We spent the last week in beautiful Valencia, attending this year’s EACL. Besides presenting four of our papers and a poster, we had some great time sightseeing the town, swimming in the freezing sea, and drinking some traditional sangria and agua de Valencia! We are quite happy with this vacation before the real summer vacation!
Last week we visited Valencia, a beautiful port city on Spain’s southeastern coast, but also a venue for this year’s EACL! We arrived on Sunday with just enough time to roam around Valencia for a bit before the conference. Luckily, we were all situated in the city center.
The first day of the conference started off with the Linguistic Annotation Workshop (LAW 2017) at which Maria Pia presented our poster, Two Layers of Annotation for Representing Event Mentions in News Stories, something we are dealing with in our EVERBEST project. The others were free to roam around (or bug Maria Pia) as other presentations were slated for the following days.
Tuesday was quite more engaging: we presented our papers at the Balto-Slavic NLP workshop (BSNLP 2017) and the Multiword Expressions workshop (MWE 2017). Domagoj started off with his talk, A Preliminary Study On Croatian Lexical Substitution, in which he presented the construction of the first Croatian dataset for this task and all the lessons learned.
Maria and Domagoj on the left, Paula, Maja, and Leon on the right.
Luckily for Paula, Leon, and Maja, their presentations were scheduled for after the lunch break (in the form of a lovely picnic outside the conference building). 🙂 First, Paula presented her paper, Debunking Sentiment Lexicons: A Case of Domain-Specific Sentiment Classification for Croatian, in which she explored whether it makes sense to rely on sentiment lexicons when dealing with a small amount of labeled sentiment data. On the other hand, Leon described the thorough analysis of sentiment analysis methods across multiple datasets, which he had done for his Comparison of Short-Text Sentiment Analysis methods for Croatian paper.
For the final TakeLab’s presentation of the day, Maja left her mark at the Multiword Expressions workshop. She presented her work, Combining Linguistic Features for the Detection of Croatian Multiword Expressions, in which she talked about how modeling linguistic feature interactions is important for this task, especially when dealing with a morphologically complex and relatively free word order language such as Croatian.
Having presented everything we wanted, we took our time to enjoy Valencia for a while: we walked around the city, visited the beach, said hello to sharks at the Oceanographic aquarium, and much more! This conference was nothing short of being a vacation before the summer vacation! 🙂
Maja, Domagoj, Paula, Jan, and Leon.