Presenter: Flavio Giobergia
Thursday, March 18th, 2021 17:30
Location: Microsoft Teams – click here to join
Flavio Giobergia: Cross-lingual propagation of sentiment Information using aligned word vectors
Deep learning methods have shown to be particularly effective in inferring the sentiment polarity of a snippet of text. However, such techniques typically require large labelled datasets to perform properly. While labelled data is abundant for some languages (e.g. English) it is not as readily available for other less commonly spoken languages. A possible solution to this problem is to extract sentiment information from a source language (for which enough data is available) and propagate it to the target language. In this talk, we will explore some lexicon-based propagation approaches and their limitations. Additionally, we will explore a different propagation approach that makes use of aligned word vectors.
Biography: Flavio Giobergia is a PhD student at the Department of Control and Compute Engineering. He obtained a Master’s degree in computer engineering at Politecnico di Torino and Politecnico di Milano (through the Alta Scuola Politecnica program). His current research focus is on semi-supervised and unsupervised machine learning algorithms, with a particular interest in unsupervised representation learning.