SlideShare a Scribd company logo
1 of 130
NLP & Semantic Computing Group
N L P
Different Semantic Perspectives for
Hybrid Question Answering Systems
Andre Freitas
University of Passau
OKBQA, Jeju, 2016
NLP & Semantic Computing Group
http://www.slideshare.net/andrenfreitas
These slides:
NLP & Semantic Computing Group
Outline
 Multiple Perspectives of Semantic Representation
 Lightweight Semantic Representation
 Knowledge Graph Extraction from Text
 Answering Queries with
Knowledge Graphs
 Reasoning
 Take-away Message
NLP & Semantic Computing Group
Multiple Perspectives of
Semantic Representation
NLP & Semantic Computing Group
QA & Semantics
• Question Answering is about managing semantic
representation, extraction, selection
trade-offs.
• And it is about integrating multiple components
in a complex approach.
• Semantic best-effort, systems tolerant to
noisy, inconsistent, vague, data.
5
NLP & Semantic Computing Group
 “Most semantic models have dealt with particular types of
constructions, and have been carried out under very simplifying
assumptions, in true lab conditions.”
 “If these idealizations are removed it is not clear at all that modern
semantics can give a full account of all but the simplest
models/statements.”
Formal World Real World
Baroni et al. 2013
Semantics for a Complex World
6
NLP & Semantic Computing Group
Why Not RDF?
• Follows a more “database-type” of
representation perspective.
• Gap towards representing text.
NLP & Semantic Computing Group
Choices of Semantic Representation
• Logical
• Frames: verbs | nouns
• Binary relations: binary | n-ary
• Named entities
• Language Models
• Syntactic structures
• Bag-of-words
Concept-level
representation
Background
knowledge
Extraction
complexity
8
NLP & Semantic Computing Group
Information Extraction
• Logical
• Frames:
 verbs | nouns
• Binary relations:
 binary | n-ary
• Named entities
• Syntactic Structures & LMs
• Bag-of-words
• Semantic parsing
• Semantic role labeling
• Relation extraction:
– closed/open
• Named entity recognition
• Syntactic/N-gram Parsing
• Indexing9
NLP & Semantic Computing Group
Information Extraction
• Logical
• Frames:
 verbs | nouns
• Binary relations:
 binary | n-ary
• Named entities
• Syntactic Structures & LMs
• Bag-of-words
• Semantic parsing
• Semantic role labeling
• Relation extraction:
– closed/open
• Named entity recognition
• Syntactic/N-gram Parsing
• Indexing
Use all of them!
NLP & Semantic Computing Group
Representation focal points
• Types of knowledge to focus at the
representation:
 Facts vs Definitions vs Opinions
 Temporality
 Spatiality
 Modality
 Polarity
 Rhetorical structures
 …
11
NLP & Semantic Computing Group
Lightweight Semantic
Representation
NLP & Semantic Computing Group
Objective
• Provide a lightweight knowledge
representation model which:
 Can represent textual discourse information.
• Maximizes the capture of textual
information.
 Is convenient to extract from text.
 Is convenient to access (query and browse).
13
NLP & Semantic Computing Group
Lightweight Semantic Representation
Representing Texts as Contextualized Entity-Centric
Linked Data Graphs, WebS 2013
NLP & Semantic Computing Group
Representation Assumptions
• Data integration:
 Named entities (instances)
 Abstract classes (unary predicates)
• Rich taxonomical structures.
• Context representation as a first class citizen.
• Open vocabulary.
• Word instead of sense/concept.
15
NLP & Semantic Computing Group
Representation Assumptions
• Data integration:
 Named entities (instances)
 Abstract classes (unary predicates)
• Rich taxonomical structures.
• Context representation as a first class citizen.
• Open vocabulary.
• Word instead of sense/concept.
16
NLP & Semantic Computing Group
Representation of Complex Relations
General Electric Company, or GE , is an American multinational conglomerate
corporation incorporated in Schenectady , New York
17
NLP & Semantic Computing Group
Data Integration points
General Electric Company, or GE , is an American multinational conglomerate
corporation incorporated in Schenectady , New York
Named entities are lower
entropy integration points
Pivot
points18
NLP & Semantic Computing Group
Data Integration points
General Electric Company, or GE , is an American multinational conglomerate
corporation incorporated in Schenectady , New York
Named entities are also
low entropy entry points
for answering queries
Pivot
points19
NLP & Semantic Computing Group
Data Integration points
General Electric Company, or GE , is an American multinational conglomerate
corporation incorporated in Schenectady , New York
Also abstract classes …
Pivot
points20
NLP & Semantic Computing Group
Data Integration points
General Electric Company, or GE , is an American multinational conglomerate
corporation incorporated in Schenectady , New York
They are also a very
convenient way to
represent.
Pivot
points21
NLP & Semantic Computing Group
Representation Assumptions
• Data integration:
 Named entities (instances)
 Abstract classes (unary predicates)
• Rich taxonomical structures.
• Context representation as a first class citizen.
• Open vocabulary.
• Word instead of sense/concept.
22
NLP & Semantic Computing Group
Taxonomy Extraction
 Are predicates with more complex compositional patterns which
describe sets.
 Parsing complex nominals.
American multinational conglomerate corporation
On the Semantic Representation and Extraction of Complex
Category Descriptors, NLDB 2014
multinational conglomerate corporation
corporation
conglomerate corporation
is a
is a
is a
Pivot
points
NLP & Semantic Computing Group
Representation Assumptions
• Data integration:
 Named entities (instances)
 Abstract classes (unary predicates)
• Rich taxonomical structures.
• Context representation as a first class citizen.
• Open vocabulary.
• Word instead of sense/concept.
24
NLP & Semantic Computing Group
Context Representation
General Electric Company, or GE , is an American multinational conglomerate
corporation incorporated in Schenectady , New York
Reification as a first class
representation element
25
NLP & Semantic Computing Group
Context Representation
General Electric Company, or GE , is an American multinational conglomerate
corporation incorporated in Schenectady , New York
Temporality, spatiality,
modality, rhetorical relations
…
26
NLP & Semantic Computing Group
Rhetorical Structures using Reification
• cause:
 e.g. “because scraping the bottom with a metal utensil will scratch
the surface.”
• circumstance
 e.g. “After completing your operating system reinstallation,”
• concession
 e.g. “Although the hotel is situated adjacent to a beach,”
• condition
 e.g. “If you can break the $ 1000 dollar investment range,”
• contrast
 e.g. “but you can do better with 2.4ghz or 900mhz phones.”
• purpose
 e.g.“in order for the rear passengers to get in the vehicle.”
• …
27
NLP & Semantic Computing Group
Representation Assumptions
• Data integration:
 Named entities (instances)
 Abstract classes (unary predicates)
• Rich taxonomical structures.
• Context representation as a first class citizen.
• Open vocabulary.
• Word instead of sense/concept.
28
NLP & Semantic Computing Group
Open Vocabulary
General Electric Company, or GE , is an American multinational conglomerate
corporation incorporated in Schenectady , New York
Temporality, spatiality,
modality, rhetorical relations
…
29
NLP & Semantic Computing Group
Open Vocabulary
• Easier to extract but difficult to consume.
• We pay the price at query time.
• How to operate over a large-scale semantically
heterogeneous knowledge-graphs?
30
NLP & Semantic Computing Group
Representation Assumptions
• Data integration:
 Named entities (instances)
 Abstract classes (unary predicates)
• Rich taxonomical structures.
• Context representation as a first class citizen.
• Open vocabulary.
• Word instead of sense/concept.
31
NLP & Semantic Computing Group
Words instead of Senses
• Motivation: Disambiguation is a tough problem.
• Sense granularity can be, at many situations,
arbitrary (too context dependent).
• We treat a word as a superposition of senses,
almost in a “quantum mechanical sense”.
32
NLP & Semantic Computing Group
SenseSuperposition
Coecke et al. (2010): Category theory and
Lambek calculus.
NLP & Semantic Computing Group
Revisited RDF (for Representing Texts)
• Data Model Types: Instance, Class, Property…
• RDFS: Taxonomic representation.
• Reification for contextual relations (subordinations).
• Blank nodes for n-ary relations.
• Triple.
• Labels over URIs.
34
NLP & Semantic Computing Group
Abstract Meaning Representations – AMR,
Maximal Use of PropBank Frame Files
Alternative Representations
NLP & Semantic Computing Group
Distributional
Semantics
NLP & Semantic Computing Group
Distributional Semantic Models
 Semantic Model with low acquisition effort
(automatically built from text)
Simplification of the representation
 Enables the construction of comprehensive
commonsense/semantic KBs
 What is the cost?
Some level of noise
(semantic best-effort)
Limited semantic model
37
NLP & Semantic Computing Group
Distributional Semantics as Commonsense Knowledge
Commonsense is here
θ
car
dog
cat
bark
run
leash
Semantic Approximation is here
38
NLP & Semantic Computing Group
I find it rather odd that people are already trying to tie the
Commission's hands in relation to the proposal for a
directive, while at the same calling on it to present a Green
Paper on the current situation with regard to optional and
supplementary health insurance schemes.
I find it a little strange to now obliging the Commission to a
motion for a resolution and to ask him at the same time to
draw up a Green Paper on the current state of voluntary
insurance and supplementary sickness insurance.
=?
Beyond Single Word Vector Models:
Compositionality
NLP & Semantic Computing Group
Compositional Semantics
 Can we extend DS to account for the meaning of
phrases and sentences?
 Compositionality: The meaning of a complex
expression is a function of the meaning of its
constituent parts.
NLP & Semantic Computing Group
Compositional Semantics
Words in which the
meaning is directly
determined by their
distributional behaviour
(e.g., nouns).
Words that act as functions
transforming the
distributional profile of
other words (e.g., verbs,
adjectives, …).
41
NLP & Semantic Computing Group
Compositional-Distributional Semantics
NLP & Semantic Computing Group
Recursive Neural Networks for
Structure Prediction
43
NLP & Semantic Computing Group
New Model: Recursive Neural Tensor Network
• Goal: Function that composes two vectors.
• More expressive than any other RNN so far.
44 Socher et al.
NLP & Semantic Computing Group
Socher et al.
NLP & Semantic Computing Group
Compositional-distributional model for
Categories
46
NLP & Semantic Computing Group
Embedding
Knowledge Graphs
47
NLP & Semantic Computing Group
The vector space is
segmented48
Dimensional reduction
mechanism!
A Distributional Structured Semantic Space for
Querying RDF Graph Data, IJSC 2012
NLP & Semantic Computing Group
Compositional-distributional model for
paraphrases
A Compositional-Distributional Semantic Model for
Searching Complex Entity Categories, *SEM (2016)
NLP & Semantic Computing Group
Knowledge Graph Extraction from Text
NLP & Semantic Computing Group
Graphene
NLP & Semantic Computing Group
Graph Extraction Pipeline
Text
Transformation
N-ary
Relation
Extraction
Text
Simplification
Graph
Serialization
Taxonomy
Extraction
Storage
RST
Classification
ML-based
Rule-based
Rule-based
ML-based
52
NLP & Semantic Computing Group
Minimalistic Text Transformations
Text
Transformation
N-ary
Relation
Extraction
Text
Simplification
Graph
Serialization
Taxonomy
Extraction
Storage
RST
Classification
ML-based
Rule-based
Rule-based
ML-based
53
NLP & Semantic Computing Group
Minimalistic Text Transformations
• Co-reference Resolution
 Pronominal co-references.
• Passive
 We have been approached by the investment banker.
 The investment banker approached us.
• Genitive modifier
 Malaysia's crude palm oil output is estimated to have
risen.
 The crude palm oil output of Malasia is estimated to
have risen. 54
NLP & Semantic Computing Group
Text Simplification
Text
Transformation
N-ary
Relation
Extraction
Text
Simplification
Graph
Serialization
Taxonomy
Extraction
Storage
RST
Classification
ML-based
Rule-based
Rule-based
ML-based
55
NLP & Semantic Computing Group
Text Simplification for KG
Extraction
“Defeating Republican nominee Mitt Romney,
Obama, who was the first African American to hold
the office, was reelected president in November
2012.”
 relations are spread across clauses
 relations are presented in non-canonical form
56
NLP & Semantic Computing Group
Text Simplification for KG
Extraction
• Insertion of a text simplification step
 Obama was reelected president in November 2012.
 Obama was the first African American to hold the office.
