Discover published sets by community

A subtask of information extraction that seeks to locate and classify named entities mentioned in text into predefined categories such as the names of persons, organizations, locations, expressions of times, quantities, monetary values, percentages, etc.

Also known as shallow parsing, it's a process of extracting phrases from unstructured text, which is then used as an input for NER to identify named entities.

A common tagging format for marking up text chunks where 'I' stands for Inside, 'O' for Outside, and 'B' for Beginning of a named entity. The tags are used for representing entities in a tokenized text.

Short for 'Conference on Natural Language Learning', it's known for its NER challenges which have contributed to the progression of NER methods by providing a common framework for evaluation.

A type of discriminative probabilistic model often used for structured prediction in NER, which takes into account the context within which entities appear to improve recognition accuracy.

A list or a dictionary of named entities used as a reference to aid in entity recognition. It's a lookup table that can be used to improve NER by providing external knowledge.

The task of resolving the ambiguity of entities in text, determining which 'entity' out of several possible candidates the term refers to, often using context or external databases.

A type of recurrent neural network (RNN) that can capture context from both past and future data points within a sequence, widely used in NER for its effectiveness in context-aware entity prediction.

The process of using domain knowledge to extract features from raw data that will be used to train machine learning models, especially crucial for NER to determine relevant attributes of named entities.

A machine learning method where a model developed for one task is reused as the starting point for a different but related task. In NER, it allows leveraging pre-trained models to improve entity recognition performance.

A transformer-based model designed to pre-train deep bidirectional representations by jointly conditioning on both left and right context in all layers, which has shown significant improvements in NER tasks.

Occurs when an entity is incorrectly identified or classified by a NER system, indicating that the model recognized an entity that is not actually present in the text.

Occurs when a NER system fails to identify or classify an actual entity in the text, which should have been recognized by the model.

Precision measures the correctness of the entities that the NER system has identified while Recall measures the system's ability to identify all relevant instances of entities in the text. Both metrics are used to evaluate NER performance.