What MT tells us about sentiments in other languages
This is an invited talk that I gave at ConsILR 2022.
Abstract: Machine translation engines are commonly used to translate user generated content on social media sites like Twitter and Facebook. In some cases, the translation happens automatically and the users do not even get the chance to see the original text. However, automatic translation of sentiments and emotions is difficult especially when the input is noisy, as it is the case with user generated content. This talk will discuss various challenges encountered when translating user generated content and proposes a number of solutions.