[Theory] Applications of NLP

[head-iie-vnr]

• [x] Examples
• [x] Components used

[head-iie-vnr]

Examples

Given text, it can answer any questions.

• Sentiment Analysis
• Chatbot
• Speech Recognition
• Machine Translations

Earlier Questions & answers had to be given earlier. Now based on the text, the Model can learn the semantics & does the needful for the user.

Example

• Money Robot speaking in Denvar airpor (Camera, Context Generation, Conversaion initiation using AI)
• Unlock door using Voice & Camera IOT (using private cloud & OpenStack)
• Home lifestyle with Smart Home Solutions (with IOT, generat information about: Rom usage, Fridge door open. Car sate, open doors, turn AC on, Speed limit advise, Give updaes on phone., Send notification when an item is out of stock, Turn off sprinkelrs when someone is playing, Visitor identification) .Generative AI can help to speak to Visitor parents.
• Case study on understanding Human Brain. It is already very effective, but did not get approved
• Camera that contantly check human mood in the campus (happy, yawning), shirts with tracker EEG is used (Electro Encephalo Gram) // Observation. Just puting the device made them to be alert.
• Looking at face check which direction they are looking at.

Idea: Teacher & Parent counceling.

[head-iie-vnr]

Sure! Let's break down each of these concepts in detail:

Ambiguity

Lexical Ambiguity

Lexical ambiguity occurs when a word has more than one meaning. The context in which the word is used is necessary to determine which meaning is intended.

• Example: The word "chip" can refer to:
  • A small piece of something (e.g., a wood chip).
  • A fried potato slice (e.g., a potato chip).
  • An electronic component (e.g., a microchip).

Syntactic Ambiguity

Syntactic ambiguity happens when a sentence can be parsed in more than one way due to its structure.

• Example: "Visiting relatives can cause problems."
  • This can mean that the act of visiting relatives can cause problems.
  • Or it can mean that relatives who are visiting can cause problems.

Referential Ambiguity

Referential ambiguity arises when it's unclear which entity a pronoun or a noun phrase refers to.

• Example: "I want to buy a drink for everyone here. I wanted to buy a newspaper."
  • "Everyone here" is ambiguous without context. Who is "everyone"? It needs clarification to resolve the ambiguity.

Tokenization

Tokenization is the process of breaking down text into individual units, such as words, phrases, or symbols. These units are called tokens.

• Example: Tokenizing the sentence "I love natural language processing!" would yield the tokens: ["I", "love", "natural", "language", "processing", "!"].

Normalization

Normalization involves transforming text into a standard format. This can include lowercasing, removing punctuation, and handling contractions.

• Example: The text "I'm learning NLP!" could be normalized to "i am learning nlp".

Stemming

Stemming reduces words to their base or root form, but it doesn't always result in a meaningful word.

• Example:
  • "Visible" -> "visbl"
  • "Democracy" -> "democr"
  • The goal is to strip suffixes to bring the word to its base form, which might not always be a valid word.

Lemmatization

Lemmatization reduces words to their base or root form but ensures that the base form is a meaningful word.

• Example:
  • "affection" -> "affect"
  • "democracy" -> "democratization"
  • This process uses vocabulary and morphological analysis to ensure the root word is valid.

Stopwords

Stopwords are common words that are often removed during text processing because they don't contribute much to the meaning of a sentence.

• Examples: "I", "me", "my", "the", "and", "is".
• Removing stopwords can help in focusing on the important words in a sentence.

Named Entity Recognition (NER)

NER is the process of identifying and classifying proper names in text into predefined categories such as names of people, organizations, locations, etc.

• Example: In the sentence "Barack Obama was the 44th President of the United States," NER would identify:
  • "Barack Obama" as a Person
  • "United States" as a Location

Chunking

Chunking, also known as shallow parsing, involves grouping adjacent tokens into larger chunks, often phrases, such as noun phrases (NP) or verb phrases (VP).

• Example: For the sentence "The quick brown fox jumps over the lazy dog":
  • NP: "The quick brown fox"
  • VP: "jumps over"
  • NP: "the lazy dog"

Chunking helps in identifying and extracting meaningful phrases from a sentence, which can be useful for various NLP tasks.

[head-iie-vnr]

Next: Activity

Text to Image.

[head-iie-vnr]

There are many tools for text to image.

• Can be useful when google search is not giving quality results
• Can help during PPT prep, directly based on PPT content.

Test to Image Tools

• midjourney Provides 7 days free trial. Can take input (text/image)
• https://copilot.microsoft.com/ > Designer. Gives text & images. Good for codelabs.

Tools

https://ai.invideo.io/onboard app.videogen.io

[head-iie-vnr]

The voice will be generated by machine.

[head-iie-vnr]

Voice Cloning

Few use cases

• Milk boiling detection (CNN for image detection)
• Hand patting. (MIT Inventor Android application builder,

Home Work

• Explore MIT Inventor.

IOT : Intelligence through high data generation automation: can happen without IOT.