Back to HomePage
THIS PAGE IS OUT-OF-DATE. ONE DAY IT WILL BE AUTOMATICALLY GENERATED
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
- Ontology: An Ontology is a formal representation of knowledge within a specific domain.
- Context:
- Typically the formalism will support IS-A and HAS-A relationships between concepts.
- Concepts may also contain attributes, where an instance within the ontology is described by a set of values associated to one or more of these attributes.
- A main thrust in ontologies is the ability for a new system to quickly make use of the ontology.
- Example(s):
- The BIOPAX ontology at http://www.biopax.org arose out of the need of several research laboratories to share their knowledge about metabolic pathways. The BIOPAX ontology is encoded in OWL.
- See: Knowledge Representation; OWL, Semantic Web, Fen, 2001
.
- OWL: OWL is an Ontology Language.
P
- Paraphrase: A Paraphrase is a restatement of a Statement in another way.
- Phrase: A Phrase is a group of Words that form a syntactic unit but has no Subject-Predicate combination and so cannot stand alone as a Sentence.
- Polysemy: Polysemy is a type of relationship between one (written) word and all others in the lexicon. A word is polysemous if it can convey more than one meaning to the reader.
- Predicate: See: Predicate Function, Predicate Phrase
- Predicate Function: A Predicate Function is a Function that returns either a true or false value.
- Predicate Adjective: A Predicate Adjective is an Adjective that follows a Linking Verb (e.g. is, seems), and which agrees with the Subject in number, gender, and case.
- Predicate Argument: A Predicate Argument is one of the values accepted by a Predicate Function.
- Predicate Calculus: See First-Order Logic
. - Predicate Noun: A Predicate Noun a noun or pronoun which follows a Linking Verb and which is the same as the Subject.
- Predicate Phrase: A Predicate Phrase is a Verb Phrase that expresses what is said about a Subject.
- Preposition: A Preposition is an Adposition that is placed before the modifying concept.
- Prepositional Phrase:* A Prepositional Phrase is a Phrase composed of a Preposition and its modifier.
- PROGOL: PROGOL is a Machine Learning Algorithm.
- Prolog: Prolog is a Computing Language that implements automated Deductive Reasoning.
- Context:
- Prolog comes from the phrase "Programming in Logic".
- It was originally designed by A. Colmerauer and P. Roussel in 1971 for natural-language processing but has since been applied to several other AI problems.
- See: Deductive Reasoning, LISP
.
- Pronoun Resolution: See Anaphora Resolution
. - PropBank: PropBank is a Corpus derived from the Penn Treebank Corpus that has been enriched with Proposition structures.
- Context:
- In the formalism, semantic arguments are encoded with (A0-A5,AA), adjuncts with (AM-), references with (R-), and Verbs with (V).
- Verb senses come from VerbNet.
- See: Adjucative Argument, Proposition, Semantic Argument, Semantic Role Labeling, http://www.cis.upenn.edu/~ace/, Palmer, Gildea, and Kingsburyet, 2003
.
- Proposition: A Proposition is a Predicate (verb) and its set of Arguments.
- AKA: Syntactic Constituent
- Context:
- A sentence often contains more than one proposition.
- See: Semantic Role Labeling
.
- Propositional Learner: A Propositional Learner is a Machine Learning Algorithm that is capable of inducing Propositional Rules.
- Context:
- Propositional Logic: Propositional Logic is a system of Logic that operates on individual members of the domain.
- Pronoun Resolution: Pronoun Resolution is the task of identifying the proper noun related to a pronoun in a paragraph.
Q
R
S
- Second-Order Logic: Second-order logic is a Logic System that allows quantification over predicates. AKA: Second-Order Predicate Calculus. Context: The system is impractical for computational purpose, but is helpful in the theoretical analysis of uncountable theories, such as Cantor sets.
- Semantic Analysis: Semantic Analysis is any NLP Task where some meaning is identified within a Document.
- Context:
- Example(s):
- See: Lexical Semantic Analysis, Semantic Parsing, Syntactic Analysis
. - Semantic Argument: A semantic argument is an argument defined by verb-specific roles.
