Natural Language Processing (NLP) 10/30/02 CS470/670 NLP (10/30/02) 1

Similar documents
Question Answering. CS486 / 686 University of Waterloo Lecture 23: April 1 st, CS486/686 Slides (c) 2014 P. Poupart 1

Past Involvement of IHH in Supporting the Global Jihad and Radical Islam - Additional Information 1

Islamic Militarism and Terrorism in the Modern World. Roots of Hate

«Violent Islamist Extremism : The European Experience» Committee on Homeland Security and Government Affairs U.S. Senate Washington, June 27, 2007

War on Terrorism Notes

Pew Global Attitudes Project 2010 Spring Survey Topline Results Pakistan Report

Information Extraction. CS6200 Information Retrieval (and a sort of advertisement for NLP in the spring)

Macmillan/McGraw-Hill SCIENCE: A CLOSER LOOK 2011, Grade 3 Correlated with Common Core State Standards, Grade 3

What is al-qaeda? 9/11: Pre-Visit

CS 671 ICT For Development 19 th Sep 2008

Prentice Hall World Geography: Building A Global Perspective 2003 Correlated to: Colorado Model Content Standards for Geography (Grade 9-12)

Al Qaeda Financing and Conflict Diamonds A Sentinel TMS Analysis

Intelligence and Terrorism Information Center at the Center for Special Studies (C.S.S.)

KEEP THIS COPY FOR REPRODUCTION Pý:RPCS.15i )OCUMENTATION PAGE 0 ''.1-AC7..<Z C. in;2re PORT DATE JPOTTYPE AND DATES COVERID

***** [KST : Knowledge Sharing Technology]

9/11 BEFORE, DAY OF, AND AFTER WHAT HAPPENED AND WHY?

Universitas Saraviensis Project Seminar Text Mining for Historical Documents Antonia Scheidel February An Introduction To Ontologies

The Fallacy in Intelligent Design

Global View Assessments Fall 2013

Mapping to the CIDOC CRM Basic Overview. George Bruseker ICS-FORTH CIDOC 2017 Tblisi, Georgia 25/09/2017

Analysis of ISIS's Claims of Responsibility for Terrorist Attacks Carried Out Abroad. Overview 1

Identifying Anaphoric and Non- Anaphoric Noun Phrases to Improve Coreference Resolution

ANAPHORIC REFERENCE IN JUSTIN BIEBER S ALBUM BELIEVE ACOUSTIC

Intelligence and Terrorism Information Center

Saudi Arabia: Terror threat reduced for time being

Factsheet about 9/11. Page 1

ISTANBUL BLASTS--Two. Published on South Asia Analysis Group ( Submitted by asiaadmin2 on Mon, 09/24/ :14

Macmillan/McGraw-Hill SCIENCE: A CLOSER LOOK 2011, Grade 4 Correlated with Common Core State Standards, Grade 4

Anaphora Resolution in Hindi Language

Automatic Evaluation for Anaphora Resolution in SUPAR system 1

Issue Overview: Jihad

The Global Jihad System Unites Against Israel and the West. Threats to attack Israeli targets worldwide, as well as in the. United States and Europe 1

FOURTH GRADE. WE LIVE AS CHRISTIANS ~ Your child recognizes that the Holy Spirit gives us life and that the Holy Spirit gives us gifts.

Grade yourself on the OER. Test Friday on Unit 1

SOCIAL MEDIA AND RADICALIZATION

Visual Analytics Based Authorship Discrimination Using Gaussian Mixture Models and Self Organising Maps: Application on Quran and Hadith

Chapter 8: Political Geography KEY ISSUES #3 & #4

What Does the Enemy Want?

Extracting the Semantics of Understood-and- Pronounced of Qur anic Vocabularies Using a Text Mining Approach

The killing of two Al-Qaeda leaders in Iraq and its implications

Network-based. Visual Analysis of Tabular Data. Zhicheng Liu, Shamkant Navathe, John Stasko

Tools Andrew Black CS 305 1

Macmillan/McGraw-Hill SCIENCE: A CLOSER LOOK 2011, Grade 1 Correlated with Common Core State Standards, Grade 1

9/11. Before, The Day of, and After. Write a journal entry telling me 5 things that happened on 9/11. Label it Journal #1

Intelligence and Terrorism Information Center at the Israel Intelligence Heritage & Commemoration Center (IICC)

Security threat from Afghanistan: Under- or overrated?