 Obama was defeating Mitt Romney.
 Mitt Romney was Republican nominee.
57
NLP & Semantic Computing Group
Syntax-driven sentence
simplification approach
Task:
• Reduce the linguistic complexity of a text while retaining the
original information/meaning using a set of syntax-based
rewrite operations (deletion, insertion, reordering, sentence
splitting).
Idea:
• Simplify a sentence by separating out components that supply
only secondary information into simpler stand-alone context
sentences, thus yielding one or more reduced core sentences.
NLP & Semantic Computing Group
Approach
• Linguistic analysis of sentences from the English Wikipedia
to identify constructs which provide only secondary information:
• non-restrictive relative clauses
• non-restrictive and restrictive appositive phrases
• participial phrases offset by commas
• adjective and adverb phrases delimited by punctuation
• particular prepositional phrases
• lead noun phrases
• intra-sentential attributions
• parentheticals
• conjoined clauses with specific features
• particular punctuation
•Rule-based simplification rules.
Improving Relation Extraction by Syntax-based
Sentence Simplification (2016)
NLP & Semantic Computing Group
N-ary Relation Extraction
Text
Transformation
N-ary
Relation
Extraction
Text
Simplification
Graph
Serialization
Taxonomy
Extraction
Storage
RST
Classification
Rule-based
Rule-based
ML-based
60
OpenIE, University of Washington
NLP & Semantic Computing Group
Taxonomy Extraction
Text
Transformation
N-ary
Relation
Extraction
Text
Simplification
Graph
Serialization
Taxonomy
Extraction
Storage
RST
Classification
Rule-based
Rule-based
ML-based
61
Representation and Extraction of Complex
Category Descriptors, NLDB 2014
NLP & Semantic Computing Group
RST Classification
Text
Transformation
N-ary
Relation
Extraction
Text
Simplification
Graph
Serialization
Taxonomy
Extraction
Storage
RST
Classification
Rule-based
Rule-based
ML-based
62
NLP & Semantic Computing Group
Rhetorical Structure Extraction
63
TEXT-LEVEL RST-STYLE DISCOURSE PARSER (Feng and Hirst, 2012)
Structure classification Relation classification
NLP & Semantic Computing Group
Answering Queries with
Knowledge Graphs
NLP & Semantic Computing Group
Now our graph supports semantic
approximations as a first-class operation
NLP & Semantic Computing Group
Approach Overview
Query Planner
Ƭ-Space
(embedding graphs)
Wikipedia
Commonsense
knowledge
RDF
Explicit Semantic
Analysis
Core semantic approximation &
composition operations
Query AnalysisQuery Query Features
Query Plan
66
NLP & Semantic Computing Group
Core Principles
 Minimize the impact of Ambiguity, Vagueness,
Synonymy.
 Address the simplest matchings first (semantic pivoting).
 Semantic Relatedness as a primitive operation.
 Distributional semantics models as commonsense
knowledge representation.
 Lightweight syntactic constraints.
67
NLP & Semantic Computing Group
• Step 2: Query NER
 Rules-based: POS Tag + IDF
Who is the daughter of Bill Clinton married to?
(PROBABLY AN INSTANCE)
Query Pre-Processing
(Question Analysis)
68
NLP & Semantic Computing Group
• Step 3: Determine answer type
 Rules-based.
Who is the daughter of Bill Clinton married to?
(PERSON)
Query Pre-Processing
(Question Analysis)
69
NLP & Semantic Computing Group
• Transform natural language queries into a
pseudo-logical form.
“Who is the daughter of Bill Clinton married to?”
Query Pre-Processing
(Question Analysis)
70
NLP & Semantic Computing Group
Query Pre-Processing
(Question Analysis)
Bill Clinton daughter married to
(INSTANCE)
Person
ANSWER
TYPE
QUESTION FOCUS
71
• Step 5: Determine the query pattern
 Rules based.
• Remove stop words.
• Merge words into entities.
• Reorder structure from core entity position.
NLP & Semantic Computing Group
• Step 5: Determine the query pattern
 Rules based.
• Remove stop words.
• Merge words into entities.
• Reorder structure from core entity position.
Query Pre-Processing
(Question Analysis)
Bill Clinton daughter married to
(INSTANCE)
Person
(PREDICATE) (PREDICATE) Query Features
72
NLP & Semantic Computing Group
• Map query features into a query plan.
• A query plan contains a sequence of:
 Search operations.
 Navigation operations.
Query Planning
(INSTANCE) (PREDICATE) (PREDICATE) Query Features
 (1) INSTANCE SEARCH (Bill Clinton)
 (2) DISAMBIGUATE ENTITY TYPE
 (3) GENERATE ENTITY FACETS
 (4) p1 <- SEARCH RELATED PREDICATE (Bill Clintion, daughter)
 (5) e1 <- GET ASSOCIATED ENTITIES (Bill Clintion, p1)
 (6) p2 <- SEARCH RELATED PREDICATE (e1, married to)
 (7) e2 <- GET ASSOCIATED ENTITIES (e1, p2)
 (8) POST PROCESS (Bill Clintion, e1, p1, e2, p2)
Query Plan
73
NLP & Semantic Computing Group
Core Entity Search
Bill Clinton daughter married to Person
:Bill_Clinton
Query:
KB:
Entity search
74
NLP & Semantic Computing Group
Distributional Semantic Search
Bill Clinton daughter married to Person
:Bill_Clinton
Query:
:Chelsea_Clinton
:child
:Baptists
:religion
:Yale_Law_School
:almaMater
...
(PIVOT ENTITY)
(ASSOCIATED
TRIPLES)
75
KB:
NLP & Semantic Computing Group
Distributional Semantic Search
Bill Clinton daughter married to Person
:Bill_Clinton
Query:
:Chelsea_Clinton
:child
:Baptists
:religion
:Yale_Law_School
:almaMater
...
sem_rel(daughter,child)=0.054
sem_rel(daughter,child)=0.004
sem_rel(daughter,alma mater)=0.001
Which properties are semantically related to ‘daughter’?
76
KB:
NLP & Semantic Computing Group
Distributional Semantic Search
Bill Clinton daughter married to Person
:Bill_Clinton
Query:
:Chelsea_Clinton
:child
77
KB:
NLP & Semantic Computing Group
Distributional Semantic Search
Bill Clinton daughter married to Person
:Bill_Clinton
Query:
:Chelsea_Clinton
:child
(PIVOT ENTITY)
78
KB:
NLP & Semantic Computing Group
Distributional Semantic Search
Bill Clinton daughter married to Person
:Bill_Clinton
Query:
:Chelsea_Clinton
:child
:Mark_Mezvinsky
:spouse
79
KB:
Note the lazy
disambiguation
NLP & Semantic Computing Group
80
NLP & Semantic Computing Group
What is the highest mountain?
Second Query Example
(CLASS) (OPERATOR) Query Features
mountain - highest PODS
81
NLP & Semantic Computing Group
Entity Search
Mountain highest
:Mountain
Query:
:typeOf
(PIVOT ENTITY)
82
KB:
NLP & Semantic Computing Group
Extensional Expansion
Mountain highest
:Mountain
Query:
:Everest
:typeOf
(PIVOT ENTITY)
:K2:typeOf
...
83
KB:
NLP & Semantic Computing Group
Distributional Semantic Matching
Mountain highest
:Mountain
Query:
:Everest
:typeOf
(PIVOT ENTITY)
:K2:typeOf
...
:elevation
:location
...
:deathPlaceOf
84
KB:
NLP & Semantic Computing Group
Get all numerical values
Mountain highest
:Mountain
Query:
:Everest
:typeOf
(PIVOT ENTITY)
:K2:typeOf
...
:elevation
:elevation
8848 m
8611 m
85
KB:
NLP & Semantic Computing Group
Apply operator functional definition
Mountain highest
:Mountain
Query:
:Everest
:typeOf
(PIVOT ENTITY)
:K2:typeOf
...
:elevation
:elevation
8848 m
8611 m
SORT
TOP_MOST
86
KB:
NLP & Semantic Computing Group
Results
87
NLP & Semantic Computing Group
StarGraph
• Open source NoSQL platform for building and
interacting with large and sparse knowledge
graphs.
• Semantic approximation as a built-in operation.
• Scalable query execution performance.
88
NLP & Semantic Computing Group
Heuristics for the selection of the
semantic pivot is critical!
• Discussed here just superficially:
 Information-theoretical justification.
How hard is the Query? Measuring the Semantic Complexity of Schema-
Agnostic Queries, IWCS (2015).
Schema-agnositc queries over large-schema databases: a distributional
semantics approach, PhD Thesis (2015).
On the Semantic Mapping of Schema-agnostic Queries: A Preliminary
Study, NLIWoD (2015).
89
NLP & Semantic Computing Group
Indra
 Multilingual platform for experimentation with
different word vector models.
"Indra's net" is the net of the Vedic god Indra, whose net hangs over
his palace on Mount Meru, the axis mundi of Hindu cosmology and
Hindu mythology. Indra's net has a multifaceted jewel at each vertex,
and each jewel is reflected in all of the other jewels.
In the Avatamsaka Sutra, the image of "Indra's net" is used to describe
the interconnectedness of the universe.
NLP & Semantic Computing Group
NLP & Semantic Computing Group
Indra
NLP & Semantic Computing Group
NLP & Semantic Computing Group
Bridging Structured & Unstructured Data
• NER + Text + Passage Retrieval Ranking
 Simple and powerful QA basis.
• Lazy disambiguation.
94
NLP & Semantic Computing Group
Treo Answers Jeopardy Queries (Video)
NLP & Semantic Computing Group
Ranking Candidate Answers
• But what if there are multiple candidate answers!
Q: Who was Queen Victoria’s second son?
• Answer Type: Person
• Passage:
The Marie biscuit is named after Marie Alexandrovna, the daughter of
Czar Alexander II of Russia and wife of Alfred, the second son of Queen
Victoria and Prince Albert
96Dan Jurafky’s slides
NLP & Semantic Computing Group
Ranking Candidate Answers
• But what if there are multiple candidate answers!
Q: Who was Queen Victoria’s second son?
• Answer Type: Person
• Passage:
The Marie biscuit is named after Marie Alexandrovna, the daughter of
Czar Alexander II of Russia and wife of Alfred, the second son of Queen
Victoria and Prince Albert
97Dan Jurafky’s slides
NLP & Semantic Computing Group
Feature Engineering
The Marie biscuit is named after Marie Alexandrovna, the daughter of
Czar Alexander II of Russia and wife of Alfred, the second son of Queen
Victoria and Prince Albert
followed by a ‘,’ followed by an
apposition
Who was Queen Victoria’s second son?
contains an entity in
the query
has a four-word
overlap
type =
PERSON
matches AnswerType
98
NLP & Semantic Computing Group
Propositionalisation
e0 followedBy(,) followedByAppositionContainingQueryEntities() answer …
Alfred true true true
… … …
passage
entity (e0)
entity (en)
…
The Marie biscuit is named after Marie Alexandrovna, the daughter of
Czar Alexander II of Russia and wife of Alfred, the second son of Queen
Victoria and Prince Albert
answer
99
NLP & Semantic Computing Group
Reasoning for Text Entailment
NLP & Semantic Computing Group
Beyond Word Vector Models
give birth
mother
car
θ
Distributional semantics can
give us a hint about the
concepts’ semantic proximity...
...but it still can’t tell us what
exactly the relationship
between them is
give birth
mother
???
NLP & Semantic Computing Group
Beyond Word Vector Models
give birth
mother
???
give birth
mother
???
NLP & Semantic Computing Group
Beyond Word Vector Models:
Intensional Reasoning
 Representing structured intensional-level
knowledge.
 Creation of an intensional-level reasoning
model.
104
NLP & Semantic Computing Group
Commonsense Reasoning
 Selective (focussed) reasoning
 - Selecting the relevant facts in the context of the
inference
 Reducing the search space.
Scalability
NLP & Semantic Computing Group
Extended WordNet (XWN)
NLP & Semantic Computing Group
Commonsense Data (ConceptNet)
http://conceptnet5.media.mit.edu/
107
NLP & Semantic Computing Group
Distributional semantic relatedness as a
Selectivity Heuristics
Distributional
heuristics
108
target
source
answer
NLP & Semantic Computing Group
Distributional semantic relatedness as a
Selectivity Heuristics
Distributional
heuristics
109
target
source
answer
NLP & Semantic Computing Group
Distributional semantic relatedness as a
Selectivity Heuristics
Distributional
heuristics
110
target
source
answer
NLP & Semantic Computing Group
John Smith EngineerInstance-level
occupation
Does John Smith have a degree?
NLP & Semantic Computing Group
NLP & Semantic Computing Group
NLP & Semantic Computing Group
NLP & Semantic Computing Group
NLP & Semantic Computing Group
NLP & Semantic Computing Group
NLP & Semantic Computing Group
NLP & Semantic Computing Group
NLP & Semantic Computing Group
NLP & Semantic Computing Group
NLP & Semantic Computing Group
NLP & Semantic Computing Group
NLP & Semantic Computing Group
NLP & Semantic Computing Group
NLP & Semantic Computing Group
NLP & Semantic Computing Group
A Distributional Semantics Approach for
Selective Reasoning on Commonsense Graph
Knowledge Bases, NLDB (2015).
NLP & Semantic Computing Group
Bringing it into the Real World
NLP & Semantic Computing Group
Semeval 2017
NLP & Semantic Computing Group
Take-away Message
• Choosing the sweet-spot in terms of semantic
representation is critical for the construction of robust
QA systems.
 Work at a word-based representation instead of a sense
representation.
 Text simplification/clausal disembedding critical for
relation extraction.
 Need for a standardized semantic representation for
relations extracted from texts.
NLP & Semantic Computing Group
Take-away Message
• Text entailment:
 Intensional-level reasoning.
 Natural logic.
 Distributional semantics.
• Distributional semantics:
 Robust, language-agnostic semantic matching.
 Selective reasoning over commonsense KBs.

