saw him working with language originally on the phonological level dealing with speech impairments in children subsequently exploring semantics and syntax more.

According to the Professor, being able to travel in order to attend a number of conferences is what he enjoys as it is not just about “doing things for yourself” but “sharing” your knowledge with the world. Not only has he been able to participate at conferences in all different parts of the world such as India, Mexico, Japan, Germany, etc. but he has been involved in organizing them as well (EACL, PAN@CLEF, numerous workshops on Natural Language Processing, etc.). In handling with this particular aspect of his job, Paolo is driven by “the sense of community” as conferences are unique for bringing scientists from around the world together sharing and discussing ideas the way we did back in Valencia as well as receiving some feedback.

As for teaching, Rosso believes there has to be a reasonable amount of it for a researcher to be able to stay committed to their studies. He is happy with the amount of teaching he is currently doing at the Technical University of Valencia where he has been Associate Professor since 2006 and has been teaching Algorithms and Data Structures, Programming, Applications of Computational Linguistics, etc. at the Bachelor and Master levels. What keeps Rosso motivated to carry on teaching, which he admits has its ups and downs, is being able to share his knowledge and expertise with his students and work on some really intricate issues that PhD candidates are welcomed to suggest as well. He is also currently involved in supervising a few PhD students.

There are certainly some requirements a scientist has to follow to get his work “quantified”. Therefore publishing the results of whatever you have been doing is essential to get funding for further research and generally to “improve the state of the art”. Reviewing his colleagues’ papers is what Paolo has done for a few journals including Information Processing and Management; Artificial Intelligence, Knowledge and Information Systems, etc.

Even though there are certain constraints in this job and doing it one can’t necessarily expect to become rich, what Paolo seems to enjoy most is freedom to explore what he finds interesting and useful and “being his own boss”. As studies of social media texts (in the context of authorship profiling, sentiment analysis, etc.) are becoming increasingly popular, Paolo has been working on corresponding issues as well. Irony detection has been of interest as well as being able to detect the polarity of customers’ opinion and most importantly, distinguishing genuinely positive messages and those including elements of irony is what is essential for marketing, security, etc.  Deception detection is another urgent topic that Rosso has been engaged in working on national projects in conjunction with colleagues from other countries. One of his latest projects is called “Arabic Author Profiling for Cyber-Security”. Terrorist threats are a major global concern and as a lot of communication is taking place via social media, it is tremendously important to detect potential threats automatically and to design individuals’ profiles including not only their gender but native language as well.

Nowadays as we are embracing the interdisciplinary approach in science, the methods of performing research have to be reviewed as well as specialists from different fields (e.g., linguists and computer science specialists) find themselves working together on tackling extremely important problems. Rosso calls this type of collaboration “enriching” even though he admits there might be certain difficulties. As he goes on to explain, in computer science they deal with big data sets (e.g., millions of tweets) and focus more on quantitative analysis of “shallow” features, while linguists work on creating datasets with metadata (e.g., personality tests) making use of the qualitative analysis. Paolo argues that there should be no misunderstanding emerging as computer scientists might seem to “simplify” the problem and instead of issuing criticism, we should “compliment each other”.

This is what we are hoping to achieve as we are planning to collaborate with Paolo Rosso and his colleagues while employing a range of machine learning approaches to the data that we have been collecting here at RusProfilingLab as well as organizing PAN shared task on Gender identification in Russian Texts held in conjunction with the FIRE 2017 Forum for Information Retrieval Evaluation in Bangalore, India. Everyone is welcome to participate in the task!

We hope that Paolo has really enjoyed his visit to Voronezh and will be coming back to have more productive discussions as well as to explore our country more as he says that travelling and being able to see the world and meet all kinds of people is what inspires him.