Tuesdays 14:50-17:45 (and a few Fridays) Vestergade 10-A12

DOWNLOAD OR READ : THE LOGIC BOOK PDF EBOOK EPUB MOBI

Overview 1. On June 29, 2014, ISIS leader Abu Bakr al-baghdadi declared the establishment of the

Outline of today s lecture

Utah South Area Family History Training

AP Human Geography. Chapter 7 Guided Reading 2 nd Half

Preventing Nuclear Terrorism

Rise of the Muslim Brotherhood

OSS PROFILE NAME: ABDUL RASUL SAYYAF. COUNTRY: Afghanistan

Jihadist Brides, Victims of the West

IN THE UNITED STATES DISTRICT COURT FOR THE EASTERN DISTRICT OF VIRGINIA. Alexandria Division

ISLAM IN CAMBODIA: Resurgence or Extremism?

Intelligence and Terrorism Information Center at the Israel Intelligence Heritage & Commemoration Center

TED ANTALYA MODEL UNITED NATIONS 2019

Congressional Testimony

STATEMENT OF JARRET BRACHMAN BEFORE THE HOUSE ARMED SERVICES COMMITTEE SUBCOMMITTEE ON TERRORISM, UNCONVENTIONAL THREATS AND CAPABILITIES

College and Career Readiness Anchor Standards for Reading. Step Into the Time 36 Step Into the Place 92, 108, 174, 292, 430

Anaphora Resolution in Biomedical Literature: A

ON THE ROLE OF METHODOLOGY: ADVICE TO THE ADVISORS

Pastor Search Survey Text Analytics Results. An analysis of responses to the open-end questions

Prentice Hall United States History Survey Edition 2013

Global Affairs May 13, :00 GMT Print Text Size. Despite a rich body of work on the subject of militant Islam, there is a distinct lack of

He got what he deserved, say Canadians about bin Laden s death

Curriculum Evaluation Tool

Islam in other Nations

Keywords: Knowledge Organization. Discourse Community. Dimension of Knowledge. 1 What is epistemology in knowledge organization?

Universiti Teknologi MARA. Ontology of Social Interaction Ethics in Al Adab Al - Mufrad by Using Semantic Web

Reuse: a symbiosis between developers and researchers

Bledar Toska, University of Vlora, Albania. Ohrid, June 2017

Pearson myworld Geography Western Hemisphere 2011

The Intelligence Function. Issues in Crime and Justice CJ 4610 PA 5315 Professor James J. Drylie Week 2

Al-Qaeda warns of more attacks

Understanding Terror Networks. By Marc Sageman. Philadelphia: University of Pennsylvania Press, Pp ISBN

Hybrid Approach to Pronominal Anaphora Resolution in English Newspaper Text

Church Leader Survey. Source of Data

CHAPTER I INTRODUCTION. which words are related to other word of the same language. Formal differences

Running head: VISUAL EXPLORATION OF SEMANTIC MARKERS OF FAITH. Visual Exploration of the Semantic Markers of Faith. Author Note

Prentice Hall The American Nation: Beginnings Through 1877 '2002 Correlated to: Chandler USD Social Studies Textbook Evaluation Instrument (Grade 8)

PHILOSOPHY-PHIL (PHIL)

War in Afghanistan War in Iraq Arab Spring War in Syria North Korea 1950-

The Impact of Oath Writing Style on Stylometric Features and Machine Learning Classifiers

Functionalism and the Chinese Room. Minds as Programs

Name: Advisory: Period: Introduction to Muhammad & Islam Reading & Questions Monday, May 8

31/05/2013 Contact :

Real-time case study on links between development and humanitarian programming for Rohingya refugees in Cox s Bazaar, Bangladesh

The North African Franchise: AQIM s Threat to U.S. Security. Strategic Insights, Volume VIII, Issue 5 (December 2009) By Captain Russell J.

Stochastic Opponent Modeling Agents: A Case Study with Hamas

International experience. Local knowledge.

