Diana Nicoleta Popa - From lexical towards contextualized meaning representation

Continuous word representations (word type embeddings) are at the basis of most modern natural language processing systems, providing competitive results particularly when input to deep learning models. However, important questions are raised concerning the challenges they face in dealing with complex natural language phenomena and regarding their ability to capture natural language variability. A first part of the thesis investigates a method for encoding complex phenomena such as entailment within a vector space by enforcing information inclusion. The second part of the thesis proposes a model for incorporating contextual knowledge into word representations by leveraging linguistic information.