More Related Content

What's hot

Semantic Perspectives for Contemporary Question Answering Systems
Semantic Perspectives for Contemporary Question Answering SystemsSemantic Perspectives for Contemporary Question Answering Systems
Semantic Perspectives for Contemporary Question Answering SystemsAndre Freitas
 
Question Answering over Linked Data (Reasoning Web Summer School)
Question Answering over Linked Data (Reasoning Web Summer School)Question Answering over Linked Data (Reasoning Web Summer School)
Question Answering over Linked Data (Reasoning Web Summer School)Andre Freitas
 
A Distributional Semantics Approach for Selective Reasoning on Commonsense Gr...
A Distributional Semantics Approach for Selective Reasoning on Commonsense Gr...A Distributional Semantics Approach for Selective Reasoning on Commonsense Gr...
A Distributional Semantics Approach for Selective Reasoning on Commonsense Gr...Andre Freitas
 
Ontology mapping for the semantic web
Ontology mapping for the semantic webOntology mapping for the semantic web
Ontology mapping for the semantic webWorawith Sangkatip
 
Open IE tutorial 2018
Open IE tutorial 2018Open IE tutorial 2018
Open IE tutorial 2018Andre Freitas
 
Introduction to Ontology Engineering with Fluent Editor 2014
Introduction to Ontology Engineering with Fluent Editor 2014Introduction to Ontology Engineering with Fluent Editor 2014
Introduction to Ontology Engineering with Fluent Editor 2014Cognitum
 
Data Integration Ontology Mapping
Data Integration Ontology MappingData Integration Ontology Mapping
Data Integration Ontology MappingPradeep B Pillai
 
Language Models for Information Retrieval
Language Models for Information RetrievalLanguage Models for Information Retrieval
Language Models for Information RetrievalNik Spirin
 
Language Models for Information Retrieval
Language Models for Information RetrievalLanguage Models for Information Retrieval
Language Models for Information RetrievalDustin Smith
 
Datalog+-Track Introduction & Reasoning on UML Class Diagrams via Datalog+-
Datalog+-Track Introduction & Reasoning on UML Class Diagrams via Datalog+-Datalog+-Track Introduction & Reasoning on UML Class Diagrams via Datalog+-
Datalog+-Track Introduction & Reasoning on UML Class Diagrams via Datalog+-RuleML
 
Trust Models for RDF Data: Semantics and Complexity - AAAI2015
Trust Models for RDF Data: Semantics and Complexity - AAAI2015Trust Models for RDF Data: Semantics and Complexity - AAAI2015
Trust Models for RDF Data: Semantics and Complexity - AAAI2015Valeria Fionda
 
NE7012- SOCIAL NETWORK ANALYSIS
NE7012- SOCIAL NETWORK ANALYSISNE7012- SOCIAL NETWORK ANALYSIS
NE7012- SOCIAL NETWORK ANALYSISrathnaarul
 
ABSTAT: Ontology-driven Linked Data Summaries with Pattern Minimalization
ABSTAT: Ontology-driven Linked Data Summaries with Pattern MinimalizationABSTAT: Ontology-driven Linked Data Summaries with Pattern Minimalization
ABSTAT: Ontology-driven Linked Data Summaries with Pattern MinimalizationBlerina Spahiu
 
RuleML2015 - Tutorial - Powerful Practical Semantic Rules in Rulelog - Funda...
RuleML2015 - Tutorial -  Powerful Practical Semantic Rules in Rulelog - Funda...RuleML2015 - Tutorial -  Powerful Practical Semantic Rules in Rulelog - Funda...
RuleML2015 - Tutorial - Powerful Practical Semantic Rules in Rulelog - Funda...RuleML
 
An Evolution of Deep Learning Models for AI2 Reasoning Challenge
An Evolution of Deep Learning Models for AI2 Reasoning ChallengeAn Evolution of Deep Learning Models for AI2 Reasoning Challenge
An Evolution of Deep Learning Models for AI2 Reasoning ChallengeTraian Rebedea
 
Tutorial - Introduction to Rule Technologies and Systems
Tutorial - Introduction to Rule Technologies and SystemsTutorial - Introduction to Rule Technologies and Systems
Tutorial - Introduction to Rule Technologies and SystemsAdrian Paschke
 
Ontology engineering: Ontology alignment
Ontology engineering: Ontology alignmentOntology engineering: Ontology alignment
Ontology engineering: Ontology alignmentGuus Schreiber
 

What's hot (20)

Semantic Perspectives for Contemporary Question Answering Systems
Semantic Perspectives for Contemporary Question Answering SystemsSemantic Perspectives for Contemporary Question Answering Systems
Semantic Perspectives for Contemporary Question Answering Systems
 
Question Answering over Linked Data (Reasoning Web Summer School)
Question Answering over Linked Data (Reasoning Web Summer School)Question Answering over Linked Data (Reasoning Web Summer School)
Question Answering over Linked Data (Reasoning Web Summer School)
 
A Distributional Semantics Approach for Selective Reasoning on Commonsense Gr...
A Distributional Semantics Approach for Selective Reasoning on Commonsense Gr...A Distributional Semantics Approach for Selective Reasoning on Commonsense Gr...
A Distributional Semantics Approach for Selective Reasoning on Commonsense Gr...
 
Ontology mapping for the semantic web
Ontology mapping for the semantic webOntology mapping for the semantic web
Ontology mapping for the semantic web
 
Ontology
OntologyOntology
Ontology
 
Open IE tutorial 2018
Open IE tutorial 2018Open IE tutorial 2018
Open IE tutorial 2018
 
Introduction to Ontology Engineering with Fluent Editor 2014
Introduction to Ontology Engineering with Fluent Editor 2014Introduction to Ontology Engineering with Fluent Editor 2014
Introduction to Ontology Engineering with Fluent Editor 2014
 
Data Integration Ontology Mapping
Data Integration Ontology MappingData Integration Ontology Mapping
Data Integration Ontology Mapping
 
Language Models for Information Retrieval
Language Models for Information RetrievalLanguage Models for Information Retrieval
Language Models for Information Retrieval
 
Language Models for Information Retrieval
Language Models for Information RetrievalLanguage Models for Information Retrieval
Language Models for Information Retrieval
 
Datalog+-Track Introduction & Reasoning on UML Class Diagrams via Datalog+-
Datalog+-Track Introduction & Reasoning on UML Class Diagrams via Datalog+-Datalog+-Track Introduction & Reasoning on UML Class Diagrams via Datalog+-
Datalog+-Track Introduction & Reasoning on UML Class Diagrams via Datalog+-
 
Learning ontologies
Learning ontologiesLearning ontologies
Learning ontologies
 
Trust Models for RDF Data: Semantics and Complexity - AAAI2015
Trust Models for RDF Data: Semantics and Complexity - AAAI2015Trust Models for RDF Data: Semantics and Complexity - AAAI2015
Trust Models for RDF Data: Semantics and Complexity - AAAI2015
 
NE7012- SOCIAL NETWORK ANALYSIS
NE7012- SOCIAL NETWORK ANALYSISNE7012- SOCIAL NETWORK ANALYSIS
NE7012- SOCIAL NETWORK ANALYSIS
 
Ontology-based Classification and Faceted Search Interface for APIs
Ontology-based Classification and Faceted Search Interface for APIsOntology-based Classification and Faceted Search Interface for APIs
Ontology-based Classification and Faceted Search Interface for APIs
 
ABSTAT: Ontology-driven Linked Data Summaries with Pattern Minimalization
ABSTAT: Ontology-driven Linked Data Summaries with Pattern MinimalizationABSTAT: Ontology-driven Linked Data Summaries with Pattern Minimalization
ABSTAT: Ontology-driven Linked Data Summaries with Pattern Minimalization
 
RuleML2015 - Tutorial - Powerful Practical Semantic Rules in Rulelog - Funda...
RuleML2015 - Tutorial -  Powerful Practical Semantic Rules in Rulelog - Funda...RuleML2015 - Tutorial -  Powerful Practical Semantic Rules in Rulelog - Funda...
RuleML2015 - Tutorial - Powerful Practical Semantic Rules in Rulelog - Funda...
 
