Professional Documents
Culture Documents
Dr. Ranjani Parthasarathi Professor Dept. of Information Science & Technology CEG, Anna University, Chennai
1/25/2012
1/25/2012
A story
The following scenario illustrates a possible use of an opportunistic network deployed after an earthquake. One of its helpers, a surveillance system, looks at a public area scene with many objects. The image is passed to another helper that analyzes it, and recognizes one of the objects as an overturned car. Another helper decides that the license plate number of the car should be obtained, and (maybe another) image analysis helper provides this information. The plate number is used by another helper to check in a vehicle database whether the car is equipped with the OnStar communication system. If it is, the appropriate OnStar center facility is contacted, becomes a helper, and obtains a connection with the OnStar device in the car. The OnStar device in the car becomes a helper and is asked to contact BANs (body area networks) on and within bodies of car occupants. Each BAN available in the car becomes a helper and reports on the vital signs of its owner. The reports from BANs are analyzed by prioritizing helpers that schedule the responder teams to ensure that people in the most serious condition are rescued sooner than others. With the exception of the BAN link that is just a bit futuristic (its widespread availability could be measured in years not in decades), all other helper capabilities are already quite common.
A simple activity
Identify the knowledge embedded/ implicit for the above scenario to work !
1/25/2012
What is Knowledge
Numerous definitions
Working Knowledge: how organisations manage what they know Harvard Business School Press, 1998, 2000 Knowledge is information combined with experience, context, interpretation, and reflection. It is a high-value form of information that is ready to apply to decisions and actions." T. Davenport et al., 1998 Knowledge is information evaluated and organized by the human mind so that it can be used purposefully, e.g., conclusions or explanations." Rousa, 2002.
1/25/2012
In organizations, it often becomes embedded not only in documents or repositories but also in organizational routines, processes, practices and norms. - Thomas H. Davenport, Laurence Prusak (1998) "Knowledge is... a mental grasp of fact(s) of reality, reached either by perceptual observation or by a process of reason based on perceptual observation." Rand, 1967.
1/25/2012
How KR Works
Intelligence requires knowledge Computational models of intelligence require models of knowledge Use formalisms to write down knowledge
Expressive enough to capture human knowledge Precise enough to be understood by machines
1/25/2012
1/25/2012
Early KR Schemes
Model-based representations reflecting the structure of the domain, and then reason based on the model.
Semantic Nets Frames Scripts
10
11
Semantic Nets
labeled, directed graph nodes represent objects, concepts, or situations
labels indicate the name nodes can be instances (individual objects) or classes (generic nodes)
12
Some slides adapted from Dorr, www.umiacs.umd.edu/~christof/courses/cmsc723-fall04 , Kurfess: www.csc.calpoly.edu/~fkurfess/Courses/CSC481/W03/Slides/3-Knowledge-Representation.ppt and Hirschberg: www1.cs.columbia.edu/~julia/cs4705/syllabus.html
13
Generic/Individual
Generic describes the idea--the notion
static
A lot of NLP using semantic nets involves instantiating generic nets based on a given piece of text.
