Bilder vom Neuenheimer Feld, Heidelberg und der Universität Heidelberg

Lehrveranstaltungen
heiCO
Ressourcen	Fachschaft
Studien-FAQ	Technik-FAQ

Knowledge Graphs

Kursbeschreibung

Studiengang	Modulkürzel	Leistungs- bewertung
BA-2010[100%\|75%]	CS-CL	6 LP
BA-2010[50%]	BS-CL	6 LP
BA-2010[25%]	BS-CL	4 LP
BA-2010	AS-FL	8 LP
Master	SS-CL, SS-FAL	8LP

Dozenten/-innen	Vivi Nastase
Veranstaltungsart	PS/HS
Erster Termin	26.10.2017
Zeit und Ort	Do, 09:15–10:45, INF 306 / SR 14 (SR)
Commitment-Frist	tbd.

Teilnahmevoraussetzungen

Formale Semantik
Neural Methods in NLP

Leistungsnachweis

Course participation, one paper presentation and leading the discussion, final project.

Inhalt

Knowledge graphs are a powerful and versatile representation of lexical and world knowledge. They are a very special ecosystem of entities/concepts and relations, that define each other and that can be explored and exploited jointly to help with varied NLP tasks, such as question answering, summarization, reasoning.

In this course we will explore the issues of building knowledge graphs, representing them in various ways, learning within knowledge graphs, and learning to further grow them with or without external information.

Kursübersicht

Seminarplan

Datum	Sitzung	Materialien
Oct 26th	Introduction	Course admin Introduction to KGs (we are using the slides from the AAAI 2017 tutorial on Knowledg Graphs)
Nov 2nd	Knowledge Graphs: the nodes, a.k.a. concepts and entities	KG: nodes NLP for KGs A list of papers for presentations Information about the course project
Nov 9nd	Knowledge Graphs: the edges, a.k.a. the relations	IE for KGs (extracting relations)
Nov 16th	Special lecture on building a KG from Wikipedia	A multifaceted KG from Wikipedia
Nov 23th	Knowledge Graph construction (I)	Probabilistic methods for growing a KG
Nov 30th	Knowledge Graph construction (II)	Representations of KGs in low-dimensional continuous vectors spaces (KG embeddings)
Dec 7th	Evaluating Knowledge Graphs and conclusions
Dec 14th	(student paper presentations begin) [Miriam] Salience Rank: Efficient Keyphrase Extraction with Topic Modeling Teneva and Cheng, ACL 2017 [Lilitta]Multi-Task Learning of Keyphrase Boundary Classification Augenstein and Søgaard, ACL 2017 [Raphael] World Knowledge for Reading Comprehension: Rare Entity Prediction with Hierarchical LSTMs Using External Descriptions Long et al., EMNLP 2017
Jan 11th	(student paper presentations) [Mohammad] PositionRank: An Unsupervised Approach to Keyphrase Extraction from Scholarly Documents Florescu and Caragea, ACL 2017 [Juri] DeepPath: A Reinforcement Learning Method for Knowledge Graph Reasoning Xiong et al., EMNLP 2017
Jan 18th	(student paper presentations) [Muyan] Coupled semi-supervised learning for information extraction Carlson et al., WSDM 2010 [Leo] Combining distant and partial supervision for relation extraction Angeli et al., EMNLP 2014
Jan 25th	(student paper presentations) [Stoyan] Representing text for joint embedding of text and knowledge bases Toutanova et al., EMNLP 2015 [Robert] Knowledge vault: a web-scale approach to probabilistic knowledge fusionDong et al., KDD 2014
Feb 1st	(student paper presentations) [Siting] Task-oriented learning of word embeddings for semantic relation classification Hashimoto et al., CoNLL 2015
Feb 8th	(student paper presentations) [Tatjana] Modeling relations and their mentions without labeled text Riedel et al., ECML-PKDD 2010 [Nadja] Canonicalizing Open Knowledge Bases Galárraga et al., CIKM 2014

Literatur

To be posted as the course goes along.