Some slides adapted from Dorr, christof/courses/cmsc723-fall04, Kurfess: fkurfess/Courses/CSC-...

CSC 9010: Special Topics, Natural Language Processing. Spring, 2005. Matuszek & Papalaskari 1

Some slides adapted from Dorr, www.umiacs.umd.edu/~christof/courses/cmsc723-fall04 , Kurfess: www.csc.calpoly.edu/~fkurfess/Courses/CSC-481/W03/Slides/3-Knowledge-Representation.ppt and Hirschberg: www1.cs.columbia.edu/~julia/cs4705/syllabus.html

Semantics and Semantic Analysis

CSC 9010: Special Topics. Natural Language Processing.

Paula Matuszek, Mary-Angela Papalaskari

Spring, 2005



Meaning• So far, we have focused on the structure of

language, not on what things mean• We have been doing natural language

processing, but not natural language understanding.

• So what is natural language understanding?– Answering an essay question on an exam?– Deciding what to order at a restaurant by reading a menu?– Realizing you’ve been insulted?– Appreciating a sonnet?

• As hard as answering "What is artificial intelligence?"



Meaning, cont• On the practical side, we want to “understand”

natural language because morphology- and syntax-based methods will only take us so far in some things:– Machine translation– Generation– Question answering

• So we saw how we could use n-grams to choose the more likely translation for a word given other words. But n-grams can’t give us the potential translations in the first place.



Semantics• What kinds of things can we not do well with

the tools we have already looked at?– Retrieve information in response to unconstrained

questions: e.g., travel planning– Accurate translations– Play the "chooser" side of 20 Questions– Read a newspaper article and answer questions

about it

• These tasks require that we also consider semantics: the meaning of our tokens and their sequences



So What IS Meaning?• From NLP viewpoint, meaning is a mapping

from linguistic forms to some kind of representation of knowledge of the world

• It is interpreted within the framework of some sort of action to be taken.

• Often we manipulate symbols all the way through; the “meaning” is put in by the human user. Translations, for instance.

• But not always – voice command systems, for instance, may map from the representation into actions.



Example• Question: Is there a restaurant in King of

Prussia serving vegetarian dinners?• From a restaurant database

– Lemon Grass is a sister restaurant to the one in West Philadelphia serving traditional Thai dishes in addition to a complete vegetarian Thai menu. The service and atmosphere are quite pleasant. A welcome addition to the King of Prussia area.

• What do we need to know to answer this question from this text?

• Can we unambiguously answer it?



Example, continued• Is there a restaurant in King of Prussia serving

vegetarian dinners? Yes.• Lemon Grass is a sister restaurant to the one in West Philadelphia

serving traditional Thai dishes in addition to a complete vegetarian Thai menu. The service and atmosphere are quite pleasant. A welcome addition to the King of Prussia area.

• What do we need to know? – Vegetarian Thai menu = vegetarian dinners– Welcome addition to the KoP area = in KoP

• Can we unambiguously answer it?– No. But we can be fairly certain.

• The only parse that makes sense

• Restaurants normally do serve the dinner meal.



Knowledge Representation for NLP• So what representation?• A useful KR needs to give us a way to encode the

knowledge relevant to our problem in a way which allows us to use it. Any representation which does this will work.

• How do we decide what we want to represent?– Entities, categories, events, time, aspect– Predicates, relationships among entities,

arguments (constants, variables)– And…quantifiers, operators (e.g. temporal)

• How much structure do we capture?



Structured Knowledge Representations

• Model-based representations reflect the structure of the domain, and then reason based on the model.– Semantic Nets– Frames– Scripts

• Sometimes called associative networks



Basics of Associative Networks• All include

– Concepts– Various kinds of links between concepts

• “has-part” or aggregation• “is-a” or specialization• More specialized depending on domain

• Typically also include– Inheritance– Some kind of procedural attachment



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)

links represent relationships the relationships contain the structural information of

the knowledge to be represented the label indicates the type of the relationship



Semantic Net Examples

Ship

Hull Propulsion

Steamboat

HasAHasA

InstanceOf

giveBob

Candy

Mary

Person

agent recipient

object

InstanceInstance



Generic/Individual• Generic describes the idea--the “notion”

– static

• Individual or instance describes a real entity– must conform to notion of generic– dynamic– “individuate” or “instantiate”

• A lot of NLP using semantic nets involves instantiating generic nets based on a given piece of text.



Individuation examplegivePerson

Thing

Personagent recipient

object

giveBob

Candy

Maryagent recipient

object

Generic RepresentationProcess the sentence“Bob gave Mary some candy”.