Carolina Bachenheimer-Schaefer, Thorsten Reibel, Jürgen Schilder & Ilija Zivadinovic Global Application and Solution Team

"Military action will bring great costs for the region," Rouhani said, and "it is necessary to apply all efforts to prevent it."

Anatomy of an Insurgency

Al-Qaeda in the Islamic Maghreb (AQIM)

Periodical Review: Summary of Information from. the Jihadist forums. This report summarizes the most prominent events brought up in the Jihadist

African Caucus Topic A: Combatting the Rise of Terrorism in Africa. Chairs: Mariana Araujo, Shalom Rubino

Transcription:

Natural Language Processing (NLP) 10/30/02 CS470/670 NLP (10/30/02) 1

NLP Definition a range of computational techniques CS470/670 NLP (10/30/02) 2

NLP Definition (cont d) a range of computational techniques for analyzing and representing naturally occurring texts CS470/670 NLP (10/30/02) 3

NLP Definition (cont d) a range of computational techniques for analyzing and representing naturally occurring texts at one or more levels of linguistic analysis CS470/670 NLP (10/30/02) 4

Levels of Language Understanding Pragmatic Discourse Semantic Syntactic Lexical Morphological CS470/670 NLP (10/30/02) 5

NLP Definition (cont d) a range of computational techniques for analyzing and representing naturally occurring texts at one or more levels of linguistic analysis for the purpose of achieving human-like language processing CS470/670 NLP (10/30/02) 6

NLP Definition (cont d) a range of computational techniques for analyzing and representing naturally occurring texts at one or more levels of linguistic analysis for the purpose of achieving human-like language processing for knowledge intensive applications CS470/670 NLP (10/30/02) 7

Goals of Information Extraction A robust information extraction system CS470/670 NLP (10/30/02) 8

Goals of Information Extraction A robust information extraction system Recognize concepts and the implicit relations amongst them CS470/670 NLP (10/30/02) 9

Goals of Information Extraction A robust information extraction system Recognize concepts and the implicit relations amongst them Convert vast amounts of textual data into a semantic representation CS470/670 NLP (10/30/02) 10

Goals of Information Extraction A robust information extraction system Recognize concepts and the implicit relations amongst them Convert vast amounts of textual data into a semantic representation Provide knowledge discovery tools for multiple analyst activities CS470/670 NLP (10/30/02) 11

Goals of Information Extraction A robust information extraction system Recognize concepts and the implicit relations amongst them Convert vast amounts of textual data into a semantic representation Provide knowledge discovery tools for multiple analyst activities visual exploration data-mining via NLP queries link analysis CS470/670 NLP (10/30/02) 12

High Level Task Description Evaluate the application of automatic knowledge extraction to link analysis CS470/670 NLP (10/30/02) 13

High Level Task Description Evaluate the application of automatic knowledge extraction to link analysis Specialization of generic relations Prototype IE to Link Analysis tool CS470/670 NLP (10/30/02) 14

High Level Task Description Evaluate the application of automatic knowledge extraction to link analysis Specialization of generic relations Prototype IE to Link Analysis tool Identify current technological barriers CS470/670 NLP (10/30/02) 15

High Level Task Description Evaluate the application of automatic knowledge extraction to link analysis Specialization of generic relations Prototype IE to Link Analysis tool Identify current technological barriers Establish high-payoff research directions CS470/670 NLP (10/30/02) 16

High Level Task Description Evaluate the application of automatic knowledge extraction to link analysis Specialization of generic relations Prototype IE to Link Analysis tool Identify current technological barriers Establish high-payoff research directions Produce substantive report on current state-of-theart CS470/670 NLP (10/30/02) 17

KNOW-IT Overview Automatically identifies and extracts concepts and relations involving people, events, places, and organizations, etc from massive volumes of digital textual data For purpose of building / adding to Knowledge Bases for use by human & automated reasoners General technology capability currently used for various text types & domains can be specialized for specific applications CS470/670 NLP (10/30/02) 18

KNOW-IT s Building Blocks: Natural Language Processing + Knowledge Extraction + Graphical Visualization CS470/670 NLP (10/30/02) 19

KNOW-IT components Concepts 60 + Proper Noun Categories CS470/670 NLP (10/30/02) 20