An Evolution of Deep Learning Models for AI2 Reasoning Challenge
An Evolution of Deep Learning Models for AI2 Reasoning ChallengeAn Evolution of Deep Learning Models for AI2 Reasoning Challenge
An Evolution of Deep Learning Models for AI2 Reasoning Challenge
 
Tutorial - Introduction to Rule Technologies and Systems
Tutorial - Introduction to Rule Technologies and SystemsTutorial - Introduction to Rule Technologies and Systems
Tutorial - Introduction to Rule Technologies and Systems
 
Ontology engineering: Ontology alignment
Ontology engineering: Ontology alignmentOntology engineering: Ontology alignment
Ontology engineering: Ontology alignment
 

Viewers also liked

Event-based MultiMedia Search and Retrieval for Question Answering
Event-based MultiMedia Search and Retrieval for Question AnsweringEvent-based MultiMedia Search and Retrieval for Question Answering
Event-based MultiMedia Search and Retrieval for Question AnsweringBenoit HUET
 
WiSS Challenge - Day 2
WiSS Challenge - Day 2WiSS Challenge - Day 2
WiSS Challenge - Day 2Andre Freitas
 
WISS QA Do it yourself Question answering over Linked Data
WISS QA Do it yourself Question answering over Linked DataWISS QA Do it yourself Question answering over Linked Data
WISS QA Do it yourself Question answering over Linked DataAndre Freitas
 
Open domain Question Answering System - Research project in NLP
Open domain  Question Answering System - Research project in NLPOpen domain  Question Answering System - Research project in NLP
Open domain Question Answering System - Research project in NLPGVS Chaitanya
 
Georgios Meditskos and Stamatia Dasiopoulou | Question Answering over Pattern...
Georgios Meditskos and Stamatia Dasiopoulou | Question Answering over Pattern...Georgios Meditskos and Stamatia Dasiopoulou | Question Answering over Pattern...
Georgios Meditskos and Stamatia Dasiopoulou | Question Answering over Pattern...semanticsconference
 
Lecture: Question Answering
Lecture: Question AnsweringLecture: Question Answering
Lecture: Question AnsweringMarina Santini
 
Query Expansion and Context: Thoughts on Language, Meaning and Knowledge Orga...
Query Expansion and Context: Thoughts on Language, Meaning and Knowledge Orga...Query Expansion and Context: Thoughts on Language, Meaning and Knowledge Orga...
Query Expansion and Context: Thoughts on Language, Meaning and Knowledge Orga...Giannis Tsakonas
 
NLP pipeline in machine translation
NLP pipeline in machine translationNLP pipeline in machine translation
NLP pipeline in machine translationMarcis Pinnis
 
SPARQL - Basic and Federated Queries
SPARQL - Basic and Federated QueriesSPARQL - Basic and Federated Queries
SPARQL - Basic and Federated QueriesKnud Möller
 
Translation Types
Translation TypesTranslation Types
Translation TypesElena Shapa
 

Viewers also liked (12)

Matchine translation
Matchine translationMatchine translation
Matchine translation
 
Indian Writing in English
Indian Writing in EnglishIndian Writing in English
Indian Writing in English
 
Event-based MultiMedia Search and Retrieval for Question Answering
Event-based MultiMedia Search and Retrieval for Question AnsweringEvent-based MultiMedia Search and Retrieval for Question Answering
Event-based MultiMedia Search and Retrieval for Question Answering
 
WiSS Challenge - Day 2
WiSS Challenge - Day 2WiSS Challenge - Day 2
WiSS Challenge - Day 2
 
WISS QA Do it yourself Question answering over Linked Data
WISS QA Do it yourself Question answering over Linked DataWISS QA Do it yourself Question answering over Linked Data
WISS QA Do it yourself Question answering over Linked Data
 
Open domain Question Answering System - Research project in NLP
Open domain  Question Answering System - Research project in NLPOpen domain  Question Answering System - Research project in NLP
Open domain Question Answering System - Research project in NLP
 
Georgios Meditskos and Stamatia Dasiopoulou | Question Answering over Pattern...
Georgios Meditskos and Stamatia Dasiopoulou | Question Answering over Pattern...Georgios Meditskos and Stamatia Dasiopoulou | Question Answering over Pattern...
Georgios Meditskos and Stamatia Dasiopoulou | Question Answering over Pattern...
 
Lecture: Question Answering
Lecture: Question AnsweringLecture: Question Answering
Lecture: Question Answering
 
Query Expansion and Context: Thoughts on Language, Meaning and Knowledge Orga...
Query Expansion and Context: Thoughts on Language, Meaning and Knowledge Orga...Query Expansion and Context: Thoughts on Language, Meaning and Knowledge Orga...
Query Expansion and Context: Thoughts on Language, Meaning and Knowledge Orga...
 
NLP pipeline in machine translation
NLP pipeline in machine translationNLP pipeline in machine translation
NLP pipeline in machine translation
 
SPARQL - Basic and Federated Queries
SPARQL - Basic and Federated QueriesSPARQL - Basic and Federated Queries
SPARQL - Basic and Federated Queries
 
Translation Types
Translation TypesTranslation Types
Translation Types
 

Similar to Different Semantic Perspectives for Question Answering Systems

Effective Semantics for Engineering NLP Systems
Effective Semantics for Engineering NLP SystemsEffective Semantics for Engineering NLP Systems
Effective Semantics for Engineering NLP SystemsAndre Freitas
 
The Role Of Ontology In Modern Expert Systems Dallas 2008
The Role Of Ontology In Modern Expert Systems   Dallas   2008The Role Of Ontology In Modern Expert Systems   Dallas   2008
The Role Of Ontology In Modern Expert Systems Dallas 2008Jason Morris
 
AI Beyond Deep Learning
AI Beyond Deep LearningAI Beyond Deep Learning
AI Beyond Deep LearningAndre Freitas
 
Building AI Applications using Knowledge Graphs
Building AI Applications using Knowledge GraphsBuilding AI Applications using Knowledge Graphs
Building AI Applications using Knowledge GraphsAndre Freitas
 
Using construction grammar in conversational systems
Using construction grammar in conversational systemsUsing construction grammar in conversational systems
Using construction grammar in conversational systemsCJ Jenkins
 
Explanations in Dialogue Systems through Uncertain RDF Knowledge Bases
Explanations in Dialogue Systems through Uncertain RDF Knowledge BasesExplanations in Dialogue Systems through Uncertain RDF Knowledge Bases
Explanations in Dialogue Systems through Uncertain RDF Knowledge BasesDaniel Sonntag
 
Knowledge representation Problem in AI.pptx
Knowledge representation Problem in AI.pptxKnowledge representation Problem in AI.pptx
Knowledge representation Problem in AI.pptxSandeepGupta229023
 
09- Syed Rehan-ai-ppt2.pptx
09- Syed Rehan-ai-ppt2.pptx09- Syed Rehan-ai-ppt2.pptx
09- Syed Rehan-ai-ppt2.pptxNandhiniV68
 
How To Make Linked Data More than Data
How To Make Linked Data More than DataHow To Make Linked Data More than Data
How To Make Linked Data More than DataAmit Sheth
 
ONTOLOGY BASED DATA ACCESS
ONTOLOGY BASED DATA ACCESSONTOLOGY BASED DATA ACCESS
ONTOLOGY BASED DATA ACCESSKishan Patel
 
Jarrar: ORM in Description Logic
Jarrar: ORM in Description Logic  Jarrar: ORM in Description Logic
Jarrar: ORM in Description Logic Mustafa Jarrar
 
Frame-Script and Predicate logic.pptx
Frame-Script and Predicate logic.pptxFrame-Script and Predicate logic.pptx
Frame-Script and Predicate logic.pptxnilesh405711
 
Rasa NLU and ML Interpretability
Rasa NLU and ML InterpretabilityRasa NLU and ML Interpretability
Rasa NLU and ML Interpretabilityztopol
 
Toward The Semantic Deep Web
Toward The Semantic Deep WebToward The Semantic Deep Web
Toward The Semantic Deep WebSamiul Hoque
 
DODDLE-OWL: A Domain Ontology Construction Tool with OWL
DODDLE-OWL: A Domain Ontology Construction Tool with OWLDODDLE-OWL: A Domain Ontology Construction Tool with OWL
DODDLE-OWL: A Domain Ontology Construction Tool with OWLTakeshi Morita
 
HYPONYMY EXTRACTION OF DOMAIN ONTOLOGY CONCEPT BASED ON CCRFS AND HIERARCHY C...
HYPONYMY EXTRACTION OF DOMAIN ONTOLOGY CONCEPT BASED ON CCRFS AND HIERARCHY C...HYPONYMY EXTRACTION OF DOMAIN ONTOLOGY CONCEPT BASED ON CCRFS AND HIERARCHY C...
HYPONYMY EXTRACTION OF DOMAIN ONTOLOGY CONCEPT BASED ON CCRFS AND HIERARCHY C...dannyijwest
 
Hyponymy extraction of domain ontology
Hyponymy extraction of domain ontologyHyponymy extraction of domain ontology
Hyponymy extraction of domain ontologyIJwest
 
Semantic technology in nutshell 2013. Semantic! are you a linguist?
Semantic technology in nutshell 2013. Semantic! are you a linguist?Semantic technology in nutshell 2013. Semantic! are you a linguist?
Semantic technology in nutshell 2013. Semantic! are you a linguist?Heimo Hänninen
 
Fueling the future with Semantic Web patterns - Keynote at WOP2014@ISWC
Fueling the future with Semantic Web patterns - Keynote at WOP2014@ISWCFueling the future with Semantic Web patterns - Keynote at WOP2014@ISWC
Fueling the future with Semantic Web patterns - Keynote at WOP2014@ISWCValentina Presutti
 

Similar to Different Semantic Perspectives for Question Answering Systems (20)

Effective Semantics for Engineering NLP Systems
Effective Semantics for Engineering NLP SystemsEffective Semantics for Engineering NLP Systems
Effective Semantics for Engineering NLP Systems
 
The Role Of Ontology In Modern Expert Systems Dallas 2008
The Role Of Ontology In Modern Expert Systems   Dallas   2008The Role Of Ontology In Modern Expert Systems   Dallas   2008
The Role Of Ontology In Modern Expert Systems Dallas 2008
 
AI Beyond Deep Learning
AI Beyond Deep LearningAI Beyond Deep Learning
AI Beyond Deep Learning
 
Building AI Applications using Knowledge Graphs
Building AI Applications using Knowledge GraphsBuilding AI Applications using Knowledge Graphs
Building AI Applications using Knowledge Graphs
 