Some slides adapted from Dorr, www.umiacs.umd.edu/~christof/courses/cmsc723-fall04 , Kurfess: www.csc.calpoly.edu/~fkurfess/Courses/CSC481/W03/Slides/3-Knowledge-Representation.ppt and Hirschberg: www1.cs.columbia.edu/~julia/cs4705/syllabus.html
14
Individuation example
Person agent give recipient Person
object
Thing Process the sentence Bob gave Mary some candy. Generic Representation
Bob agent
Mary
Candy
Instantiation
Some slides adapted from Dorr, www.umiacs.umd.edu/~christof/courses/cmsc723-fall04 , Kurfess: www.csc.calpoly.edu/~fkurfess/Courses/CSC481/W03/Slides/3-Knowledge-Representation.ppt and Hirschberg: www1.cs.columbia.edu/~julia/cs4705/syllabus.html
15
Frames
Represent related knowledge about a subject Frame has a title and a set of slots
Title is what the frame is the concept Slots capture relationships of the concept to other things
16
Some slides adapted from Dorr, www.umiacs.umd.edu/~christof/courses/cmsc723-fall04 , Kurfess: www.csc.calpoly.edu/~fkurfess/Courses/CSC481/W03/Slides/3-Knowledge-Representation.ppt and Hirschberg: www1.cs.columbia.edu/~julia/cs4705/syllabus.html
17
Usage of Frames
Most operations with frames do one of two things: Fill slots
Process a piece of text to identify an entity for which we have a frame Fill as many slots as possible
Scripts
Describe typical events or sequences Components are
script variables (players, props) entry conditions transactions exit conditions
19
Script for a typical sequence of activities at a restaurant Often has a frame behind it; script is essentially instantiating the frame
Some slides adapted from Dorr, www.umiacs.umd.edu/~christof/courses/cmsc723-fall04 , Kurfess: www.csc.calpoly.edu/~fkurfess/Courses/CSC481/W03/Slides/3-Knowledge-Representation.ppt and Hirschberg: www1.cs.columbia.edu/~julia/cs4705/syllabus.html
20
EAT-AT-RESTAURANT Script Props: (Restaurant, Money, Food, Menu, Tables, Chairs) Roles: (Hungry-Persons, Wait-Persons, Chef-Persons) Point-of-View: Hungry-Persons Time-of-Occurrence: (Times-of-Operation of Restaurant) Place-of-Occurrence: (Location of Restaurant) Event-Sequence: first: Enter-Restaurant Script then: if (Wait-To-Be-Seated-Sign or Reservations) then Get-Maitre-d's-Attention Script then: Please-Be-Seated Script then: Order-Food-Script then: Eat-Food-Script unless (Long-Wait) when Exit-Restaurant-Angry Script CSC 9010: Special then: if (Food-Quality was better than Palatable) Topics, Natural then Language Processing. Compliments-To-The-Chef Script Spring, then: 2005. Pay-For-It-Script Matuszek & finally: Papalaskari Leave-Restaurant Script
Restaurant Script
21
[Rogers 1999]
Comments on Scripts
Obviously takes a lot of time to develop them initially.
The script itself has much of the knowledge May be serious overkill for most NLP tasks
We need this level of detail if we want to include answers based on reasoning like Most restaurants do serve dinner.
Some slides adapted from Dorr, www.umiacs.umd.edu/~christof/courses/cmsc723-fall04 , Kurfess: www.csc.calpoly.edu/~fkurfess/Courses/CSC481/W03/Slides/3-Knowledge-Representation.ppt and Hirschberg: www1.cs.columbia.edu/~julia/cs4705/syllabus.html
22
Enter - WWW
Web 1.0
Personal Websites Content Management Systems Altavista, Google CiteSeer, Project Gutenberg Message Boards Buddy Lists, Address Books -
Web 2.0
Blogs Wikis
Semantic Web
1/25/2012
1/25/2012
Productivity of Search
The 4.0 Web 2020 - 2030 Intelligent Reasoning Web Semantic Search
PC Era
1980 - 1990
Amount of data
Intelligent Web
Web 4.0
2020 - 2030 Intelligent personal agents
The Internet
FTP PCs USENET
Distributed Search SWRL OWL 2010 - 2020 SPARQL Semantic Databases OpenID AJAX Semantic Search ATOM Widgets Social Web RSS Mashups P2P RDF Office 2.0 Javascript Flash SOAP XML Social Media Sharing 2000 - 2010 Weblogs Java The Web HTML SaaS Social Networking HTTP Directory Portals Wikis VR Keyword Search Lightweight Collaboration The PC BBS Websites Gopher 1990 - 2000 MacOS SQL MMOs Groupware Databases SGML Windows File Servers
