Page 23 - CT_AI_Class-7

ethical minds

NLP systems can pick up biases from the data they are trained on. If the data is biased, the system
might give unfair or harmful results.

How Machines Understand Human Language?

Machines do not naturally understand human language like humans do, but they can be trained

to interpret it using the following steps:
1. Data pre-processing (Cleaning): Raw text is collected and fact bits
cleaned by removing errors, unnecessary symbols and Google Translate was first
irrelevant data to make it suitable for analysis. released in April 2006 and

2. Syntax/syntactic analysis: The structure of sentences originally used Statistical

is analysed to understand grammar, word order and Machine Translation (SMT)
relationships between words. before switching to AI-powered
neural translation (NMT) in
3. Semantics (Meaning analysis): The system interprets the 2016 for better accuracy.

meaning of words and sentences, considering context to
understand what is being communicated.

4. Deep learning: Advanced algorithms learn from large
amounts of data to recognise patterns, improve understanding and generate more accurate
responses over time.

Applications of NLP

NLP is used in many real-life applications, such as:

 Voice assistants: NLP helps voice assistants like Google Assistant
and Amazon Alexa understand spoken commands and respond to
users in a natural way.

 Language translation: NLP enables tools like Google Translate to
convert text or speech from one language to another quickly and
accurately.

 Customer support: NLP is used in chatbots and virtual assistants on
websites and apps to answer customer queries, provide information
and offer support without human involvement.

AI Domains and Applications 21