Using construction grammar in conversational systems
Using construction grammar in conversational systemsUsing construction grammar in conversational systems
Using construction grammar in conversational systems
 
Explanations in Dialogue Systems through Uncertain RDF Knowledge Bases
Explanations in Dialogue Systems through Uncertain RDF Knowledge BasesExplanations in Dialogue Systems through Uncertain RDF Knowledge Bases
Explanations in Dialogue Systems through Uncertain RDF Knowledge Bases
 
Knowledge representation Problem in AI.pptx
Knowledge representation Problem in AI.pptxKnowledge representation Problem in AI.pptx
Knowledge representation Problem in AI.pptx
 
09- Syed Rehan-ai-ppt2.pptx
09- Syed Rehan-ai-ppt2.pptx09- Syed Rehan-ai-ppt2.pptx
09- Syed Rehan-ai-ppt2.pptx
 
How To Make Linked Data More than Data
How To Make Linked Data More than DataHow To Make Linked Data More than Data
How To Make Linked Data More than Data
 
How To Make Linked Data More than Data
How To Make Linked Data More than DataHow To Make Linked Data More than Data
How To Make Linked Data More than Data
 
ONTOLOGY BASED DATA ACCESS
ONTOLOGY BASED DATA ACCESSONTOLOGY BASED DATA ACCESS
ONTOLOGY BASED DATA ACCESS
 
Jarrar: ORM in Description Logic
Jarrar: ORM in Description Logic  Jarrar: ORM in Description Logic
Jarrar: ORM in Description Logic
 
Frame-Script and Predicate logic.pptx
Frame-Script and Predicate logic.pptxFrame-Script and Predicate logic.pptx
Frame-Script and Predicate logic.pptx
 
Rasa NLU and ML Interpretability
Rasa NLU and ML InterpretabilityRasa NLU and ML Interpretability
Rasa NLU and ML Interpretability
 
Toward The Semantic Deep Web
Toward The Semantic Deep WebToward The Semantic Deep Web
Toward The Semantic Deep Web
 
DODDLE-OWL: A Domain Ontology Construction Tool with OWL
DODDLE-OWL: A Domain Ontology Construction Tool with OWLDODDLE-OWL: A Domain Ontology Construction Tool with OWL
DODDLE-OWL: A Domain Ontology Construction Tool with OWL
 
HYPONYMY EXTRACTION OF DOMAIN ONTOLOGY CONCEPT BASED ON CCRFS AND HIERARCHY C...
HYPONYMY EXTRACTION OF DOMAIN ONTOLOGY CONCEPT BASED ON CCRFS AND HIERARCHY C...HYPONYMY EXTRACTION OF DOMAIN ONTOLOGY CONCEPT BASED ON CCRFS AND HIERARCHY C...
HYPONYMY EXTRACTION OF DOMAIN ONTOLOGY CONCEPT BASED ON CCRFS AND HIERARCHY C...
 
Hyponymy extraction of domain ontology
Hyponymy extraction of domain ontologyHyponymy extraction of domain ontology
Hyponymy extraction of domain ontology
 
Semantic technology in nutshell 2013. Semantic! are you a linguist?
Semantic technology in nutshell 2013. Semantic! are you a linguist?Semantic technology in nutshell 2013. Semantic! are you a linguist?
Semantic technology in nutshell 2013. Semantic! are you a linguist?
 
Fueling the future with Semantic Web patterns - Keynote at WOP2014@ISWC
Fueling the future with Semantic Web patterns - Keynote at WOP2014@ISWCFueling the future with Semantic Web patterns - Keynote at WOP2014@ISWC
Fueling the future with Semantic Web patterns - Keynote at WOP2014@ISWC
 

More from Andre Freitas

AI & Scientific Discovery in Oncology: Opportunities, Challenges & Trends
AI & Scientific Discovery in Oncology: Opportunities, Challenges & TrendsAI & Scientific Discovery in Oncology: Opportunities, Challenges & Trends
AI & Scientific Discovery in Oncology: Opportunities, Challenges & TrendsAndre Freitas
 
AI Systems @ Manchester
AI Systems @ ManchesterAI Systems @ Manchester
AI Systems @ ManchesterAndre Freitas
 
SemEval-2017 Task 5: Fine-Grained Sentiment Analysis on Financial Microblogs ...
SemEval-2017 Task 5: Fine-Grained Sentiment Analysis on Financial Microblogs ...SemEval-2017 Task 5: Fine-Grained Sentiment Analysis on Financial Microblogs ...
SemEval-2017 Task 5: Fine-Grained Sentiment Analysis on Financial Microblogs ...Andre Freitas
 
A Semantic Web Platform for Automating the Interpretation of Finite Element ...
A Semantic Web Platform for Automating the Interpretation of Finite Element ...A Semantic Web Platform for Automating the Interpretation of Finite Element ...
A Semantic Web Platform for Automating the Interpretation of Finite Element ...Andre Freitas
 
How Semantic Technologies can help to cure Hearing Loss?
How Semantic Technologies can help to cure Hearing Loss?How Semantic Technologies can help to cure Hearing Loss?
How Semantic Technologies can help to cure Hearing Loss?Andre Freitas
 
Towards a Distributional Semantic Web Stack
Towards a Distributional Semantic Web StackTowards a Distributional Semantic Web Stack
Towards a Distributional Semantic Web StackAndre Freitas
 
Talking to your Data: Natural Language Interfaces for a schema-less world (Ke...
Talking to your Data: Natural Language Interfaces for a schema-less world (Ke...Talking to your Data: Natural Language Interfaces for a schema-less world (Ke...
Talking to your Data: Natural Language Interfaces for a schema-less world (Ke...Andre Freitas
 
Introduction to Distributional Semantics
Introduction to Distributional SemanticsIntroduction to Distributional Semantics
Introduction to Distributional SemanticsAndre Freitas
 
On the Semantic Representation and Extraction of Complex Category Descriptors
On the Semantic Representation and Extraction of Complex Category DescriptorsOn the Semantic Representation and Extraction of Complex Category Descriptors
On the Semantic Representation and Extraction of Complex Category DescriptorsAndre Freitas
 
Coping with Data Variety in the Big Data Era: The Semantic Computing Approach
Coping with Data Variety in the Big Data Era: The Semantic Computing ApproachCoping with Data Variety in the Big Data Era: The Semantic Computing Approach
Coping with Data Variety in the Big Data Era: The Semantic Computing ApproachAndre Freitas
 
Question Answering over Linked Data: Challenges, Approaches & Trends (Tutoria...
Question Answering over Linked Data: Challenges, Approaches & Trends (Tutoria...Question Answering over Linked Data: Challenges, Approaches & Trends (Tutoria...
Question Answering over Linked Data: Challenges, Approaches & Trends (Tutoria...Andre Freitas
 
Natural Language Queries over Heterogeneous Linked Data Graphs: A Distributio...
Natural Language Queries over Heterogeneous Linked Data Graphs: A Distributio...Natural Language Queries over Heterogeneous Linked Data Graphs: A Distributio...
Natural Language Queries over Heterogeneous Linked Data Graphs: A Distributio...Andre Freitas
 
A Compositional-distributional Semantic Model over Structured Data
A Compositional-distributional Semantic Model over Structured DataA Compositional-distributional Semantic Model over Structured Data
A Compositional-distributional Semantic Model over Structured DataAndre Freitas
 

More from Andre Freitas (13)

AI & Scientific Discovery in Oncology: Opportunities, Challenges & Trends
AI & Scientific Discovery in Oncology: Opportunities, Challenges & TrendsAI & Scientific Discovery in Oncology: Opportunities, Challenges & Trends
AI & Scientific Discovery in Oncology: Opportunities, Challenges & Trends
 
AI Systems @ Manchester
AI Systems @ ManchesterAI Systems @ Manchester
AI Systems @ Manchester
 
SemEval-2017 Task 5: Fine-Grained Sentiment Analysis on Financial Microblogs ...
SemEval-2017 Task 5: Fine-Grained Sentiment Analysis on Financial Microblogs ...SemEval-2017 Task 5: Fine-Grained Sentiment Analysis on Financial Microblogs ...
SemEval-2017 Task 5: Fine-Grained Sentiment Analysis on Financial Microblogs ...
 
A Semantic Web Platform for Automating the Interpretation of Finite Element ...
A Semantic Web Platform for Automating the Interpretation of Finite Element ...A Semantic Web Platform for Automating the Interpretation of Finite Element ...
A Semantic Web Platform for Automating the Interpretation of Finite Element ...
 
How Semantic Technologies can help to cure Hearing Loss?
How Semantic Technologies can help to cure Hearing Loss?How Semantic Technologies can help to cure Hearing Loss?
How Semantic Technologies can help to cure Hearing Loss?
 
Towards a Distributional Semantic Web Stack
Towards a Distributional Semantic Web StackTowards a Distributional Semantic Web Stack
Towards a Distributional Semantic Web Stack
 
Talking to your Data: Natural Language Interfaces for a schema-less world (Ke...
Talking to your Data: Natural Language Interfaces for a schema-less world (Ke...Talking to your Data: Natural Language Interfaces for a schema-less world (Ke...
Talking to your Data: Natural Language Interfaces for a schema-less world (Ke...
 
Introduction to Distributional Semantics
Introduction to Distributional SemanticsIntroduction to Distributional Semantics
Introduction to Distributional Semantics
 
On the Semantic Representation and Extraction of Complex Category Descriptors
On the Semantic Representation and Extraction of Complex Category DescriptorsOn the Semantic Representation and Extraction of Complex Category Descriptors
On the Semantic Representation and Extraction of Complex Category Descriptors
 
Coping with Data Variety in the Big Data Era: The Semantic Computing Approach
Coping with Data Variety in the Big Data Era: The Semantic Computing ApproachCoping with Data Variety in the Big Data Era: The Semantic Computing Approach
Coping with Data Variety in the Big Data Era: The Semantic Computing Approach
 
Question Answering over Linked Data: Challenges, Approaches & Trends (Tutoria...
Question Answering over Linked Data: Challenges, Approaches & Trends (Tutoria...Question Answering over Linked Data: Challenges, Approaches & Trends (Tutoria...
Question Answering over Linked Data: Challenges, Approaches & Trends (Tutoria...
 
Natural Language Queries over Heterogeneous Linked Data Graphs: A Distributio...
Natural Language Queries over Heterogeneous Linked Data Graphs: A Distributio...Natural Language Queries over Heterogeneous Linked Data Graphs: A Distributio...
Natural Language Queries over Heterogeneous Linked Data Graphs: A Distributio...
 