Proper Noun Categorization Scheme Geographic Affiliation Organization Human Document Equipment Scientific Temporal Misc. Entity City Port Airport Island County Province Country Continent Region Water Geo. Misc. Religion Nationality Company Company Type Government U.S. Government Organization Person Title Document Software Hardware Machines Disease Drugs Chemicals Date Time Misc. CS470/670 NLP (10/30/02) 21

KNOW-IT components Concepts 60 + Proper Noun Categories WordNet Synsets CS470/670 NLP (10/30/02) 22

CS470/670 NLP (10/30/02) 23

KNOW-IT components Concepts 60 + Proper Noun Categories WordNet Synsets Relations 40 + generic semantic relations CS470/670 NLP (10/30/02) 24

Semantic Relations Relations AGNT (act, animate) PART (entity-x, entity-y) PTIM (T, time) CAUS (state-x, state-y) PURP (act-x, act-y) (state/entity, act-y) Definition animate is performer (agent) of action entity-x has part entity-y T occurred at specific time x has a cause y act-x has purpose act-y state has purpose act-y CS470/670 NLP (10/30/02) 25

KNOW-IT components Concepts 60 + Proper Noun Categories WordNet Synsets Relations 40 + generic semantic relations Concept-Relation-Concept CS470/670 NLP (10/30/02) 26

Concept-Relation Extraction HEADLINE: Albanian suspected to have links to bin Laden arrested SOURCE: Agence France Presse, 01/10/99 Maksim Ciciku was arrested by the Albanian police in Tirana. Ciciku met Osama bin Laden in April 1994. CS470/670 NLP (10/30/02) 27

Concept-Relation Extraction HEADLINE: Albanian suspected to have links to bin Laden arrested SOURCE: Agence France Presse, 01/10/99 Maksim Ciciku was arrested by the Albanian police in Tirana. Ciciku met Osama bin Laden in April 1994. CG_1 OBJ ( arrest, Maksim Ciciku person ) AGNT ( arrest, Albanian police ) CHRC ( police, Albanian nationality ) LOC ( arrest, Tirana city ) CG_2 AGNT ( meet, Maksim Ciciku person ) OBJ ( meet, Osama bin Laden person ) PTIM ( meet, April 1994 ) CS470/670 NLP (10/30/02) 28

CS470/670 NLP (10/30/02) 29

Adapting KNOW-IT for Link Analysis Extraction in KNOW-IT is broad and shallow based on linguistic regularities not domain-dependent rules But the technology can be extended to narrow and deep applications for Link Analysis terrorism domain for HPKB CS470/670 NLP (10/30/02) 30

Specialization Methodology Map 2 or more general C-R-C extraction rules into a more specific link rule, e.g. for SUPPORT: CS470/670 NLP (10/30/02) 31

Specialization Methodology Map 2 or more general C-R-C extraction rules into a more specific link rule, e.g. for SUPPORT: C1 -R-C2 + C2 -R-C3 CS470/670 NLP (10/30/02) 32

Specialization Methodology: Map 2 or more general C-R-C extraction rules into a more specific link rule, e.g. for SUPPORT: C1 - AGNT -C2 + C2 - OBJ - C3 CS470/670 NLP (10/30/02) 33

Specialization Methodology Map 2 or more general C-R-C extraction rules into a more specific link rule, e.g. for SUPPORT: C1 -AGNT -C2 + C2 - OBJ - C3 <international agent*> AGNT <support verb*> + <support verb*> OBJ <X 54> CS470/670 NLP (10/30/02) 34

Where, International agent = any Proper Noun whose category is an element of the set {7, 40, 411, 412, 413, 414, 415, 416, 417, 50, 501, 51, 52, 53, 54} AND Support verb = any element of the synsets containing verbs such as: {fund, back, support, aid, help, assist, sponsor, subsidize, patronize, cosponsor, bankroll, champion, defend} CS470/670 NLP (10/30/02) 35

Then,. extremist Harkatul Jihad group, reportedly backed by Saudi dissident Osama bin Laden... CS470/670 NLP (10/30/02) 36

Then,. extremist Harkatul Jihad group, reportedly backed by Saudi dissident Osama bin Laden... AGENT (back, Osama bin Laden person) OBJECT (back, Hartakul Jihad terrorist_group group) CS470/670 NLP (10/30/02) 37

