The aim of the course is to provide an introduction to the fundamental concepts related to the Linguistic aspects of human languages, and Natural Language Processing (NLP) techniques; moreover, in the course, some NLP applications will be presented, e.g. Machine Translation. After successfully completing the course, students will be able to: -describe basic linguistic aspects of human languages. -explain the common computational vector space models for words applied in language technology. -describe the challenges related to word vector models. -know the main neural language models and apply them for different applications.
Course Prerequisites
Basic knowledge of statistics, programming languages, and machine learning.
Teaching Methods
The course will be constituted of both lectures introducing the main topics and sessions where open source tools will be explained and employed. Seminars held by experts at national and international levels may be part of the course.
Assessment Methods
Written examination with closed and open questions. The written examination is aimed at assessing the level of understanding of the basic aspects taught during the course; it is constituted by a set of open and closed questions.
Texts
Emily M. Bender, "Linguistic Fundamentals for Natural Language Processing", Synthesis lectures on human language technologies, Morgan&Claypool Publishers, 2013. Daniel Jurafsky and James Martin, "Speech and Language Processing, 2nd Edition", Prentice Hall, 2008. Yoav Goldberg, "Neural Network Methods for Natural Language Processing", Synthesis lectures on human language technologies, Morgan&Claypool Publishers, 2017.
Contents
This course will first provide the notions of the morphological and syntactic structure of human languages, useful in creating more linguistically aware NLP systems. The course will then introduce some NLP tasks and text representation techniques. Starting from statistical methods to modern neural approaches, an overview of fundamental techniques will be presented and practiced, such as the n-gram model, Word2Vec, the encoder-decoder paradigm, and neural language models. Open-source software for NLP will be introduced and used.
