U.S. patent application number 15/093337 was filed with the patent office on 2016-11-10 for method and apparatus for referring expression generation.
The applicant listed for this patent is ARRIA DATA2TEXT LIMITED. Invention is credited to Ehud Baruch REITER.
Application Number | 20160328381 15/093337 |
Document ID | / |
Family ID | 53368642 |
Filed Date | 2016-11-10 |
United States Patent
Application |
20160328381 |
Kind Code |
A1 |
REITER; Ehud Baruch |
November 10, 2016 |
METHOD AND APPARATUS FOR REFERRING EXPRESSION GENERATION
Abstract
Methods, apparatuses, and computer program products are
described herein that are configured to perform referring
expression generation. In some example embodiments, a method is
provided that comprises identifying an intended referent to be
referred to in a textual output. The method of this embodiment may
also include determining that a salient ancestor of the intended
referent is lower in a part-of hierarchy than a lowest common
ancestor. The method of this embodiment may also include causing
the salient ancestor to be set as a current target referent and a
new salient ancestor to be determined for the current target
referent. In some example embodiments, the default descriptor of
each current target referent is added to the referring noun phrase
and the part-of hierarchy is traversed via salient ancestor links
until the new salient ancestor of the current target referent is
higher than or equal to the lowest common ancestor.
Inventors: |
REITER; Ehud Baruch;
(Aberdeen, GB) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
ARRIA DATA2TEXT LIMITED |
Aberdeen |
|
GB |
|
|
Family ID: |
53368642 |
Appl. No.: |
15/093337 |
Filed: |
April 7, 2016 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
14634119 |
Feb 27, 2015 |
9355093 |
|
|
15093337 |
|
|
|
|
PCT/US2012/053183 |
Aug 30, 2012 |
|
|
|
14634119 |
|
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G06F 40/295 20200101;
G06F 40/289 20200101; G06F 40/56 20200101 |
International
Class: |
G06F 17/27 20060101
G06F017/27 |
Claims
1-20. (canceled)
21. A natural language generation method for generating a referring
noun phrase for an intended referent found in one or more messages,
the method comprising: arranging, using a processor, one or more
messages, wherein messages represent a phrase or a simple sentence
and are created in an instance in which an input data stream
comprises data that satisfies one or more message requirements, and
wherein at least a portion of the input data stream comprises
non-linguistic data; determining a lowest common ancestor for an
intended referent in a message of the one or more messages;
generating a referring noun phrase for the intended referent to be
included in a textual output by traversing a part-of hierarchy;
generating the textual output comprising the referring noun phrase
such that it is displayable on a user interface, wherein the
textual output linguistically describes at least a portion of the
input data stream; and displaying the textual output via a display
device.
22. A method according to claim 21, wherein the one or more parts
of the part-of hierarchy that are traversed based on one or more
salient ancestor links.
23. A method according to claim 21, further comprising: determining
that the intended referent is marked as salient; and causing the
referring noun phrase to solely comprise the default descriptor of
the intended referent.
24. A method according to claim 21, further comprising: determining
that the salient ancestor is equal to the lowest common ancestor;
and causing the referring noun phrase to comprise the default
descriptor of the intended referent.
25. A method according to claim 21, further comprising: determining
the previously referred-to entity based on a last mentioned entity
in a discourse model.
26. A method according to claim 25, wherein the previously
referred-to entity is set to a root component of the part-of
hierarchy in an instance in which the previously referred-to entity
is set to null.
27. A method according to claim 21, wherein the default descriptor
of an entity is at least one of a default descriptor, a class name
or a type name.
28. A method according to claim 21, further comprising: determining
that one or more parts of the part-of hierarchy traversed are
marked as ignore, wherein a default descriptor of the one or more
parts of the part-of hierarchy that are marked as to be ignored are
not included in the referring noun phrase.
29. A method according to claim 21, wherein the default descriptor
of an entity further comprises at least one of a class name or a
type name.
30. A method according to claim 21, wherein the referring noun
phrase comprises a predetermined maximum number of premodifiers of
the default descriptor of the intended referent, wherein the
predetermined number of premodifiers comprise one or more default
descriptors of the one or more parts of the part-of hierarchy that
are traversed.
31. A method according to claim 30, wherein the referring noun
phrase comprises a set of postmodifiers of the default descriptor
of the intended referent, wherein the set of postmodifiers comprise
one or more default descriptors of one or more parts of the part-of
hierarchy that are traversed that were not included as
premodifiers.
