text summarization: applications
this article is mainly a summarization of Yasemin Altun’s presentation in May 2014 on how google applies text summarization.
text summarization is highly related to google knowledge graph project：
entities description within red circle use text summarization from wiki to give a one sentence description of the entity.
She mentioned google then mainly focus on Entity-centric summarization， describe the entities through news-worthy events.
- Identify most relevant information with respect to a given entity.
- (Annotate the content with references to entities)
- Multi-document summarization to leverage redundancy
- a. Extractive (selecting sentences from the documents)
- b. Abstractive (generate the sentences)
- Beyond academic definition: For machine or user consumption
text summarization applies to five categories in google：
- Entity timelines
- Storylines of events
- Sentence compression
- Event understanding
- Summarization of user-generated content
Given an entity and a time period, provide a summary of the most memorable events involving this entity.
describe the entity through the time dimension.
Storylines of events
Identify and summarize events that leads to the event of interest
Given a sentence, generate a shorter one while preserving the essential information.
Learn how events are referred to in text and represent event mentioning text in predicate argument structure.
events description are subjective, event understanding is used to keep the neutrality of event description and represents them in predicate arguments formats.
Summarization of user-generated content
Identify main aspects and summarize for machine or user consumption.
summarize youtube’s user comments on the video.