Web 3.0
Web 2.0
Web 1.0
IRC Email
PC Era
1980 - 1990
File Systems
Better search
Products
Places
Interests
Services
Deeper integration
Activities Web Pages
Richer content
Documents
Better personalization
Multimedia
Subscriber to
Jane Person
Friend of
Source of
Tagging
Statistics
Top-Down (Contemporary)
Automatically generate semantic metadata for vertical domains Create services that provide this as an overlay to nonsemantic Web Nobody has to learn RDF/OWL
Tagging
Semantic Web
Top-down
Statistics Linguistics Bottom-up
Artificial intelligence
Web users
(profiles, preferences)
Shared ontology
Multimedia resources
35
Ontology
An ontology formally defines a common set of terms that are used to describe and represent a domain (e.g., librarianship, medicine, etc.) Ontologies include computer-usable definitions of basic concepts in the domain and the relationships among them Ontologies are usually expressed in a logicbased language
Dr.T.V.Geetha, Anna University 38
RDF Triples
Subject
Predicate
Object
the subject, which is an RDF URI reference or a blank node the predicate, which is an RDF URI reference the object, which is an RDF URI reference, a literal or a blank node
Source: http://www.w3.org/TR/rdf-concepts/#section-triples
Ontologies
ID
Definition
Definition
More on Ontology
Aristotle - Ontology
Before: study of the nature of being Since Aristotle: study of knowledge representation and reasoning Terminology: Genus: (Classes) Species: (Subclasses) Differentiae: (Characteristics which allow to group or distinguish objects from each other)
What is ontology?
Philosophy (400BC):
Neches (91):
Ontology defines basic terms and relations comprising the vocabulary of a topic area as well as the rules for combining terms and relations to define extensions to the vocabulary
Gruber (93): Explicit
specification of a conceptualization
Borst (97):
43
Machine readable
Consensual knowledge
44
Terms: Labels for concepts Protein, Gene, Relationships: Semantic links between concepts
Is-a-kind, is-a, part-of, name-of,
45
Concept - Instance
Concept / Class / Universal (Metaphysics)
an abstract or general idea inferred or derived from specific instances
46
Developing an ontology
Defining classes in the ontology
Concepts in a domain of discourse (classes -sometimes called concepts)
Restrictions on slots
(facets - sometimes called role restrictions)
Knowledge Management
Ontological Engineering
Knowledge Engineering Software/Data Engineering
Object Modeling
48
Benefits
Communication between people Interoperability between software agents
Reuse of domain knowledge
49
Types of ontologies
[Guarino et al. 1999] - N. Guarino, C. Masolo, G. Vetere. OntoSeek: Content-Based Access to the Web. In: IEEE Intelligent Systems, 14(3), 70--80, 1999.
50
Ontology -Examples
fruit
fruit apple citrus pear
tropical temperate
lime
lemon orange
Dr.T.V.Geetha, Anna
52 University
52
Ontology- Example II
subClassOf
domain
Person
subClassOf
range
hasSuperVisor
hasSuperVisor
Living Thing
eats
Plant
Grass Herbivore
Tree Cow
54
Ontologies Example IV
Geographical Entity (GE)
is-a flow_through
Natural GE
Inhabited GE
capital_of
mountain
instance_of
river
country
located_in
city
Zugspitze
height (m)
Neckar
length (km)
2962
367
Stuttgart
Berlin
Design: Philipp Cimiano
55
Ontology Example V
Object
is_a-1 is_a-1
Person
is_a-1 is_a-1
knows
Topic
described_in related_to
Document
is_a-1 is_a-1
writes
Student
is_a-1
Researcher
is_a-1
Letter
RULES:
T described_in is_about D T T P is_about knows D
Tel
Affiliation
Ram
+49 721 608 . AIFB
writes
Representation Languages: RDF(S); OWL; Predicate Logic; FDr.T.V.Geetha, Anna Logic University 56
Ontology Example VI
Chemical
Molecule
Compound
Element
Ion
Atom
Molecular Compound
Ionic Compound
Molecular Element
Ionic Molecule
Non-Metal
Metal
Metaloid
EcoCyc
58
59
60
strong semantics
Conceptual Model
RDF/S XTM Extended ER
Is Subclass of
Semantic Interoperability
Thesaurus
ER DB Schemas, XML Schema
Taxonomy
Relational Model, XML
weak semantics