32. An apparatus comprising: at least one processor; and at least
one memory including computer program code, the at least one memory
and the computer program code configured to, with the at least one
processor, cause the apparatus to at least: arrange one or more
messages, wherein messages represent a phrase or a simple sentence
and are created in an instance in which an input data stream
comprises data that satisfies one or more message requirements,
wherein at least a portion of the input data stream comprises
non-linguistic data; determine a lowest common ancestor for an
intended referent in a message of the one or more messages;
generate a referring noun phrase for a salient ancestor of the
intended referent to be included in a textual output by traversing
a part-of hierarchy; generate the textual output comprising the
referring noun phrase such that it is displayable on a user
interface, wherein the textual output linguistically describes at
least a portion of the input data stream; and display the textual
output via a display device.
33. An apparatus according to claim 32, wherein the at least one
memory including the computer program code is further configured
to, with the at least one processor, cause the apparatus to:
determine that the intended referent is marked as salient; and
cause the referring noun phrase to solely comprise the default
descriptor of the intended referent.
34. An apparatus according to claim 32, wherein the at least one
memory including the computer program code is further configured
to, with the at least one processor, cause the apparatus to:
determine that the salient ancestor is equal to the lowest common
ancestor; and cause the referring noun phrase to comprise the
default descriptor of the intended referent.
35. An apparatus according to claim 32, wherein the at least one
memory including the computer program code is further configured
to, with the at least one processor, cause the apparatus to:
determine the previously referred-to entity based on a last
mentioned entity in a discourse model.
36. A non-transitory computer readable memory medium having program
code instructions stored thereon, the program code instructions
which when executed by an apparatus cause the apparatus at least
to: arrange one or more messages, wherein messages represent a
phrase or a simple sentence and are created in an instance in which
an input data stream comprises data that satisfies one or more
message requirements, wherein at least a portion of the input data
stream comprises non-linguistic data; determine a lowest common
ancestor for an intended referent in a message of the one or more
messages; generate a referring noun phrase for the intended
referent to be included in a textual output by traversing a part-of
hierarchy; generate the textual output comprising the referring
noun phrase such that it is displayable on a user interface,
wherein the textual output linguistically describes at least a
portion of the input data stream; and display the textual output
via a display device.
37. A computer program product according to claim 36, further
comprising program code instructions configured to: determine that
the intended referent is marked as salient; and cause the referring
noun phrase to solely comprise the default descriptor of the
intended referent.
38. A computer program product according to claim 36, further
comprising program code instructions configured to: determine that
the salient ancestor is equal to the lowest common ancestor; and
cause the referring noun phrase to comprise the default descriptor
of the intended referent.
39. A computer program product according to claim 36, further
comprising program code instructions configured to: determine the
previously referred-to entity based on a last mentioned entity in a
discourse model.
40. A computer program product according to claim 39, wherein the
previously referred-to entity is set to a root component of the
part-of hierarchy in an instance in which the previously
referred-to entity is set to null.
41. A method according to claim 21, wherein generating the
referring noun phrase further comprises: removing a first element
in the queue, wherein the first element is designated as a head
noun in the referring noun phase.
42. A method according to claim 41, wherein generating the
referring noun phrase further comprises: removing a next element in
the queue; and setting the next element as a premodifier to the
head noun; and in instance in which a predetermined premodifier
count threshold is satisfied, removing an element from the queue
and setting it as a post modifier to the head noun.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is a continuation of U.S. application Ser.
No. 14/634,119, filed Feb. 27, 2015, which is a continuation of
International Application No. PCT/US2012/053183, filed Aug. 30,
2012, all of which are hereby incorporated herein by reference.
TECHNOLOGICAL FIELD
[0002] Embodiments of the present invention relate generally to
natural language generation technologies and, more particularly,
relate to a method, apparatus, and computer program product for
referring expression generation.
BACKGROUND
[0003] In some examples, a natural language generation (NLG) system
is configured to transform raw input data that is expressed in a
non-linguistic format into a format that can be expressed
linguistically, such as through the use of natural language. For
example, raw input data may take the form of a value of a stock
market index over time and, as such, the raw input data may include
data that is suggestive of a time, a duration, a value and/or the
like. Therefore, an NLG system may be configured to input the raw
input data and output text that linguistically describes the value
of the stock market index; for example, "Securities markets rose
steadily through most of the morning, before sliding downhill late
in the day."