Instantiation



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

Typically can be organized hierarchically Most frame systems have an is-a slot allows the use of inheritance

Slots can contain all kinds of items Rules, facts, images, video, questions, hypotheses, other frames In NLP, typically capture relationships to other frames or entities

Slots can also have procedural attachments on creation, modification, removal of the slot value



Simple Frame ExampleSlot Name Filler

Restaurant Lemon Grass

Cuisine Thai, Vegetarian

Price Expensive

Service Excellent

Atmosphere Good

Location KoP

Web page Ask Google.



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

• Use contents of slots– Look up answers to questions– Generate new text

[Rogers 1999]



Scripts• Describe typical events or sequences

• Components are– script variables (players, props)– entry conditions– transactions– exit conditions

• Create instance by filling in variables



Restaurant Script Example• generic template for restaurants

– different types– default values

• script for a typical sequence of activities at a restaurant

• Often has a frame behind it; script is essentially instantiating the frame

[Rogers 1999]



Restaurant ScriptEAT-AT-RESTAURANT Script

Props: (Restaurant, Money, Food, Menu, Tables, Chairs)Roles: (Hungry-Persons, Wait-Persons, Chef-Persons)Point-of-View: Hungry-PersonsTime-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 then: if (Food-Quality was better than Palatable) then Compliments-To-The-Chef Script then: Pay-For-It-Script finally: Leave-Restaurant Script

[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.”



First Order Predicate Calculus (FOPC)• Terms

– Constants: Lemon Grass– Functions: LocationOf(Lemon Grass)– Variables: x in LocationOf(x)

• Predicates: Relations that hold among objects– Serves(Lemon Grass,VegetarianFood)

• Logical Connectives: Permit compositionality of meaning– I only have $5 and I don’t have a lot of time– Have(I,$5) Have(I,LotofTime)



FOPC Semantics• Sentences in FOPC can be assigned

truth values True or False



Variables and Quantifiers• Existential (): There exists

– A restaurant that serves Vegetarian food in KoP– (x) Restaurant(x) Serves(x,VegetarianFoor)

Near(LocationOf(x), KoP)

• Universal (): For all– All vegetarian restaurants serve vegetarian food– (x) VegetarianRestaurant(x) ->

Serves(x,VegetarianFood)



FOPC Examples• John gave Mary a book

– Previously: Give(John,Mary,book)

• Better:– (x) Giving(x) Giver(John,x)

Givee(Mary,x) Given(book,x)

• Full Definition of Give:– (w,x,y,z) Giving(x) Giver(w,x)

Givee(z,x) Given(y,x)



Choosing a Representation

• We would like our representation to support:– Verifiability– Unambiguous Representation– Canonical Form– Inference– Expressiveness



Verifiability• System can match input representation

against representations in knowledge base. If it finds a match, it can return Yes; Otherwise No.

• Does Lemon Grass serve vegetarian food?Serves(Lemon Grass,vegetarian food)



Unambiguous Representation• Single linguistic input can have different

meaning representations• Each representation unambiguously

characterizes one meaning.• Example: small cars and motorcycles are

allowed – car(x) & small(x) & motorcycle(y) & small(y) &

allowed(x) & allowed(y)– car(x) & small(x) & motorcycle(y) & allowed(x) &

allowed(y)



Ambiguity and Vagueness• An expression is ambiguous if, in a given

context, it can be disambiguated to have a specific meaning, from a number of discrete, possible meanings. E.g., bank (financial institution) vs bank (river bank)

• An expression is vague, if it refers to a range of a scalar variable, such that, even in a specific context, it’s hard to specify the range entirely. E.g., he’s tall, it’s warm, etc.



Representing Similar Concepts

• Distinct inputs could have the same meaning– Does Lemon Grass have vegetarian dishes?– Do they have vegetarian food at Lemon Grass?– Are vegetarian dishes served at Lemon Grass?– Does Lemon Grass serve vegetarian fare?

• Alternatives– Four different semantic representations– Store all possible meaning representations in KB



Canonical Form• Solution: Inputs that mean same thing

have same meaning representation

• Is this easy? No!– Vegetarian dishes, vegetarian food,

vegetarian fare– Have, serve

• What to do?



How to Produce a Canonical Form• Systematic Meaning Representations can be

derived from thesaurus– food ___– dish ___|____one overlapping meaning sense– fare ___|

• We can systematically relate syntactic constructions– [S [NP Lemon Grass] serves [NP

vegetarian dishes]]– [S [NP vegetarian dishes] are served at [NP

Lemon Grass]]



Inference• Consider a more complex request

– Can vegetarians eat at Lemon Grass?– Vs: Does Lemon Grass serve vegetarian food?