Then,. extremist Harkatul Jihad group, reportedly backed by Saudi dissident Osama bin Laden... AGENT (back, Osama bin Laden person) OBJECT (back, Hartakul Jihad terrorist_group group) SUPPORT (Osama bin Laden person, Hartakul Jihad terrorist_group group) CS470/670 NLP (10/30/02) 38

03/14/1999 (AFP) Bangladesh bomb blast toll 10, opposition wants judicial probe the extremist Harkatul Jihad group, reportedly backed by Saudi dissident Osama bin Laden... CS470/670 NLP (10/30/02) 39

03/14/1999 (AFP) Bangladesh bomb blast toll 10, opposition wants judicial probe the extremist Harkatul Jihad group, reportedly backed by Saudi dissident Osama bin Laden... the DT extremist JJ Harkatul_Jihad NP 1 group NN,, reportedly RB backed VBD by IN Saudi NP 2 dissident IN Osama_bin_LadeNP 3 CS470/670 NLP (10/30/02) 40

03/14/1999 (AFP) Bangladesh bomb blast toll 10, opposition wants judicial probe the extremist Harkatul Jihad group, reportedly backed by Saudi dissident Osama bin Laden... the DT extremist JJ Harkatul_Jihad NP 1 group NN,, reportedly RB backed VBD by IN Saudi NP 2 dissident IN Osama_bin_LadeNP 3 <PN> 1 54 Harkatul Jihad 2 17 Saudi 3 30 Osama bin Laden </PN> CS470/670 NLP (10/30/02) 41

03/14/1999 (AFP) Bangladesh bomb blast toll 10, opposition wants judicial probe the extremist Harkatul Jihad group, reportedly backed by Saudi dissident Osama bin Laden... the DT extremist JJ Harkatul_Jihad NP 1 group NN,, reportedly RB backed VBD by IN Saudi NP 2 dissident IN Osama_bin_LadeNP 3 <PN> 1 54 Harkatul Jihad 2 17 Saudi 3 30 Osama bin Laden </PN> CG0: AGNT (back, Osama bin Laden person) OBJECT (back, Harkatul Jihad terrorist_group group) CHRC (Harkatul Jihad terrorist_group group, extremist) MANR (back, reportedly) ISA (Osama bin Laden person, Saudi nationality dissident) CS470/670 NLP (10/30/02) 42

03/14/1999 (AFP) Bangladesh bomb blast toll 10, opposition wants judicial probe the extremist Harkatul Jihad group, reportedly backed by Saudi dissident Osama bin Laden... the DT extremist JJ Harkatul_Jihad NP 1 group NN,, reportedly RB backed VBD by IN Saudi NP 2 dissident IN Osama_bin_LadeNP 3 <PN> 1 54 Harkatul Jihad 2 17 Saudi 3 30 Osama bin Laden </PN> CG0: AGNT (back, Osama bin Laden person) OBJECT (back, Harkatul Jihad terrorist_group group) CHRC (Harkatul Jihad terrorist_group group, extremist) MANR (back, reportedly) ISA (Osama bin Laden person, Saudi nationality dissident) CG0: AGNT (back, Osama bin Laden person) OBJECT (back, Harkatul Jihad terrorist_group group)... SUPPORT(Osama bin Laden person, Harkatul Jihad terrorist_group group) CS470/670 NLP (10/30/02) 43

03/14/1999 (AFP) Bangladesh bomb blast toll 10, opposition wants judicial probe the extremist Harkatul Jihad group, reportedly backed by Saudi dissident Osama bin Laden... support (Osama bin Laden person, Hartakul Jihad terrorist_group group) Osama bin Laden support Harkatul Jihad group CS470/670 NLP (10/30/02) 44

03/12/1999 (AFP) Bangladesh arrest Afghan war veteran over bomb attack from Monirul Hassan, a member of the Harkatul Jihad group who was reportedly trained by the Taliban militia in Afghanistan... CS470/670 NLP (10/30/02) 45

03/12/1999 (AFP) Bangladesh arrest Afghan war veteran over bomb attack from Monirul Hassan, a member of the Harkatul Jihad group who was reportedly trained by the Taliban militia in Afghanistan... CG0 ISA AFFL AGNT OBJ LOC (Monirul Hassan person, member) (member, Harkatul Jihad terrorist_group group) (train, Taliban militia organization) (train, Monirul Hassan person) (train, Afghanistan country) CS470/670 NLP (10/30/02) 46

