Erweiterte Suche | Seiten im Überblick

Semantic relations in knowledge repositories and in texts

Kursbeschreibung

Dozenten/-innen	Vivi Nastase
Veranstaltungsart	Proseminar
Erster Termin	26.04.2017
Zeit und Ort	Mi. 14:15–15:45 INF 325 SR 24

Leistungsnachweis

• 20% questions and participation
• 40% presentation
• 40% project

Inhalt

Semantic relations describe how things interact or how they are related. We can find such relations in ontologies that capture knowledge about the world and how the objects in the world are related, and in texts where entities and their relations and interactions are mentioned. We need both these sources of semantic relations -- we can use texts to populate ontologies, and we can use ontologies to make sense of new texts.

The course will focus mostly on semantic relations between nominals, and we'll explore these relations from both a theoretical and a practical perspective. The theoretical exploration sketches the historical path which has brought us to the contemporary view and interpretation of semantic relations. We will discuss a wide range of relation inventories proposed by linguists and by language processing people. Such inventories vary by domain, granularity and suitability for downstream applications.

On the practical side, we will investigate the recognition and acquisition of relations from texts. For supervised learning methods, we will review some of the available datasets, the variety of features which can describe relation instances, and learning algorithms found appropriate for the task. Next, we'll look into weakly supervised and unsupervised learning methods of acquiring relations from large corpora with little or no previously annotated data. We will see how enduring the bootstrapping algorithm based on seed examples or patterns has proved to be, and how it has been adapted to tackle Web-scale text collections. We also see a few machine learning techniques which can perform fast and reliable relation extraction by taking advantage of data redundancy and variability.

We will look at other semantic relations that serve to organize knowledge -- such as entailment relations -- and how they fit into the semantic relation landscape.

Kursübersicht

Seminarplan

Datum	Sitzung	Materialien
26.04.2017	Introduction and historical overview	Part I
3.05.2017	Semantic relations between nominals, semantic relations between concepts   List of papers for presentations	Part II Read more about this in the book chapters
10.05.2017	Supervised learning of semantic relations: entity and relation features, data, approaches   Data and other project details for the course project	Part III Read more about this in the book chapter
17.05.2017	Unsupervised learning of semantic relations	Part IV Read more about this in the book chapter
24.05.2017	Deep learning for semantic relations	Part V
7.06.2017	[Caroline] Relation classification via convolutional deep neural network Zeng et al., COLING 2014	presentation
14.06.2017	[Theresa] Open language learning for information extraction Mausam et al., EMNLP 2012 [additional reading] Identifying relations for open information extraction Fader et al., EMNLP 2011	presentation
21.06.2017	[Andreas] A Dependency-Based Neural Network for Relation Classification Liu et al., ACL 2015	presentation
28.06.2017	[Calvin] Learning word representations by jointly modeling syntagmatic and paradigmatic relations Sun et al., ACL 2015 [additional presentation -- Vivi] Semantic relations in continuous low-dimensional vector spaces	presentation   slides
5.07.2017	[Janos] Coupled semi-supervised learning for information extraction Carlson et al., WSDM 2010	presentation
12.07.2017	Projects -- information here
17.07.2017 (MONDAY!)	Test data released for the class projects
19.07.2017 (the usual Wednesday)	Projects due!
26.07.2017	Project presentations (10 minutes each) and awards ceremony!

» weitere Kursmaterialien