A Compositional-distributional Semantic Model over Structured Data
A Compositional-distributional Semantic Model over Structured DataA Compositional-distributional Semantic Model over Structured Data
A Compositional-distributional Semantic Model over Structured Data
 

Recently uploaded

TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...
TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...
TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...ssifa0344
 
Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...
Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...
Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...Sérgio Sacani
 
Bacterial Identification and Classifications
Bacterial Identification and ClassificationsBacterial Identification and Classifications
Bacterial Identification and ClassificationsAreesha Ahmad
 
Isotopic evidence of long-lived volcanism on Io
Isotopic evidence of long-lived volcanism on IoIsotopic evidence of long-lived volcanism on Io
Isotopic evidence of long-lived volcanism on IoSérgio Sacani
 
SCIENCE-4-QUARTER4-WEEK-4-PPT-1 (1).pptx
SCIENCE-4-QUARTER4-WEEK-4-PPT-1 (1).pptxSCIENCE-4-QUARTER4-WEEK-4-PPT-1 (1).pptx
SCIENCE-4-QUARTER4-WEEK-4-PPT-1 (1).pptxRizalinePalanog2
 
Chemical Tests; flame test, positive and negative ions test Edexcel Internati...
Chemical Tests; flame test, positive and negative ions test Edexcel Internati...Chemical Tests; flame test, positive and negative ions test Edexcel Internati...
Chemical Tests; flame test, positive and negative ions test Edexcel Internati...ssuser79fe74
 
Botany 4th semester series (krishna).pdf
Botany 4th semester series (krishna).pdfBotany 4th semester series (krishna).pdf
Botany 4th semester series (krishna).pdfSumit Kumar yadav
 
GBSN - Biochemistry (Unit 1)
GBSN - Biochemistry (Unit 1)GBSN - Biochemistry (Unit 1)
GBSN - Biochemistry (Unit 1)Areesha Ahmad
 
Pests of cotton_Borer_Pests_Binomics_Dr.UPR.pdf
Pests of cotton_Borer_Pests_Binomics_Dr.UPR.pdfPests of cotton_Borer_Pests_Binomics_Dr.UPR.pdf
Pests of cotton_Borer_Pests_Binomics_Dr.UPR.pdfPirithiRaju
 
Forensic Biology & Its biological significance.pdf
Forensic Biology & Its biological significance.pdfForensic Biology & Its biological significance.pdf
Forensic Biology & Its biological significance.pdfrohankumarsinghrore1
 
Animal Communication- Auditory and Visual.pptx
Animal Communication- Auditory and Visual.pptxAnimal Communication- Auditory and Visual.pptx
Animal Communication- Auditory and Visual.pptxUmerFayaz5
 
Formation of low mass protostars and their circumstellar disks
Formation of low mass protostars and their circumstellar disksFormation of low mass protostars and their circumstellar disks
Formation of low mass protostars and their circumstellar disksSérgio Sacani
 
Kochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRL
Kochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRLKochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRL
Kochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRLkantirani197
 
Pests of mustard_Identification_Management_Dr.UPR.pdf
Pests of mustard_Identification_Management_Dr.UPR.pdfPests of mustard_Identification_Management_Dr.UPR.pdf
Pests of mustard_Identification_Management_Dr.UPR.pdfPirithiRaju
 
Biological Classification BioHack (3).pdf
Biological Classification BioHack (3).pdfBiological Classification BioHack (3).pdf
Biological Classification BioHack (3).pdfmuntazimhurra
 
Presentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptxPresentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptxgindu3009
 
GBSN - Microbiology (Unit 2)
GBSN - Microbiology (Unit 2)GBSN - Microbiology (Unit 2)
GBSN - Microbiology (Unit 2)Areesha Ahmad
 
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...Sérgio Sacani
 
Creating and Analyzing Definitive Screening Designs
Creating and Analyzing Definitive Screening DesignsCreating and Analyzing Definitive Screening Designs
Creating and Analyzing Definitive Screening DesignsNurulAfiqah307317
 
GUIDELINES ON SIMILAR BIOLOGICS Regulatory Requirements for Marketing Authori...
GUIDELINES ON SIMILAR BIOLOGICS Regulatory Requirements for Marketing Authori...GUIDELINES ON SIMILAR BIOLOGICS Regulatory Requirements for Marketing Authori...
GUIDELINES ON SIMILAR BIOLOGICS Regulatory Requirements for Marketing Authori...Lokesh Kothari
 

Recently uploaded (20)

TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...
TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...
TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...
 
Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...
Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...
Discovery of an Accretion Streamer and a Slow Wide-angle Outflow around FUOri...
 
Bacterial Identification and Classifications
Bacterial Identification and ClassificationsBacterial Identification and Classifications
Bacterial Identification and Classifications
 
Isotopic evidence of long-lived volcanism on Io
Isotopic evidence of long-lived volcanism on IoIsotopic evidence of long-lived volcanism on Io
Isotopic evidence of long-lived volcanism on Io
 
SCIENCE-4-QUARTER4-WEEK-4-PPT-1 (1).pptx
SCIENCE-4-QUARTER4-WEEK-4-PPT-1 (1).pptxSCIENCE-4-QUARTER4-WEEK-4-PPT-1 (1).pptx
SCIENCE-4-QUARTER4-WEEK-4-PPT-1 (1).pptx
 
Chemical Tests; flame test, positive and negative ions test Edexcel Internati...
Chemical Tests; flame test, positive and negative ions test Edexcel Internati...Chemical Tests; flame test, positive and negative ions test Edexcel Internati...
Chemical Tests; flame test, positive and negative ions test Edexcel Internati...
 
Botany 4th semester series (krishna).pdf
Botany 4th semester series (krishna).pdfBotany 4th semester series (krishna).pdf
Botany 4th semester series (krishna).pdf
 
GBSN - Biochemistry (Unit 1)
GBSN - Biochemistry (Unit 1)GBSN - Biochemistry (Unit 1)
GBSN - Biochemistry (Unit 1)
 
Pests of cotton_Borer_Pests_Binomics_Dr.UPR.pdf
Pests of cotton_Borer_Pests_Binomics_Dr.UPR.pdfPests of cotton_Borer_Pests_Binomics_Dr.UPR.pdf
Pests of cotton_Borer_Pests_Binomics_Dr.UPR.pdf
 
Forensic Biology & Its biological significance.pdf
Forensic Biology & Its biological significance.pdfForensic Biology & Its biological significance.pdf
Forensic Biology & Its biological significance.pdf
 
Animal Communication- Auditory and Visual.pptx
Animal Communication- Auditory and Visual.pptxAnimal Communication- Auditory and Visual.pptx
Animal Communication- Auditory and Visual.pptx
 
Formation of low mass protostars and their circumstellar disks
Formation of low mass protostars and their circumstellar disksFormation of low mass protostars and their circumstellar disks
Formation of low mass protostars and their circumstellar disks
 
Kochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRL
Kochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRLKochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRL
Kochi ❤CALL GIRL 84099*07087 ❤CALL GIRLS IN Kochi ESCORT SERVICE❤CALL GIRL
 
Pests of mustard_Identification_Management_Dr.UPR.pdf
Pests of mustard_Identification_Management_Dr.UPR.pdfPests of mustard_Identification_Management_Dr.UPR.pdf
Pests of mustard_Identification_Management_Dr.UPR.pdf
 
Biological Classification BioHack (3).pdf
Biological Classification BioHack (3).pdfBiological Classification BioHack (3).pdf
Biological Classification BioHack (3).pdf
 
Presentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptxPresentation Vikram Lander by Vedansh Gupta.pptx
Presentation Vikram Lander by Vedansh Gupta.pptx
 
GBSN - Microbiology (Unit 2)
GBSN - Microbiology (Unit 2)GBSN - Microbiology (Unit 2)
GBSN - Microbiology (Unit 2)
 
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
PossibleEoarcheanRecordsoftheGeomagneticFieldPreservedintheIsuaSupracrustalBe...
 
Creating and Analyzing Definitive Screening Designs
Creating and Analyzing Definitive Screening DesignsCreating and Analyzing Definitive Screening Designs
Creating and Analyzing Definitive Screening Designs
 
GUIDELINES ON SIMILAR BIOLOGICS Regulatory Requirements for Marketing Authori...
GUIDELINES ON SIMILAR BIOLOGICS Regulatory Requirements for Marketing Authori...GUIDELINES ON SIMILAR BIOLOGICS Regulatory Requirements for Marketing Authori...
GUIDELINES ON SIMILAR BIOLOGICS Regulatory Requirements for Marketing Authori...
 

