IKH

Inspiration for Semantic Processing

In the previous session ,you learnt that grammar was not neccessary for understanding natural language.In the next video, let’s take a look at some more examples to understand why semantics is important.

From a technical point of view, grammar is a set of the rule that need to followed while constructing a sentence. However, several sentences are bound by the rules of grammar yet are gibberish. For example, take a look at the following sentence by Norm Chomsky who is an American linguist.

-“Colorless green ideas sleep furiusly-“

This sentence is structured correctly from a grammatical point of view. The noun–(ideas–) is described by an adjective(green) and followed by a verb (sleep). Yet, the meaning of the sentence is ambiguous.

Additionally, since language is an evolving concept, grammar keeps changing to match the needs of the era. Most of the communication done nowadays is via text or social media. Here, a more relaxed and informal tone of writing is used yet, it is still packed with information. 

In both these cases, one needs to stop relying on syntactics and use semantic processing. Now that you know why semantics is important, in the next video, you will learn how to capture the meaning of words in vectors.

What do you do first when you come across a word you are unfamiliar with? 

The dictionary definitions are not quite straightforward. Understanding a definition refers to understanding all the words within the definition of the word. However, we do not rely on a dictionary every time we don’t understand the meaning of a word.

You understand the meaning of the word from the overall context of the surrounding words. For example, let us assume that one does not know the meaning of the word ‘credit”.

After reading the sentence ‘The money was credited to my bank account’, one can easily infer that the word ‘credit’ is related to the exchange of currency. The words ‘money’ and ‘account’ set a context to the sentence that implies the predicted meaning. Through intelligent predictions such as this one, the meaning of words in a sentence becomes quite intuitive.

Hence, it was rightly said by the English linguist John Firth in 1957 –

 “You shall know a word by the company it keeps.”

Distributional semantics creates word vectors such that the word’s meaning is captured from its context.

In the next segment, you will learn about the intuition of how you can capture the meaning of words using geometry.