• Why do these result in the same answer?• Inference: Draw conclusions about truth of

propositions not explicitly stored in KB• serve(Lemon Grass,VegetarianFood) =>

CanEat(Vegetarians,At Lemon Grass)



Non-Yes/No Questions• Example: I'd like to find a

restaurant where I can get vegetarian food.

• serve(x,VegetarianFood)• Matching succeeds only if variable

x can be replaced by known object in KB.



Meaning Structure of Language

• Human Languages– Display a basic predicate-argument

structure– Make use of variables– Make use of quantifiers– Display a partially compositional semantics



Compositional Semantics• Assumption: The meaning of the whole is

comprised of the meaning of its parts– George cooks. Dan eats. Dan is sick.– Cook(George) Eat(Dan) Sick(Dan)– If George cooks and Dan eats, Dan will get sick.(Cook(George) ^ eat(Dan)) Sick(Dan)

• Meaning derives from – The people and activities represented (predicates

and arguments, or, nouns and verbs)– The way they are ordered and related: syntax of

the representation, which may also reflect the syntax of the sentence



Syntax-Driven Semantics S

NP VP eat(Dan) Nom V N Dan eats

• So….can we link syntactic structures to a corresponding semantic representation to produce the ‘meaning’ of a sentence in the course of parsing it?



Specific vs. General-Purpose Rules• We don’t want to have to specify for every

possible parse tree what semantic representation it maps to

• We want to identify general mappings from parse trees to semantic representations

• One approach:– Augment the lexicon and the grammar – Devise a mapping between rules of the grammar

and rules of semantic representation



Semantic AttachmentsExtend each grammar rule with instructions on

how to map the components of the rule to a semantic representationS NP VP {VP.sem(NP.sem)}

• Each semantic function defined in terms of the semantic representation of choice

• Problem: how to define these functions and how to specify their composition so we always get the meaning representation we want from our grammar?



A Simple Example McDonald’s serves burgers.• Associating constants with constituents

– ProperNoun McDonald’s {McDonald’s}– PlNoun burgers {burgers}

• Defining functions to produce these from input– NP ProperNoun {ProperNoun.sem}– NP PlNoun {PlNoun.sem}– Assumption: meaning representations of children

are passed up to parents for non-branching constuents

• Verbs are where the action is



• V serves {E(e,x,y) Isa(e,Serving) ^ Server(e,x) ^ Served(e,y)} where e = event, x = agent, y = patient

• Will every verb have its own distinct representation?– McDonald’s hires students.– McDonald’s gave customers a bonus.– Predicate(Agent, Patient, Beneficiary)

• Typically, divide verbs into classes based on what roles participate in the verb.



Doing Compositional Semantics

• To incorporate semantics into grammar we must– Figure out right representation for each constituent

based on the parts of that constituent (e.g. Adj)– Figure out the right representation for a category

of constituents based on other grammar rules, making use of that constituent (e.g. V.sem)

• This gives us a set of function-like semantic attachments incorporated into our CFG– E.g. Nom Adj Nom {x Nom.sem(x) ^

Isa(x,Adj.sem)}



What do we do with them?• Alter a parser such as an Early-style parser

so when constituents (dot at the end of the rule) are completed, the attached semantic is function applied and meaning representation is created and stored with state

• Or, let parser run to completion and then walk through resulting tree running semantic attachments from bottom-up



Option 1 (Integrated Semantic Analysis)S NP VP {VP.sem(NP.sem)}

– VP.sem has been stored in state representing VP– NP.sem stored with the state for NP– When rule completed, retrieve value of VP.sem and of

NP.sem, and apply VP.sem to NP.sem – Store result in S.sem.

• As fragments of input parsed, semantic fragments created

• Can be used to block ambiguous representations• But: you also perform semantic analysis on

orphaned constituents that play no role in final parse



Option 2: Post-hoc Semantic Analysis

• Do semantics after syntactic parse– Don’t do any semantic work on

constituents we will back out of – can be a significant efficiency gain

– But can’t use semantics to aid the parse.



Compositional Semantics Summary • Hypothesis: Principle of Compositionality

– Semantics of NL sentences and phrases can be composed from the semantics of their subparts

• Rules can be derived which map syntactic analysis to semantic representation (Rule-to-Rule Hypothesis)

• Parsing and mapping can be done during syntactic analysis or as a separate phase.



