22
Implementing the Model of Natural Language Processing based Question Answering R.Sudharshan, M.Mohammed Feroz Khan, Final year BTECH IT , KTVR KPET M.Sabeena Ferzana ME CSE I year , Karpagam university

Implementing the Model of Natural Language Processing based Question Answering

Embed Size (px)

Citation preview

Page 1: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 1/22

Implementing the Model of Natural Language

Processing based Question Answering

R.Sudharshan,

M.Mohammed Feroz Khan,

Final year BTECH IT , KTVR KPET

M.Sabeena Ferzana

ME CSE I year , Karpagam university

Page 2: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 2/22

Introduction

Internet is a great resource for information ,

With a massive collection of information in theform of text, Users suffer delay in getting the

right information right time.

Page 3: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 3/22

People want to Questions

Examples from AltaVista query log

• who invented surf music?

• how to make stink bombs?

• where are the snow dens of yesteryear?

which English translation of the bible is used in official catholicliturgies?

• how to do clay art

• how tall is the sears tower?

Examples from Excite query log

where can i find information on puritan religion?• what are the 7 wonders of the world

• how can i eliminate stress

• What vacuum cleaner does Consumers Guide recommend

Page 4: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 4/22

What’s wrong in existing systems 

Shows Some blog

Completely

irrelevant

information

Page 5: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 5/22

Our Idea

Idea is to develop a software package to understand

your query and retrieve the exact answer / information

from the web.

If the answer is not present, suggest some relevant

information.

Page 6: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 6/22

So How it can be done

Artificial Intelligence

RoboticsNatural Language

ProcessingSearch

Information

Retrieval

Machine

TranslationLanguage Analysis

Semantics Parsing

Page 7: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 7/22

User asks

QuestionRewrite Query

Search

Knowledge

Base

Collect Summaries

Mine n-Grams

Filter n-Grams

Select best

answers

Concept

Page 8: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 8/22

Understanding of a Question

Specifying the computer what we really mean

by the question is important.

• Feature extraction from the questions

• Determining the type of the answer (Whether

it is a named entity or a passage)

Page 9: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 9/22

Query Classifications

Type Example Answer type

HowDo How did Steve Jobs die? A fact

HowFar How far away is the moon? Numerical

HowLong How long do hermit crabs live? Numerical

HowMany How many films Ajith acted? Numerical

HowManyTimes

How many times Sachin scored a century in WankedeStadium?

Numerical

How much How much calorie should a male adult need? Numerical

WhEquiv What language is mostly spoken in India? Factual Entity

WhPrep What is Rahman famous for? A fact

WhRole What state has the most Immigrants? Factual entity

Page 10: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 10/22

Query Classifications

What What was the name of the first Russian astronautto do a spacewalk?What is thumbrule?Name a flying mammal

A name , FactualEntity, Fact

WhatMeas What type of bridge is the Golden Gate Bridge?What kind of animal was Urang utan?

WhatRole What does laser stand for? Factual Text

WhatTime What year did Goa become a state? Date

When When did Tsunami happened in Indian Ocean ?When was Pincode introduced? 

Date

Where Where is Belize located? Factual Entity 

Who Who is the captain of Indian Hockey team? Factual Entity 

WhoRole Who invented the electric guitar? Factual Entity 

Why Why can't ostriches fly? Fact 

Page 11: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 11/22

Query Rewriting

Category specific transformation rules

“Where is Tajmahal located” 

Syntactical Rewrite

•Where Tajmahal is located

•Where Tajmahal located is

Semantic Rewrite

•Where is Tajmahal situated

•Where is Tajmahal placed

•Where is Tajmahal sited

Page 12: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 12/22

Semantic Lexical Analysis

• Wordnet – a semantic lexical thesaurus

• The semantic knowledge, which is obtained

from Word net, allows us to complete the

named entity lexicons in order to improve the

choice of the expected answer type.

• Useful in Named Entity Extraction

Page 13: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 13/22

Information Retrieval

• Rewrites are searched in knowledge base

usually a search engine or private textual

database

HTML , Documents ,

XML ,RSSPlain text Sentence Splitter

Page 14: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 14/22

Part of Speech tagging

[ The/DT market/NN ] for/IN 

[ system-management/NN software/NN ]

For/IN [ Digital/NNP ] [ ’s/POS hardware/NN ]is/VBZ fragmented/JJ enough/RB that/IN [

a/DT giant/NN ] such/JJ as/IN [

Computer/NNP Associates/NNPS ] should/MD

do/VB well/RB there/RB ./.

Page 15: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 15/22

Page 16: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 16/22

Page 17: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 17/22

Applications

• Question answering

• Intelligent Interactive Voice Responses

Voice responsive Intelligent Applications• Intelligent Search Engines

• Natural Language interfaces to Databases ,

RSS Feeds

Page 18: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 18/22

Who invented Zero

Ask

Zero was invented byThe Indian mathematician and astronomer Brahmagupta (598-668 AD). 

Source : http://www.math.sfu.ca/histmath/India/7thCenturyAD/brahmagupta.html 

Answers

Page 19: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 19/22

Where was bodhidharma born

Ask

Bodhidharma - born in Kancheepuram the Pallava capital in 440 AD.

Answers

Page 20: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 20/22

Conclusion

• Natural Language Processing and Information

retrieval is an exciting research area in the

field of Artificial Intelligence

• Intelligent answering capability will be an

amazing thing that could happen in near

future.

Page 21: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 21/22

References

[1] P. J. Hancox, A brief history of Natural Language Processing,http://www.cs.bham.ac.uk/~pjh/sem1a5/pt1/pt1_history.html.

[2] Fellbaum C., WordNet: An Electronic Lexical Database, MIT Press,Cambridge, MA, 1998.

[3] Androutsopoulos, G. D. Ritchie, P. Thanisch, “Natural language

Interfaces to Databases: An Introduction”, Journal of NaturalLanguage Engineering, 1995.

[4] Graesser A.C, Johnson B.K., Question Asking and Answering ,Handbook of Psycholinguistics,M.A. Gernsbcher (Eds), AcademicPress, 1994, p. 517-538.

[5] Schmid H., « Improvments in Part-of-Speech Tagging with anApplication To German », dans Natural Language Processing UsingVery Large Corpora, S. Armstrong, K.W. Chuch, P. Isabelle, E.Tzoukermann et D. Yarowski (Eds), Kluwer Academic Publisher,Dordrecht, 1999.

Page 22: Implementing the Model of Natural Language Processing based Question Answering

7/30/2019 Implementing the Model of Natural Language Processing based Question Answering

http://slidepdf.com/reader/full/implementing-the-model-of-natural-language-processing-based-question-answering 22/22

Thank You

Have the courage to follow your heart and intuitionEverything else is secondary

-Steve Jobs