strong semantics
Thesaurus
DB Schemas, XML Schema
Taxonomy
Model, XML
weak semantics
Ontology Applications
Information Retrieval
Query Expansion
Information Extraction
Template Definition, Semantic Integration
Question Answering
Question Analysis, Answer Selection
Document Clustering/Classification
Extend Bag-of-words
Knowledge Management
Check Consistency, Infer New Knowledge
64
Big Ontologies
There are several large, general ontologies that are freely available. Some examples are:
Cyc - Original general purpose ontology OntoSem a lexical KR system and ontology WordNet - a large, on-line lexical reference system World Fact Book -- 5Meg of KIF sentences! UMLS - NLMs Unified Medical Language System SUMO Standard Upper Merged Ontology
65
Building an Ontology
Ontology Elements
Concepts(classes) + their hierarchy Concept properties (slots/attributes) Property restrictions (type, cardinality, domain) Relations between concepts (disjoint, equality) Instances
67
68
Consider Reuse
With the spreading deployment of the Semantic Web, ontologies will become more widely available We rarely have to start from scratch when defining an ontology
There is almost always an ontology available from a third party that provides at least a useful starting point for our own ontology
Dr.T.V.Geetha, Anna University 69
Possible questions: Distance between two cities? What sort of connections exist between two cities? In which country is a city? How many borders are crossed?
71
Enumerate Terms
Write down in an unstructured list all the relevant terms that are expected to appear in the ontology Nouns form the basis for class names Verbs (or verb phrases) form the basis for property names Traditional knowledge engineering tools can be used to obtain the set of terms an initial structure for these terms
72
city
capital country
Connection_on_land
border
road
Connection_on_water
railway
currency
connection
Dr.T.V.Geetha, Anna University 73
Connection_in_air
Identify properties
What properties do the terms have? What would you like to say about the terms? Initially get comprehensive list of terms do not worry about overlap between concepts they represent Relations among terms Properties concepts have Whether concepts classes or slots Closely integrated steps Developing Class hierarchy Defining properties of concepts (slots)
Dr.T.V.Geetha, Anna University 74
Define Taxonomy
Relevant terms must be organized in a taxonomic hierarchy
Opinions differ on whether it is more efficient/reliable to do this in a top-down or a bottom-up fashion
76
Define facets
Define facets of Slots Value type --- string, number, boolean, enumerated, instance type Allowed values Number of values (cardinality) Single and multiple Minimum and maximum cardinality Other features slot can take
Dr.T.V.Geetha, Anna University
77
Define Properties
Often interleaved with the previous step The semantics of subClassOf demands that whenever A is a subclass of B, every property statement that holds for instances of B must also apply to instances of A It makes sense to attach properties to the highest class in the hierarchy to which they apply
Dr.T.V.Geetha, Anna University 78
Country
Has_capital
City
Start_point
Connection
Borders_with
Capital_of
Capital_city
D Subclass_of(A)
A=B
81
Issues on Slots
- inverse slots (redundant, but explicit)
Has_capital
Country
Capital_of
Capital_city
B borders_with A
A.connection(C)
82
Define Instances
Filling the ontologies with instances is a separate step Number of instances >> number of classes Thus populating an ontology with instances is not done manually
Retrieved from legacy data sources (DBs) Extracted automatically from a text corpus
Dr.T.V.Geetha, Anna University 83
1/25/2012
Nyaaya (Tarka)
Ontology
Structure of knowledge A systematic account of Existence Defines the terms used to describe and represent knowledge Allows detailed, accurate, precise, consistent, sound, and meaningful distinctions among the classes, properties, and relations.
