Named entity recognition (NER): Named entity recognition extracts proper names of persons, organizations, places, and so on.
Relation Extraction: Relation Extraction uncovers relationships between different entities mentioned in the text; for example, the relationship between a company and the location where the company is headquartered at.
Event Extraction: Like Relation Extraction, Event Extraction finds the relation between entities, but it also implies a change of state, such as who beat whom in the Australian Open.
Co-reference Resolution: Co-reference Resolution attempts to link different references in the text to the same entity. In the following example, Co-reference Resolution indicates that the pronoun ‘he’ refers to ‘President Obama’; ‘the president’ also refers to ‘President Obama’; but ‘it’ refers to ‘Russia.’
Sentiment analysis: Sentiment Analysis extracts opinion towards a particular entity or a particular aspect of an entity, along with the polarity of that sentiment (whether it is positive or negative).
Table Extraction: A lot of information on the web and in public records is expressed in what looks like tables, but may or may not be formatted as tables in HTML. Table Extraction associates apparent cells with table boundaries, titles, row and column headings, and so on.