I'm currently research assistant at Idiap Research Institute and PhD student at École polytechnique fédérale de Lausanne (EPFL) in Switzerland, under the supervision of Dr. James Henderson and Prof. Hervé Bourlard, and previously with Prof. Andrei Popescu-Belis
I'm working on extenting contextual and inter-sentential information for Neural Machine Translation (NMT). My research is funded by SUMMA project.
Previously, I received a Master's degree in Computer Science from the joint program of Universities of Fribourg, Bern, and Neuchatel in Switzerland, and Bachelor degree in Computer Science from Pontificia Universidad Católica del Perú (PUCP).
Machine Translation (MT) models learn to translate from large
amounts of parallel sentences in different languages. The focus on sentences brings a practical
simplification for the task, but also has the disadvantage of missing contextual information while
translating a document.
One key point is that the dependencies among distant words, related to discourse-level constraints, are being ignored, even inside a single sentence. This often results in a lack of cohesion and coherence in the translated text. So, the objective of my PhD thesis is to improve the automatic translation of documents by including discourse-level constraints.