Different Semantic Perspectives for Question Answering Systems

  • 1. NLP & Semantic Computing Group N L P Different Semantic Perspectives for Hybrid Question Answering Systems Andre Freitas University of Passau OKBQA, Jeju, 2016
  • 2. NLP & Semantic Computing Group http://www.slideshare.net/andrenfreitas These slides:
  • 3. NLP & Semantic Computing Group Outline  Multiple Perspectives of Semantic Representation  Lightweight Semantic Representation  Knowledge Graph Extraction from Text  Answering Queries with Knowledge Graphs  Reasoning  Take-away Message
  • 4. NLP & Semantic Computing Group Multiple Perspectives of Semantic Representation
  • 5. NLP & Semantic Computing Group QA & Semantics • Question Answering is about managing semantic representation, extraction, selection trade-offs. • And it is about integrating multiple components in a complex approach. • Semantic best-effort, systems tolerant to noisy, inconsistent, vague, data. 5
  • 6. NLP & Semantic Computing Group  “Most semantic models have dealt with particular types of constructions, and have been carried out under very simplifying assumptions, in true lab conditions.”  “If these idealizations are removed it is not clear at all that modern semantics can give a full account of all but the simplest models/statements.” Formal World Real World Baroni et al. 2013 Semantics for a Complex World 6
  • 7. NLP & Semantic Computing Group Why Not RDF? • Follows a more “database-type” of representation perspective. • Gap towards representing text.
  • 8. NLP & Semantic Computing Group Choices of Semantic Representation • Logical • Frames: verbs | nouns • Binary relations: binary | n-ary • Named entities • Language Models • Syntactic structures • Bag-of-words Concept-level representation Background knowledge Extraction complexity 8
  • 9. NLP & Semantic Computing Group Information Extraction • Logical • Frames:  verbs | nouns • Binary relations:  binary | n-ary • Named entities • Syntactic Structures & LMs • Bag-of-words • Semantic parsing • Semantic role labeling • Relation extraction: – closed/open • Named entity recognition • Syntactic/N-gram Parsing • Indexing9
  • 10. NLP & Semantic Computing Group Information Extraction • Logical • Frames:  verbs | nouns • Binary relations:  binary | n-ary • Named entities • Syntactic Structures & LMs • Bag-of-words • Semantic parsing • Semantic role labeling • Relation extraction: – closed/open • Named entity recognition • Syntactic/N-gram Parsing • Indexing Use all of them!
  • 11. NLP & Semantic Computing Group Representation focal points • Types of knowledge to focus at the representation:  Facts vs Definitions vs Opinions  Temporality  Spatiality  Modality  Polarity  Rhetorical structures  … 11
  • 12. NLP & Semantic Computing Group Lightweight Semantic Representation
  • 13. NLP & Semantic Computing Group Objective • Provide a lightweight knowledge representation model which:  Can represent textual discourse information. • Maximizes the capture of textual information.  Is convenient to extract from text.  Is convenient to access (query and browse). 13
  • 14. NLP & Semantic Computing Group Lightweight Semantic Representation Representing Texts as Contextualized Entity-Centric Linked Data Graphs, WebS 2013
  • 15. NLP & Semantic Computing Group Representation Assumptions • Data integration:  Named entities (instances)  Abstract classes (unary predicates) • Rich taxonomical structures. • Context representation as a first class citizen. • Open vocabulary. • Word instead of sense/concept. 15
  • 16. NLP & Semantic Computing Group Representation Assumptions • Data integration:  Named entities (instances)  Abstract classes (unary predicates) • Rich taxonomical structures. • Context representation as a first class citizen. • Open vocabulary. • Word instead of sense/concept. 16
  • 17. NLP & Semantic Computing Group Representation of Complex Relations General Electric Company, or GE , is an American multinational conglomerate corporation incorporated in Schenectady , New York 17
  • 18. NLP & Semantic Computing Group Data Integration points General Electric Company, or GE , is an American multinational conglomerate corporation incorporated in Schenectady , New York Named entities are lower entropy integration points Pivot points18
  • 19. NLP & Semantic Computing Group Data Integration points General Electric Company, or GE , is an American multinational conglomerate corporation incorporated in Schenectady , New York Named entities are also low entropy entry points for answering queries Pivot points19
  • 20. NLP & Semantic Computing Group Data Integration points General Electric Company, or GE , is an American multinational conglomerate corporation incorporated in Schenectady , New York Also abstract classes … Pivot points20
  • 21. NLP & Semantic Computing Group Data Integration points General Electric Company, or GE , is an American multinational conglomerate corporation incorporated in Schenectady , New York They are also a very convenient way to represent. Pivot points21
  • 22. NLP & Semantic Computing Group Representation Assumptions • Data integration:  Named entities (instances)  Abstract classes (unary predicates) • Rich taxonomical structures. • Context representation as a first class citizen. • Open vocabulary. • Word instead of sense/concept. 22
  • 23. NLP & Semantic Computing Group Taxonomy Extraction  Are predicates with more complex compositional patterns which describe sets.  Parsing complex nominals. American multinational conglomerate corporation On the Semantic Representation and Extraction of Complex Category Descriptors, NLDB 2014 multinational conglomerate corporation corporation conglomerate corporation is a is a is a Pivot points
  • 24. NLP & Semantic Computing Group Representation Assumptions • Data integration:  Named entities (instances)  Abstract classes (unary predicates) • Rich taxonomical structures. • Context representation as a first class citizen. • Open vocabulary. • Word instead of sense/concept. 24
  • 25. NLP & Semantic Computing Group Context Representation General Electric Company, or GE , is an American multinational conglomerate corporation incorporated in Schenectady , New York Reification as a first class representation element 25
  • 26. NLP & Semantic Computing Group Context Representation General Electric Company, or GE , is an American multinational conglomerate corporation incorporated in Schenectady , New York Temporality, spatiality, modality, rhetorical relations … 26
  • 27. NLP & Semantic Computing Group Rhetorical Structures using Reification • cause:  e.g. “because scraping the bottom with a metal utensil will scratch the surface.” • circumstance  e.g. “After completing your operating system reinstallation,” • concession  e.g. “Although the hotel is situated adjacent to a beach,” • condition  e.g. “If you can break the $ 1000 dollar investment range,” • contrast  e.g. “but you can do better with 2.4ghz or 900mhz phones.” • purpose  e.g.“in order for the rear passengers to get in the vehicle.” • … 27
  • 28. NLP & Semantic Computing Group Representation Assumptions • Data integration:  Named entities (instances)  Abstract classes (unary predicates) • Rich taxonomical structures. • Context representation as a first class citizen. • Open vocabulary. • Word instead of sense/concept. 28
  • 29. NLP & Semantic Computing Group Open Vocabulary General Electric Company, or GE , is an American multinational conglomerate corporation incorporated in Schenectady , New York Temporality, spatiality, modality, rhetorical relations … 29
  • 30. NLP & Semantic Computing Group Open Vocabulary • Easier to extract but difficult to consume. • We pay the price at query time. • How to operate over a large-scale semantically heterogeneous knowledge-graphs? 30
  • 31. NLP & Semantic Computing Group Representation Assumptions • Data integration:  Named entities (instances)  Abstract classes (unary predicates) • Rich taxonomical structures. • Context representation as a first class citizen. • Open vocabulary. • Word instead of sense/concept. 31
  • 32. NLP & Semantic Computing Group Words instead of Senses • Motivation: Disambiguation is a tough problem. • Sense granularity can be, at many situations, arbitrary (too context dependent). • We treat a word as a superposition of senses, almost in a “quantum mechanical sense”. 32
  • 33. NLP & Semantic Computing Group SenseSuperposition Coecke et al. (2010): Category theory and Lambek calculus.
  • 34. NLP & Semantic Computing Group Revisited RDF (for Representing Texts) • Data Model Types: Instance, Class, Property… • RDFS: Taxonomic representation. • Reification for contextual relations (subordinations). • Blank nodes for n-ary relations. • Triple. • Labels over URIs. 34
  • 35. NLP & Semantic Computing Group Abstract Meaning Representations – AMR, Maximal Use of PropBank Frame Files Alternative Representations
  • 36. NLP & Semantic Computing Group Distributional Semantics
  • 37. NLP & Semantic Computing Group Distributional Semantic Models  Semantic Model with low acquisition effort (automatically built from text) Simplification of the representation  Enables the construction of comprehensive commonsense/semantic KBs  What is the cost? Some level of noise (semantic best-effort) Limited semantic model 37
  • 38. NLP & Semantic Computing Group Distributional Semantics as Commonsense Knowledge Commonsense is here θ car dog cat bark run leash Semantic Approximation is here 38
  • 39. NLP & Semantic Computing Group I find it rather odd that people are already trying to tie the Commission's hands in relation to the proposal for a directive, while at the same calling on it to present a Green Paper on the current situation with regard to optional and supplementary health insurance schemes. I find it a little strange to now obliging the Commission to a motion for a resolution and to ask him at the same time to draw up a Green Paper on the current state of voluntary insurance and supplementary sickness insurance. =? Beyond Single Word Vector Models: Compositionality
  • 40. NLP & Semantic Computing Group Compositional Semantics  Can we extend DS to account for the meaning of phrases and sentences?  Compositionality: The meaning of a complex expression is a function of the meaning of its constituent parts.
  • 41. NLP & Semantic Computing Group Compositional Semantics Words in which the meaning is directly determined by their distributional behaviour (e.g., nouns). Words that act as functions transforming the distributional profile of other words (e.g., verbs, adjectives, …). 41
  • 42. NLP & Semantic Computing Group Compositional-Distributional Semantics
  • 43. NLP & Semantic Computing Group Recursive Neural Networks for Structure Prediction 43
  • 44. NLP & Semantic Computing Group New Model: Recursive Neural Tensor Network • Goal: Function that composes two vectors. • More expressive than any other RNN so far. 44 Socher et al.
  • 45. NLP & Semantic Computing Group Socher et al.
  • 46. NLP & Semantic Computing Group Compositional-distributional model for Categories 46
  • 47. NLP & Semantic Computing Group Embedding Knowledge Graphs 47
  • 48. NLP & Semantic Computing Group The vector space is segmented48 Dimensional reduction mechanism! A Distributional Structured Semantic Space for Querying RDF Graph Data, IJSC 2012
  • 49. NLP & Semantic Computing Group Compositional-distributional model for paraphrases A Compositional-Distributional Semantic Model for Searching Complex Entity Categories, *SEM (2016)
  • 50. NLP & Semantic Computing Group Knowledge Graph Extraction from Text
  • 51. NLP & Semantic Computing Group Graphene
  • 52. NLP & Semantic Computing Group Graph Extraction Pipeline Text Transformation N-ary Relation Extraction Text Simplification Graph Serialization Taxonomy Extraction Storage RST Classification ML-based Rule-based Rule-based ML-based 52
  • 53. NLP & Semantic Computing Group Minimalistic Text Transformations Text Transformation N-ary Relation Extraction Text Simplification Graph Serialization Taxonomy Extraction Storage RST Classification ML-based Rule-based Rule-based ML-based 53
  • 54. NLP & Semantic Computing Group Minimalistic Text Transformations • Co-reference Resolution  Pronominal co-references. • Passive  We have been approached by the investment banker.  The investment banker approached us. • Genitive modifier  Malaysia's crude palm oil output is estimated to have risen.  The crude palm oil output of Malasia is estimated to have risen. 54
  • 55. NLP & Semantic Computing Group Text Simplification Text Transformation N-ary Relation Extraction Text Simplification Graph Serialization Taxonomy Extraction Storage RST Classification ML-based Rule-based Rule-based ML-based 55
  • 56. NLP & Semantic Computing Group Text Simplification for KG Extraction “Defeating Republican nominee Mitt Romney, Obama, who was the first African American to hold the office, was reelected president in November 2012.”  relations are spread across clauses  relations are presented in non-canonical form 56
  • 57. NLP & Semantic Computing Group Text Simplification for KG Extraction • Insertion of a text simplification step  Obama was reelected president in November 2012.  Obama was the first African American to hold the office.  Obama was defeating Mitt Romney.  Mitt Romney was Republican nominee. 57
