On "Aboutness"

We’ve come across this concept a while ago at Bitext. What’s aboutness? According to the Wikipedia, aboutness is “simply meaning” or, in other words, aboutness is knowing what a piece of text is about, its content.

We wonder quite often every day what’s the aboutness of a piece of text. It is one of the tasks that we perform more often every day in the knowledge economy.

It is also one of the most time-consuming tasks (reading takes time and there’s a lot to read out there). This is what we do quite often when we search; for example, when we search “oil spill” we are saying: tell me which documents are about the “oil spill”, probably intending the recent oil spill connected to BP Plc and the USA, for example.

Having this task performed automatically, with machines and software, is certainly a business target in many areas like business intelligence, marketing and SEO, etc.

The interesting part is that extracting aboutness automatically has been the mission of Bitext since we started. Extracting aboutness can take many different forms in practice, involving things like:

-what’s the topic or topics of a piece of news (global warming, oil spill, stock prices),
-who’s involved in this story, be them people (Barack Obama, George Soros), companies (Apple, BP Plc, Amazon), etc.
-how are topics, people, companies, products, institutions connected, i.e., who did what to whom how where when, etc.
-and so on and so forth.

From this point of view, Bitext has developed the technology that discloses and extracts aboutness in different business scenarios: search, opinion mining, news categorization…

We do like the concept, it covers quite comprehensively the kind of problems we address at Bitext – The Aboutness Company!


Subscribe Here!