Course Description

This course will be divided into two parts:

1) Practical application of text generation in order to introduce two theoretical frameworks : data-driven generation (from databases and knowledge). The class will introduce the various proposed approaches (symbolic, probabilistic, statistical).

2) Sentiment analysis and multilingual opinions : the notion of comparable corpora, state of the art methods of document alignment (via bilingual dictionaries and Latent Semantic Approach), opinion and sentiment detection of texts (different typologies, the different machine learning algorithms used for this detection)