Today Irene Eleta from university of Maryland visited the lab with a seminar called Multilingual Users of Twitter: Social Ties Across Language Borders or How a Story Could Travel the World.” As me, she has been working a lot with Twitter for her PhD thesis. She is particularly interested in exploring the role of multilingual users in different social media platforms. Among her challenges , she aims to find solutions in
- classifying tweets in a certain language but quoting the name of songs/books/movies in another language.
- Detecting automatic messages in different languages
- Scripts of translators in arabic, jew, etc. (coding)
My suggestion in the first problem is to “ignore” single tweets where two languages are detected and only consider those that have a high probability of being only from one language. The reason is that many people use many english expressions without knowing English. Users may be classified as multilingual when only they are using names of movies in English or using expressions etc.
On the other hand, (maybe what I was more interested about) was about the role of multilingual people in Information propagation. The idea is to measure the real importance of these people in the moment of special events like protests, revolutions, crisis or catastrophic events such as earthquakes. Some of the questions to answer would be:
- Are previously classified multilingual people important somehow in propagating information in special events? My naive hypothesis would be “yes, they are.” Because they know other languages, multilingual people will care more about propagating information to the world so that the world can also understand what is going on….. and in particular, the language chosen to communicate with the world will be English. In order to explore this, temporal analysis will be needed.
Looking from another perspective, the detection of multilingual users and the study of their interactions can trigger the invention of new useful functionalities in many sites. For example, up to now I always have problems with language detections and spell checking using Gmail….wouldn’t it be nice if Gmail will know the “language” that you use with your friends and automatically change the spell checker? It seems for me that up to now Gmail saves the previously used spell checker…and it bother me a lot to be switching languages all the time to avoid those annoying red lines.