Inference
1/25/2012 KM Workshop AU, Dec2007
Substance (dravya) - 9 Quality (guna) - 24 Action (karma) -5 Universal/Commonality (saamaanya) Particularity (visesha) Inherence (samavaaya) Non-existence (abhaava) - 4
KM Workshop AU, Dec2007
Space (ether) sound as its Dec2007 1/25/2012 KM Workshop AU, inherent quality
1/25/2012
Gnaanam cognition/knowledge
The quality which is the cause of all kinds of transactions (communication) Two types
Remembrance (smrti)
Born of mental impressions
Apprehension (anubhava)
True & Untrue ( conforming to the object or not) ( valid / invalid )
1/25/2012 KM Workshop AU, Dec2007
Gnaanam cognition/knowledge(2)
Valid apprehension (4 types)
Perception (pratyaksha) Inference (anumaana) Analogy based (upamaana) Verbal testimony (shabda)
Perception
1/25/2012
Inference
Extract the hidden essence from observable facts Identify the implied Cause and effect Deduction and Induction
Universal to Particular Vs Particular to General
1/25/2012
1/25/2012
Inference - Formalism
Five part syllogism
Prathigna (Statement) the mountain has fire Hetu (reason) - because there is smoke Vyaapti (invariable concomitance)
Smoke is always pervaded by fire
Udaharana (example/illustration)
As in a hearth / kerosene stove
Nigamana
(conclusion)
1/25/2012
Vyatireka vyaapti
Not B implies Not A
Anvaya-vyatireka vyapti
Both
1/25/2012
Verbal Testimony
Shabda (words) Vakya (sentences) Valid sentences
Akaanksha (mutual expectation) Sannidhi (proximity) Yogyata (fitness / aptness/ no-nonsense)
1/25/2012
KRIL
Knowledge representation based on Indian Logic
1/25/2012
1/25/2012
Tamil
105
106
107
108
109
110
111
UNL Expression
[d]
[s] [w] ;tirunelveli;icl>place;1 ;nellaiappar temple;iof>temple;2 ;gandimathi amman temple;iof>temple;3 ;krishnapuram temple;iof>temple;4 ;sree vaikundam temple;iof>temple;5 ;necessity;icl>attribute;6 ;go;icl>do;7 ;pure;aoj>thing;8 ;place;icl>thing;9 [/w] [r] 0 plc 1 c1 0 man 6 c2 0 pur 7 c3 9 mod 8 c4 [/r] [/s]
112
113
114
115
116
117
118
119
agt/obj
frm/tmt/aoj
agt/cag/to
plc
pos
plc/and
int
127
plf/plt/via/agt
nam
iof
131
132
133
134
135
136
137
1/25/2012
Summary
State of art on KR Details on Ontology Construction Conceptual Inter-lingua framework
1/25/2012
!! Thank you !!
rp@annauniv.edu
1/25/2012 KM Workshop AU, Dec2007
Substance (dravya)
Substratum of qualities Nine substances
Back
1/25/2012 KM Workshop AU, Dec2007
Earth (smell) Water (cold touch) Fire (warm touch) Air (felt but not seen) Ether/Space (sound)
-Time
-Direction
-Soul (Knowledge) -Mind (emotions)
Quality (guna)
rupa (color) rasa (taste) gandha(smell) sparsha (touch) sankhya (no.) parimaNa(magnitude) Prthaktva(separateness) samyoga (conjunction) Vibhaga (disjunction) paratva (remoteness) Aparatva (proximity) gurutva (heaviness) Dravatva (fluidity) sneha (stickiness) shabda (sound) Budhi (cognition) sukha (pleasure) dukha (sorrow) Iccha (desire) dvesa (dislike) praytna (effort) dharma (merit) Adharma (demerit) samskara (tendency)
1/25/2012 KM Workshop AU, Dec2007
Back
Action (karma)
Upward Downward Contraction Expansion Motion (lateral)
Back
1/25/2012 KM Workshop AU, Dec2007
Non-existence / Negation
Pragabhaava (prior / antecedent negation)
Before creation
1/25/2012