Lecture

Information Extraction & Knowledge Inference

Graph Chat

Chat with Graph Search

Ask any question about EPFL courses, lectures, exercises, research, news, etc. or try the example questions below.

DISCLAIMER: The Graph Chatbot is not programmed to provide explicit or categorical answers to your questions. Rather, it transforms your questions into API requests that are distributed across the various IT services officially administered by EPFL. Its purpose is solely to collect and recommend relevant references to content that you can explore to help you answer your questions.

Description

This lecture covers the concepts of information extraction and knowledge inference. Information extraction involves creating a matrix with entity pairs as rows and relation types as columns, extracting relations from text patterns and knowledge bases. Knowledge inference focuses on surface patterns, Bayesian personalized ranking, and relation embeddings. It also delves into taxonomy induction, entity disambiguation, and the discovery of concepts and terms. The instructor discusses the challenges of entity linking, homonyms, and synonyms, as well as the use of entity graphs and local information for interpreting mentions in a knowledge base.

Official source

Ontological neighbourhood

Information engineering

Natural language processing: Topics in natural language processing

Related concepts (32)

Binary relation

In mathematics, a binary relation associates elements of one set, called the domain, with elements of another set, called the codomain. A binary relation over sets X and Y is a new set of ordered pairs (x, y) consisting of elements x in X and y in Y. It is a generalization of the more widely understood idea of a unary function. It encodes the common concept of relation: an element x is related to an element y, if and only if the pair (x, y) belongs to the set of ordered pairs that defines the binary relation.

Entity linking

In natural language processing, entity linking, also referred to as named-entity linking (NEL), named-entity disambiguation (NED), named-entity recognition and disambiguation (NERD) or named-entity normalization (NEN) is the task of assigning a unique identity to entities (such as famous individuals, locations, or companies) mentioned in text. For example, given the sentence "Paris is the capital of France", the idea is to determine that "Paris" refers to the city of Paris and not to Paris Hilton or any other entity that could be referred to as "Paris".

Homogeneous relation

In mathematics, a homogeneous relation (also called endorelation) on a set X is a binary relation between X and itself, i.e. it is a subset of the Cartesian product X × X. This is commonly phrased as "a relation on X" or "a (binary) relation over X". An example of a homogeneous relation is the relation of kinship, where the relation is between people. Common types of endorelations include orders, graphs, and equivalences. Specialized studies of order theory and graph theory have developed understanding of endorelations.

Converse relation

In mathematics, the converse relation, or transpose, of a binary relation is the relation that occurs when the order of the elements is switched in the relation. For example, the converse of the relation 'child of' is the relation 'parent of'. In formal terms, if and are sets and is a relation from to then is the relation defined so that if and only if In set-builder notation, The notation is analogous with that for an inverse function. Although many functions do not have an inverse, every relation does have a unique converse.

Knowledge acquisition

Knowledge acquisition is the process used to define the rules and ontologies required for a knowledge-based system. The phrase was first used in conjunction with expert systems to describe the initial tasks associated with developing an expert system, namely finding and interviewing domain experts and capturing their knowledge via rules, objects, and frame-based ontologies. Expert systems were one of the first successful applications of artificial intelligence technology to real world business problems.

Related lectures (62)

Information Extraction: Methods and Applications

Explores methods for information extraction, including traditional and embedding-based approaches, supervised learning, distant supervision, and taxonomy induction.

Knowledge Inference: Entity Disambiguation and Graph Embeddings

Explores entity disambiguation, graph embeddings, scoring functions, and learning methods.

Entity Disambiguation

Explores Entity Disambiguation, linking text mentions to a knowledge base, coherence in entity graphs, and Personalized PageRank.

Link Prediction: Knowledge Inference

Discusses link prediction in knowledge graphs and models like TransE.

Matrix Factorization: Information Extraction