We will talk in detail about Abstract Meaning Representation (AMR), a form of deep semantic parsing that allows a detailed but flexible representation of the semantics of one or more sentences, capturing "who does what to whom". The talk first introduces the formalism of AMR graphs and its properties. We will focus on how changes in textual input reflect in the AMR and its compositional properties as a means of capturing sentence variations. This will be followed by an overview of state-of-the art approaches for the main related machine learning tasks: AMR parsing and AMR-to-text. Finally, the incorporation of structure to large language models for AMR parsing will be presented as an example of SoTA AMR parsing.

Texto informado pelo autor.

* Os participantes dos seminários não poderão acessar às dependências da FGV usando bermuda, chinelos, blusa modelo top ou cropped, minissaia ou camiseta regata. O uso da máscara é facultativo, porém é obrigatória a apresentação do comprovante de vacinação (físico ou digital).