Annotation for Machine Learning

Faculteit Letteren
Jaar 2019/20
Vakcode LIX027B05
Vaknaam Annotation for Machine Learning
Niveau(s) bachelor
Voertaal Engels
Periode semester II b
ECTS 5
Rooster >>>

Uitgebreide vaknaam Annotation for Machine Learning
Leerdoelen In this course, students will learn how to define an annotation goal for a given NLP task, how to collect a dataset, define an annotation model, evaluate and adjudicate the annotations, create a gold standard corpus, perform a variety of statistical analytics over the corpus, train several machine learning (ML) algorithms on the dataset by selecting effective features drawn from the annotations, evaluate the learned ML models and compare their evaluation results, come up with a (strong) baseline model for the initial NLP task.
Omschrijving Annotated natural language data is a crucial component in natural language processing. Annotation of data is a process of adding meta information to the text in order to facilitate computers to learn natural language processing. The course will examine how annotation of natural language text can lead to increase the performance of machine learning algorithms. In particular, the course will teach a multistage process for building your own annotated natural language dataset (aka corpus) for training and testing machine learning algorithms for a particular task.
Uren per week 4
Vaksoort bachelor jr 2
Coördinator L. Abzianidze, PhD.
Docent(en) L. Abzianidze, PhD.
Verplichte literatuur
Titel Auteur ISBN Prijs
Natural Language Annotation for Machine Learning (download a free ebook from the RUG library) James Pustejovsky & Amber Stubbs 9781449306663
Entreevoorwaarden Na het behalen van Inleiding Programmeren I (LIX021P05) en Taaloptimalisatie (LCX023P05).
Opmerkingen
Opgenomen in
Opleiding Jaar Periode Type
BSc Informatiekunde 2 semester II b verplicht