Non-Compositional Language• What do we do with language whose meaning isn’t

derived from the meanings of its parts– Non-compositional modifiers: fake, former, local– Metaphor:

• You’re the cream in my coffee. She’s the cream in George’s coffee.

• The break-in was just the tip of the iceberg. This was only the tip of Shirley’s iceberg.

– Idioms: • The old man finally kicked the bucket. The old man finally

kicked the proverbial bucket.

– Deferred reference: The ham sandwich wants his check.

• Mix lexical items with special grammar rules? Take a totally different approach?



Feature and Information Extraction• We don’t always need a complete parse.

– We may be after only some specific information– And we know what that information is

• Lemon Grass is a sister restaurant to the one in West Philadelphia serving traditional Thai dishes in addition to a complete vegetarian Thai menu. The service and atmosphere are quite pleasant. A welcome addition to the King of Prussia area.

– We don’t care at all about the information in red• Other semantic methods are more useful here.



Feature Extraction• Group individual terms into more complex entities (which

in turn become tokens)• Examples

– Dates, times, names, places

– URLs, HREFs and IMG tags

– Relationships like “X is president of Y”

• Can involve quite high-level features: language of a sample of text

• Not usually a goal in itself• In this case we are generating semantics for further use

by an application



Feature Extraction• Human-meaningful features: Parse token

stream, applying pattern-matching rules– general, broadly applicable features (dates)

– domain-specific features (chemical names)

– Can involve very sophisticated domain knowledge. • Statistical features:

– document length, vocabulary used, sentence length, document complexity, etc, etc

• Often first step in document-based analysis such as classification



Information Extraction• Retrieve some specific information which is

located somewhere in this set of documents.• Don’t want all the information, just what we

have specified.• Information may occur multiple times in the

text but we just need to find it once• Often what is really wanted from a web search,

for instance.• Often involves long sentences, complex

syntax, extended discourse, multiple writers• May involve sentence fragments and other

non-grammatical text



Some Examples of Information Extraction• Financial Information

– Who is the CEO/CTO of a company?– What were the dividend payments for stocks I’m interested in for

the last five years?

• Biological Information– Are there known inhibitors of enzymes in a pathway?– Are there chromosomally located point mutations that result in a

described phenotype?

• Other typical questions– Who is expert in some area?– Is there a vegetarian restaurant in KoP?– Can I get a flight to Chicago tomorrow?



• Create a model of information to be extracted• Create knowledge base of rules for extraction

– concepts– relations among concepts

• Process information applying a series of tools• Always involves some level of domain modeling

and considerable natural language processing

Information Extraction -- How



Staged Approach

• Apply a series of tools to an input text

• At each stage an element of is identified for possible use in the next level

• The end result is a set of relations suitable for entry into a database



Cascaded Process• Tokenize

• POS Tag

• Tag named entities: Names, dates, locations

• Tag complex entities and relationships– Grammatical, such as noun phrases– Semantic, such as CEO

• Bind values to tags throughout.



Named Entities• Named entity recognition…

– Labeling all the occurrences of named entities in a text…

• People, organizations, lakes, bridges, hospitals, mountains, etc…

• This is well-understood technology, readily available and works well.

• Typically uses a combination of enumerated lists (often called gazetteers) and regular expressions.



Complex Entities and Relationships• Uses Named Entities as components• Pattern-matching rules specific to a given

domain• May be multi-pass: One pass creates entities

which are used as part of later passes.– EG: First pass locates CEOs, second pass

locates dates for CEOs being replaced.

• May involve syntactic analysis as well, especially for things like negatives and reference resolution



Why Does This Work?• The problem is very constrained.

– Only looking for a small set of items that can appear in a small set of roles

• We can ignore stuff we don’t care about.

• BUT: creating the knowledge bases can be very complex and time-consuming.



Summary• Some NLP tasks require more than morphology and

syntax; they require some model of meaning: semantics• Semantics can be modeled in a variety of different

formalisms. The goal of all of them is to map from the sample of text being processed into some representation relevant to a real-world task

• Complete generic semantic parses gives us a very rich representation of text, but typically depend on the assumption that language is compositional.

• Other tasks such as Information Extraction may not require a detailed parse, but may depend on extensive domain knowledge

• How you represent and work with semantics will always depend on the NLP task you are trying to accomplish.

Some slides adapted from Dorr, christof/courses/cmsc723-fall04, Kurfess: fkurfess/Courses/CSC-...

Documents

Transcript of Some slides adapted from Dorr, christof/courses/cmsc723-fall04, Kurfess: fkurfess/Courses/CSC-...