  • 58. NLP & Semantic Computing Group Syntax-driven sentence simplification approach Task: • Reduce the linguistic complexity of a text while retaining the original information/meaning using a set of syntax-based rewrite operations (deletion, insertion, reordering, sentence splitting). Idea: • Simplify a sentence by separating out components that supply only secondary information into simpler stand-alone context sentences, thus yielding one or more reduced core sentences.
  • 59. NLP & Semantic Computing Group Approach • Linguistic analysis of sentences from the English Wikipedia to identify constructs which provide only secondary information: • non-restrictive relative clauses • non-restrictive and restrictive appositive phrases • participial phrases offset by commas • adjective and adverb phrases delimited by punctuation • particular prepositional phrases • lead noun phrases • intra-sentential attributions • parentheticals • conjoined clauses with specific features • particular punctuation •Rule-based simplification rules. Improving Relation Extraction by Syntax-based Sentence Simplification (2016)
  • 60. NLP & Semantic Computing Group N-ary Relation Extraction Text Transformation N-ary Relation Extraction Text Simplification Graph Serialization Taxonomy Extraction Storage RST Classification Rule-based Rule-based ML-based 60 OpenIE, University of Washington
  • 61. NLP & Semantic Computing Group Taxonomy Extraction Text Transformation N-ary Relation Extraction Text Simplification Graph Serialization Taxonomy Extraction Storage RST Classification Rule-based Rule-based ML-based 61 Representation and Extraction of Complex Category Descriptors, NLDB 2014
  • 62. NLP & Semantic Computing Group RST Classification Text Transformation N-ary Relation Extraction Text Simplification Graph Serialization Taxonomy Extraction Storage RST Classification Rule-based Rule-based ML-based 62
  • 63. NLP & Semantic Computing Group Rhetorical Structure Extraction 63 TEXT-LEVEL RST-STYLE DISCOURSE PARSER (Feng and Hirst, 2012) Structure classification Relation classification
  • 64. NLP & Semantic Computing Group Answering Queries with Knowledge Graphs
  • 65. NLP & Semantic Computing Group Now our graph supports semantic approximations as a first-class operation
  • 66. NLP & Semantic Computing Group Approach Overview Query Planner Ƭ-Space (embedding graphs) Wikipedia Commonsense knowledge RDF Explicit Semantic Analysis Core semantic approximation & composition operations Query AnalysisQuery Query Features Query Plan 66
  • 67. NLP & Semantic Computing Group Core Principles  Minimize the impact of Ambiguity, Vagueness, Synonymy.  Address the simplest matchings first (semantic pivoting).  Semantic Relatedness as a primitive operation.  Distributional semantics models as commonsense knowledge representation.  Lightweight syntactic constraints. 67
  • 68. NLP & Semantic Computing Group • Step 2: Query NER  Rules-based: POS Tag + IDF Who is the daughter of Bill Clinton married to? (PROBABLY AN INSTANCE) Query Pre-Processing (Question Analysis) 68
  • 69. NLP & Semantic Computing Group • Step 3: Determine answer type  Rules-based. Who is the daughter of Bill Clinton married to? (PERSON) Query Pre-Processing (Question Analysis) 69
  • 70. NLP & Semantic Computing Group • Transform natural language queries into a pseudo-logical form. “Who is the daughter of Bill Clinton married to?” Query Pre-Processing (Question Analysis) 70
  • 71. NLP & Semantic Computing Group Query Pre-Processing (Question Analysis) Bill Clinton daughter married to (INSTANCE) Person ANSWER TYPE QUESTION FOCUS 71 • Step 5: Determine the query pattern  Rules based. • Remove stop words. • Merge words into entities. • Reorder structure from core entity position.
  • 72. NLP & Semantic Computing Group • Step 5: Determine the query pattern  Rules based. • Remove stop words. • Merge words into entities. • Reorder structure from core entity position. Query Pre-Processing (Question Analysis) Bill Clinton daughter married to (INSTANCE) Person (PREDICATE) (PREDICATE) Query Features 72
  • 73. NLP & Semantic Computing Group • Map query features into a query plan. • A query plan contains a sequence of:  Search operations.  Navigation operations. Query Planning (INSTANCE) (PREDICATE) (PREDICATE) Query Features  (1) INSTANCE SEARCH (Bill Clinton)  (2) DISAMBIGUATE ENTITY TYPE  (3) GENERATE ENTITY FACETS  (4) p1 <- SEARCH RELATED PREDICATE (Bill Clintion, daughter)  (5) e1 <- GET ASSOCIATED ENTITIES (Bill Clintion, p1)  (6) p2 <- SEARCH RELATED PREDICATE (e1, married to)  (7) e2 <- GET ASSOCIATED ENTITIES (e1, p2)  (8) POST PROCESS (Bill Clintion, e1, p1, e2, p2) Query Plan 73
  • 74. NLP & Semantic Computing Group Core Entity Search Bill Clinton daughter married to Person :Bill_Clinton Query: KB: Entity search 74
  • 75. NLP & Semantic Computing Group Distributional Semantic Search Bill Clinton daughter married to Person :Bill_Clinton Query: :Chelsea_Clinton :child :Baptists :religion :Yale_Law_School :almaMater ... (PIVOT ENTITY) (ASSOCIATED TRIPLES) 75 KB:
  • 76. NLP & Semantic Computing Group Distributional Semantic Search Bill Clinton daughter married to Person :Bill_Clinton Query: :Chelsea_Clinton :child :Baptists :religion :Yale_Law_School :almaMater ... sem_rel(daughter,child)=0.054 sem_rel(daughter,child)=0.004 sem_rel(daughter,alma mater)=0.001 Which properties are semantically related to ‘daughter’? 76 KB:
  • 77. NLP & Semantic Computing Group Distributional Semantic Search Bill Clinton daughter married to Person :Bill_Clinton Query: :Chelsea_Clinton :child 77 KB:
  • 78. NLP & Semantic Computing Group Distributional Semantic Search Bill Clinton daughter married to Person :Bill_Clinton Query: :Chelsea_Clinton :child (PIVOT ENTITY) 78 KB:
  • 79. NLP & Semantic Computing Group Distributional Semantic Search Bill Clinton daughter married to Person :Bill_Clinton Query: :Chelsea_Clinton :child :Mark_Mezvinsky :spouse 79 KB: Note the lazy disambiguation
  • 80. NLP & Semantic Computing Group 80
  • 81. NLP & Semantic Computing Group What is the highest mountain? Second Query Example (CLASS) (OPERATOR) Query Features mountain - highest PODS 81
  • 82. NLP & Semantic Computing Group Entity Search Mountain highest :Mountain Query: :typeOf (PIVOT ENTITY) 82 KB:
  • 83. NLP & Semantic Computing Group Extensional Expansion Mountain highest :Mountain Query: :Everest :typeOf (PIVOT ENTITY) :K2:typeOf ... 83 KB:
  • 84. NLP & Semantic Computing Group Distributional Semantic Matching Mountain highest :Mountain Query: :Everest :typeOf (PIVOT ENTITY) :K2:typeOf ... :elevation :location ... :deathPlaceOf 84 KB:
  • 85. NLP & Semantic Computing Group Get all numerical values Mountain highest :Mountain Query: :Everest :typeOf (PIVOT ENTITY) :K2:typeOf ... :elevation :elevation 8848 m 8611 m 85 KB:
  • 86. NLP & Semantic Computing Group Apply operator functional definition Mountain highest :Mountain Query: :Everest :typeOf (PIVOT ENTITY) :K2:typeOf ... :elevation :elevation 8848 m 8611 m SORT TOP_MOST 86 KB:
  • 87. NLP & Semantic Computing Group Results 87
  • 88. NLP & Semantic Computing Group StarGraph • Open source NoSQL platform for building and interacting with large and sparse knowledge graphs. • Semantic approximation as a built-in operation. • Scalable query execution performance. 88
  • 89. NLP & Semantic Computing Group Heuristics for the selection of the semantic pivot is critical! • Discussed here just superficially:  Information-theoretical justification. How hard is the Query? Measuring the Semantic Complexity of Schema- Agnostic Queries, IWCS (2015). Schema-agnositc queries over large-schema databases: a distributional semantics approach, PhD Thesis (2015). On the Semantic Mapping of Schema-agnostic Queries: A Preliminary Study, NLIWoD (2015). 89
  • 90. NLP & Semantic Computing Group Indra  Multilingual platform for experimentation with different word vector models. "Indra's net" is the net of the Vedic god Indra, whose net hangs over his palace on Mount Meru, the axis mundi of Hindu cosmology and Hindu mythology. Indra's net has a multifaceted jewel at each vertex, and each jewel is reflected in all of the other jewels. In the Avatamsaka Sutra, the image of "Indra's net" is used to describe the interconnectedness of the universe.
  • 91. NLP & Semantic Computing Group
  • 92. NLP & Semantic Computing Group Indra
  • 93. NLP & Semantic Computing Group
  • 94. NLP & Semantic Computing Group Bridging Structured & Unstructured Data • NER + Text + Passage Retrieval Ranking  Simple and powerful QA basis. • Lazy disambiguation. 94
  • 95. NLP & Semantic Computing Group Treo Answers Jeopardy Queries (Video)
  • 96. NLP & Semantic Computing Group Ranking Candidate Answers • But what if there are multiple candidate answers! Q: Who was Queen Victoria’s second son? • Answer Type: Person • Passage: The Marie biscuit is named after Marie Alexandrovna, the daughter of Czar Alexander II of Russia and wife of Alfred, the second son of Queen Victoria and Prince Albert 96Dan Jurafky’s slides
  • 97. NLP & Semantic Computing Group Ranking Candidate Answers • But what if there are multiple candidate answers! Q: Who was Queen Victoria’s second son? • Answer Type: Person • Passage: The Marie biscuit is named after Marie Alexandrovna, the daughter of Czar Alexander II of Russia and wife of Alfred, the second son of Queen Victoria and Prince Albert 97Dan Jurafky’s slides
  • 98. NLP & Semantic Computing Group Feature Engineering The Marie biscuit is named after Marie Alexandrovna, the daughter of Czar Alexander II of Russia and wife of Alfred, the second son of Queen Victoria and Prince Albert followed by a ‘,’ followed by an apposition Who was Queen Victoria’s second son? contains an entity in the query has a four-word overlap type = PERSON matches AnswerType 98
  • 99. NLP & Semantic Computing Group Propositionalisation e0 followedBy(,) followedByAppositionContainingQueryEntities() answer … Alfred true true true … … … passage entity (e0) entity (en) … The Marie biscuit is named after Marie Alexandrovna, the daughter of Czar Alexander II of Russia and wife of Alfred, the second son of Queen Victoria and Prince Albert answer 99
  • 100. NLP & Semantic Computing Group Reasoning for Text Entailment
  • 101. NLP & Semantic Computing Group Beyond Word Vector Models give birth mother car θ Distributional semantics can give us a hint about the concepts’ semantic proximity... ...but it still can’t tell us what exactly the relationship between them is give birth mother ???
  • 102. NLP & Semantic Computing Group Beyond Word Vector Models give birth mother ??? give birth mother ???
  • 103. NLP & Semantic Computing Group Beyond Word Vector Models: Intensional Reasoning  Representing structured intensional-level knowledge.  Creation of an intensional-level reasoning model. 104
  • 104. NLP & Semantic Computing Group Commonsense Reasoning  Selective (focussed) reasoning  - Selecting the relevant facts in the context of the inference  Reducing the search space. Scalability
  • 105. NLP & Semantic Computing Group Extended WordNet (XWN)
  • 106. NLP & Semantic Computing Group Commonsense Data (ConceptNet) http://conceptnet5.media.mit.edu/ 107
  • 107. NLP & Semantic Computing Group Distributional semantic relatedness as a Selectivity Heuristics Distributional heuristics 108 target source answer
  • 108. NLP & Semantic Computing Group Distributional semantic relatedness as a Selectivity Heuristics Distributional heuristics 109 target source answer
  • 109. NLP & Semantic Computing Group Distributional semantic relatedness as a Selectivity Heuristics Distributional heuristics 110 target source answer
  • 110. NLP & Semantic Computing Group John Smith EngineerInstance-level occupation Does John Smith have a degree?
  • 111. NLP & Semantic Computing Group
  • 112. NLP & Semantic Computing Group
  • 113. NLP & Semantic Computing Group
  • 114. NLP & Semantic Computing Group
  • 115. NLP & Semantic Computing Group
  • 116. NLP & Semantic Computing Group
  • 117. NLP & Semantic Computing Group
  • 118. NLP & Semantic Computing Group
  • 119. NLP & Semantic Computing Group
  • 120. NLP & Semantic Computing Group
  • 121. NLP & Semantic Computing Group
  • 122. NLP & Semantic Computing Group
  • 123. NLP & Semantic Computing Group
  • 124. NLP & Semantic Computing Group
  • 125. NLP & Semantic Computing Group
  • 126. NLP & Semantic Computing Group A Distributional Semantics Approach for Selective Reasoning on Commonsense Graph Knowledge Bases, NLDB (2015).
  • 127. NLP & Semantic Computing Group Bringing it into the Real World
  • 128. NLP & Semantic Computing Group Semeval 2017
  • 129. NLP & Semantic Computing Group Take-away Message • Choosing the sweet-spot in terms of semantic representation is critical for the construction of robust QA systems.  Work at a word-based representation instead of a sense representation.  Text simplification/clausal disembedding critical for relation extraction.  Need for a standardized semantic representation for relations extracted from texts.
  • 130. NLP & Semantic Computing Group Take-away Message • Text entailment:  Intensional-level reasoning.  Natural logic.  Distributional semantics. • Distributional semantics:  Robust, language-agnostic semantic matching.  Selective reasoning over commonsense KBs.