4. Intro to NLP
How Non-Natural Beings(Computers)
process
Natural Language!
5. Who’s Me?
- Currently Sr. Data Scientist at Fountain9 (YC’21)
- AI Research enthusiast in the NLP and Time Series
Forecasting domain collaborating with academia.
- Worked at
- Hindustan Coca-Cola Beverage Pvt Ltd,
- then as Solutions Consultant at various co.s including Browserstack
(at the concurrence of technology and business)
- PG Specialisation in Applied DS (Michigan Uni)
B.E. Mechanical Engineering from M.U. (Distinction)
Why Me ?
- 1 PGS, 1 NanoDegree, few internships, some good projects,
50+ paper reviews and 100+ Linkedin courses (because, why
not?)
6. What is NLP ?
Natural Language Processing is when we program computers to process and analyze large
amounts of natural language data.
What’s natural language data?
Data that we generate on a daily basis by talking, chatting, calling, emailing, commenting……
Limitless data!
NLP is understanding and interpreting this data in a way that a literate human would.
7. Why NLP?
Why use Natural Language Processing?
- Understand this large amount of data
- Use this large amount of data
- Employ this data to Make models that understand
- Our speech , language, intent, emotion
- And many more endless reasons……
8. How NLP helps?
- Understand this large amount of data
- Its vibe >> good / bad / neutral >> Sentiment Analysis
- Its meaning
- In simpler words >> Summary
- In different words >> Translation
- Its relation to each other >> Topic Modelling
9. How NLP Helps?
Eg. 1. Sentiment Scoring
(of a paragraph from a blog on AI )
2. Topic Modelling (of the same blog)
['0.019*"ai" + 0.011*"human" + 0.011*"machine" +
0.007*"intelligence" + 0.007*"learning"']
10. HOW NLP HELPS?
- Use this large amount of data
- To derive insights
- Regarding opinions >> Discourse analysis
- Regarding objects >> Entity Recognition
- To find answers >> Question Answering
https://course.spacy.io/en/chapter1
Img credits : thecleverprogrammer.com
11. HOW NLP HELPS?
To build models that
- Understand our commands >> Speech recognition
- Assist us >> Speech to text
- Talk to us >> Text to speech (or basically anything to speech)
12. How it works?
Preliminary Steps :
- Text Mining
- Frequency, Length, Uniques, Screening for X/Y
- Text Analytics
- Patterns, Structures, Combinations, Network Graphs