03/12/1999 (AFP) Bangladesh arrest Afghan war veteran over bomb attack from Monirul Hassan, a member of the Harkatul Jihad group who was reportedly trained by the Taliban militia in Afghanistan... CG0 ISA (Monirul Hassan person, member) AFFL (member, Harkatul Jihad terrorist_group group) affiliate (Monirul Hassan person, Harkatul Jihad terrorist_group group) AGNT OBJ... prep (train, Taliban militia organization) (train, Monirul Hassan person) (Taliban militia organization, Monirul Hassan person) CS470/670 NLP (10/30/02) 47

03/12/1999 (AFP) Bangladesh arrest Afghan war veteran over bomb attack from Monirul Hassan, a member of the Harkatul Jihad group who was reportedly trained by the Taliban militia in Afghanistan... affiliate (Monirul Hassan person, Harkatul Jihad terrorist_group group) prep (Taliban militia organization, Monirul Hassan person) CS470/670 NLP (10/30/02) 48

03/12/1999 (AFP) Bangladesh arrest Afghan war veteran over bomb attack from Monirul Hassan, a member of the Harkatul Jihad group who was reportedly trained by the Taliban militia in Afghanistan... affiliate (Monirul Hassan person, Harkatul Jihad terrorist_group group) prep (Taliban militia organization, Monirul Hassan person) Taliban Militia Osama bin Laden prep support Monirul Hassan affiliate Harkatul Jihad group CS470/670 NLP (10/30/02) 49

03/08/1999 (AFP) 16 soldiers killed, 21 wounded in Algerian ambush The Salafist Group for Preaching and Combat (GSPC), led by Hassan Hattab, recently distributed Created at the instigation of bin Laden, the group is especially active... CS470/670 NLP (10/30/02) 50

03/08/1999 (AFP) 16 soldiers killed, 21 wounded in Algerian ambush The Salafist Group for Preaching and Combat (GSPC), led by Hassan Hattab, recently distributed Created at the instigation of bin Laden, the group is especially active... head (Hassan Hattab person, GSPC terrorist_group) support (Osama bin Laden person, GSPC terrorist_group) CS470/670 NLP (10/30/02) 51

03/08/1999 (AFP) 16 soldiers killed, 21 wounded in Algerian ambush The Salafist Group for Preaching and Combat (GSPC), led by Hassan Hattab, recently distributed Created at the instigation of bin Laden, the group is especially active... head (Hassan Hattab person, GSPC terrorist_group) support (Osama bin Laden person, GSPC terrorist_group) Taliban Militia Osama bin Laden Hassan Hattab prep support support head Monirul Hassan affiliate Harkatul Jihad group GSPC CS470/670 NLP (10/30/02) 52

02/15/1999 (AFP) Bin Laden held to be behind an armed Algerian Islamic movement Mohamed Berrachad had worked for Hattab, who is himself a dissident from the Armed Islamic Group (GIA) In his testimony, Berrachad said Bin Laden and Hattab communicated by satellite telephone and that he had heard their conversations, said to hinge on the discrediting of Antar Zouabri's GIA by its savage massacres of civilians... Antar Zouabri lead GIA disagree Mohamed Berrachad discredit discredit affiliate Taliban Militia Osama bin Laden Hassan Hattab prep support support head Monirul Hassan affiliate Harkatul Jihad group GSPC CS470/670 NLP (10/30/02) 53

CS470/670 NLP (10/30/02) 54

CS470/670 NLP (10/30/02) 55

CS470/670 NLP (10/30/02) 56

CS470/670 NLP (10/30/02) 57

CS470/670 NLP (10/30/02) 58

CS470/670 NLP (10/30/02) 59

As a link analyzer, KNOW-IT Assists analysts in appraising a potential crisis situation by determining the key players and the nature of their relations to one another Automatically filters, extracts, organizes, and analyzes textual intelligence data Generates and visualizes networks from relevant, unstructured text Allows analysts to specialize the links by easy-towrite specification & generalization rules Provides rich output to visualization tools CS470/670 NLP (10/30/02) 60