- Semantic Parsing: Semantic Parsing is the Parsing of the underlying meaning in a document into some Semantic Representation Language.
- Semantic Relation: A Semantic Relation is a relation between two or more Concepts that is True in some Domain.
- Semantic Role Labeling: Semantic Role Labeling is the NLP Task for the identification of the Propositions associated with each Predicate in a Sentence.
- Semantic Web: A nascent endeavor to create another World Wide Web where the information must be published in a format easy for computers to access (as opposed to the current WWW where the information is meant for human consumption). The standard encoding mechanism for the Semantic Web is OWL (and formerly RDF). An example of the Semantic Web is the www.BioPAX.org site.
- Semi-Supervised Learning: Semi-supervised learning is a type of machine learning where an algorithm makes use of both labeled and unlabelled data.
- Sentence: A Sentence is a sequence of Terminal Words that conforms some Natural Language Syntax.
. - SEQUENTIAL-COVERING: SEQUENTIAL-COVERING is a supervised classification algorithm that performs a general-to-specific beam search through rule-space. The algorithm removes training examples covered by each discovered rule and then repeats until all the positive examples have been covered. The algorithm does not backtrack so the underlying LEARN-ONE-RULE must be effective.
- Statement: A Statement is something that may be True in some Domain.
- Statistically Independent: Two Events are Statistically Independent if the probability that they both occur simultaneously is equal to the product of the probability that each occurs individually.
- AKA: Stochastically Independent
- Context:
- Sometimes statistical independence is represented as P(A ^ B) and its property as P(A)·P(B).
- Alternatively, two events are independent if the discovery that one of the event has occurred does not help you determine whether the other event has also occurred: i.e., P(A|B) = P(A).
- See: Probability, Conditional Probability, Independent Variable
.
- Structured Data: Structured Data is Data that is in a format that is amenable to computer processing.
- Context:
- Example(s):
- It can be an RDBMS table,
- It can be an XML file.
- It can be an array data structure in a program.
- See: Data, Unstructured Data
.
- Subsumption: A subsumption relation specifies the relative generality of two concepts. More fomally, concept A subsumes concept B if the definitions of A and B logically imply that members of B must also be members of A.
- Synonym: A word x is a Synonym of another word y if they are both similar enough in meaning that they can be interchanged in some situations without loss of meaning.
- Context: Synonyms are often dependent on context.
- Example(s): The words 'attorney' can be a synonym of the word 'lawyer' although the word 'attorney' is more typical in American English.
- See: Antonym, Polysemy, Semantics, [[en.wikipedia.org/wiki/Synonym]].
.
- Syntactic Relation: A Syntactic Relation is a relation between Words that conforms to some Grammar.
T
- Table: In relational databases, a structure (table) that contains a set of records (tuples).
- See: Predicate, Relation.
.
- Target Verb: A Target Verb is a Verb that governs a Proposition.
- Taxonomy: A hierachical classification of concepts typically for a specific domain. The primary semantic organizing principle of taxonomies is class inclusion (is a or subsumption relationships). Examples of taxonomises include the tree of life, and library catalogues.
- Template: AKA: forms or 'frames') The common tabular-like structure that is filled in information extraction tasks. The elements of templates are often referred to as slots. Occassionally information extraction is referred to as a "template filling" or "slot filling" exercise.
- Text Mining: The automated discovery of interesting patterns from human-readable sources. Typical sources include the Web, email, corporate databases with text information, and publication databases such as Citeseer and MEDLINE. Text mining is sometimes referred to as data mining on unstructured text data.
- TF-IDF: TF-IDF is a function that estimates how well a term describes a document.
- Top-Down Learning: Refers to the technique of starting from a general rule and to proceed by specializing it.
- TREC: Text Retrieval Conference. A conference sponsored by NIST with tracks in Information Extraction, Question Answering, and other NLP tasks.
- Tuple: In relational databases, an entry within a relation.
U
V
W
X
- XML: XML is a standard that facilitates the exchange of structured data.
Useful Glossaries