Scribd
Upload
87 views

Natural Language Processing

The document discusses natural language processing (NLP), including its history and how it works. It provides examples of NLP applications and discusses challenges in NLP like segmentation, tokenization, and part-of-speech tagging. The document also covers drawbacks of NLP and areas for improvement.

Uploaded by

Zain Tech Tips
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
87 views

Natural Language Processing

The document discusses natural language processing (NLP), including its history and how it works. It provides examples of NLP applications and discusses challenges in NLP like segmentation, tokenization, and part-of-speech tagging. The document also covers drawbacks of NLP and areas for improvement.

Uploaded by

Zain Tech Tips
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 21

Team Members Course:

1- Ali Zain Intro to Computing


2- Hassan Islam Cs-101
3- Umair Ahmad
4- Imtisal Hassan BS Information Technology
Quaid-i-Azam University

Natural Language Processing 1


Table of Content
1- What is NLP?
2- How does it work?
3- Examples
3- Practical facts about NLP.
4- Challenges in NLP
5- Drawbacks and Improvements.
6- Tools for NLP.
7- Applications.
8- References
NLP Intro
Almost 6500 languages are spoken in the world.
These are the way to communicate with humans.

Just like these, Humans want to


communicate with Robots or Machines,
& they also want Machines to communicate
with each other.

3
In short :

NLP is the way to narrow the


gap B/W Humans and Robots. 4
History

NLP has its roots in 1950s.

5
How does NLP Work?

Input Speech Recognition NLU NLG Output

Voice/Text Recognizing the Understanding Generating Ans Common


Language in Binary In Binary Language

6
NLP Examples

1. Digital Assistants
 Apple Siri
 Google Assistant
 MS Cortona
 Amazon Alexa
 Samsung Bixby

GOOGLE Microsoft Apple 7


NLP Examples

2- Speech Reorganization
 Voice Typing
 Predictive Text Generation

3- Sentimental Analysis
 Ai Chatting >> Replika, Waysa
 Chat Bots >> Eliza, Parry, etc
 Social Media >> Reactions on posts
4- Advertisements
 Google, FB, show ads by looking at your
search behavior.
8
NLP Examples

5- Spelling Checking
 Auto Correction of words while typing >> Grammarly

6- Machine Translation
 Google Translator

7- Automatic Text Generation


 Copy.ai (Website)
 Copyhat (App)
8- QNA
 Answering to questions directly
9
Practical Facts

1- NLP is a part of Ai
2- NLP focuses on language
3- NLP helps to reduce language barriers
4- NLP is far better than Keyboard Searching
5- NLP is used to know behavior
6- Chat bots are examples of NLP

10
Challenges
Various Challenges in NLP
Robot is like a baby. We have to teach it
everything like a baby to make it intelligent,
fast and accurate.

As NLP is regarding Language, So our main


purpose is to teach it Languages.

11
Challenges in NLP
1- Segmentation
Separation of sentences in a paragraph for better understanding.
2- Tokenizing
Separation of words in a sentence for better understanding.
3- Stop Words
Removal of Stop words (are, and, the, or) for better understanding.
4- Stemming
Giving Machine relevant and same words to understand >> Walk, walking, walked.
5- Lemmatization
Understanding Mood, Behavior, Gender etc
12
Challenges in NLP
6- POS Tags
Understanding Speech >> Parts of Speech
7- Named Entity Tagging
Telling Machine about famous things,
e.g : famous personalities, Places, Things etc
Google’s CEO Sundar Pichai Introduced Pixel 6 in New York.

13
Drawbacks

1- It is very much hard to make a model with

the specifications mention above, if there is no


any pre-trained model.

2- The system may not be able to ans Correct.

3- There can be Ambiguity. (The ability to


understand in more than 1 way.)

4- Low Resource languages are not available.

5- Errors in Text or speech.

14
Drawbacks / Improvements
Let’s Try Google Assistant

What is my Age?
Do you know my Age?

15
Data on Internet

Improvements
Almost 80% of the data is unstructured on the
internet. And only 20-21% is structured.

1- If NLP is given more data, it can be improved.


2- Information Overloaded

Structured data Unstructured Data

16
Tools for NLP
NLP is related to Ai, And Ai is a Data Science field
And in Data Science Two Languages are Most Famous.

R & PYTHON
17
Related Fields

Machine Learning
Artificial Intelligence
Robotics
Neural Networks

18
Conclusion

Ai, ML, Robotics and many other related fields are nothing if NLP is not present in them.

NLP is the most important factor in Ai. Because it understands languages and then reply with

Intelligence inserted in it.

If there is no NLP, Means There is no Ai, Means there are no Ai Applications,

19
References:
1. YouTube 2. Websites
 Gate Smashers  Wikipedia
 Simplilearn
 IBM Technology
 Edurekha

20
21

You might also like