[0004] Data that is input into a NLG system may be provided in, for
example, a recurrent formal structure. The recurrent formal
structure may comprise a plurality of individual fields and defined
relationships between the plurality of individual fields. For
example, the input data may be contained in a spreadsheet or
database, presented in a tabulated log message or other defined
structure, encoded in a `knowledge representation` such as the
resource description framework (RDF) triples that make up the
Semantic Web and/or the like. In some examples, the data may
include numerical content, symbolic content or the like. Symbolic
content may include, but is not limited to, alphanumeric and other
non-numeric character sequences in any character encoding, used to
represent arbitrary elements of information. In some examples, the
output of the NLG system is text in a natural language (e.g.
English, Japanese or Swahili), but may also be in the form of
synthesized speech.
BRIEF SUMMARY
[0005] Methods, apparatuses, and computer program products are
described herein that are configured to perform referring
expression generation. In some example embodiments, a method is
provided that comprises identifying an intended referent to be
referred to in a textual output. The method of this embodiment may
also include determining a lowest common ancestor for the intended
referent and a previously referred-to entity within a part-of
hierarchy. The method of this embodiment may also include
determining that a salient ancestor of the intended referent is
lower in the part-of hierarchy than the lowest common ancestor in
an instance in which the intended referent is marked as not
salient. The method of this embodiment may also include causing the
salient ancestor to be set as a current target referent and a new
salient ancestor to be determined for the current target referent.
In some example embodiments, the default descriptor of each current
target referent is added to the referring noun phrase and the
part-of hierarchy is traversed via salient ancestor links until the
new salient ancestor of the current target referent is higher than
or equal to the lowest common ancestor.
BRIEF DESCRIPTION OF THE DRAWINGS
[0006] Having thus described embodiments of the invention in
general terms, reference will now be made to the accompanying
drawings, which are not necessarily drawn to scale, and
wherein:
[0007] FIG. 1 is a schematic representation of natural language
generation environment that may benefit from some example
embodiments of the present invention;
[0008] FIG. 2 illustrates an example referring expression
generation system according to some example embodiments described
herein;
[0009] FIG. 3 illustrates an example flow diagram that may be
performed by the referring expression generation system in
accordance with some example embodiments of the present
invention;
[0010] FIG. 4 illustrates an example hierarchy that may be accessed
by the referring expression generation system in accordance with
some example embodiments of the present invention;
[0011] FIG. 5 illustrates a block diagram of an apparatus that
embodies a referring expression generation system in accordance
with some example embodiments of the present invention; and
[0012] FIG. 6 illustrates a flowchart that may be performed by a
referring expression generation system in accordance with some
example embodiments of the present invention.
DETAILED DESCRIPTION
[0013] Example embodiments will now be described more fully
hereinafter with reference to the accompanying drawings, in which
some, but not all, embodiments are shown. Indeed, the embodiments
may take many different forms and should not be construed as
limited to the embodiments set forth herein; rather, these
embodiments are provided so that this disclosure will satisfy
applicable legal requirements. Like reference numerals refer to
like elements throughout. The terms "data," "content,"
"information," and similar terms may be used interchangeably,
according to some example embodiments, to refer to data capable of
being transmitted, received, operated on, and/or stored. Moreover,
the term "exemplary", as may be used herein, is not provided to
convey any qualitative assessment, but instead merely to convey an
illustration of an example. Thus, use of any such terms should not
be taken to limit the spirit and scope of embodiments of the
present invention.
[0014] Natural language texts that describe one or more entities,
such as those entities that have a complex internal structure (e.g.
machine parts, geographic locations, equipment or the like),
include a number of assets (e.g. listing of vehicles) or the like,
use referring expressions to identify particular intended referents
(e.g. components and sub components). In some examples, a referring
expression is any noun phrase, or surrogate for a noun phrase,
whose function in a text is to identify an individual person,
place, object, or a set of persons, places, objects or the like.
Referring expressions are generated based on the discourse context
(e.g. the previously generated text) and the genre of the text
(e.g. engineering maintenance manuals often use different referring
expressions from operational manuals).
[0015] However, in order to generate referring expressions that
describe an entity, a decision must be made about how much
information to include in each referring expression. For example, a
referring expression that has limited information may cause a
reader to become confused, whereas an expression with too much
information may reduce readability and effectiveness of a text. By
way of further example, in a complex system, one or more of the
following referring expressions that describe a blade component may
be generated based on a hierarchy (see e.g. FIG. 4): "the blades";
"the turbine blades"; "the right engine's turbine blades"; "the
turbine blades of the power system's right engine"; "the turbine
blades of the right engine of the Super Puma's power system";
and/or "the turbine blades of the Super Puma's right engine".
[0016] In order to generate or otherwise select a referring
expression to be included in a textual output, methods,
apparatuses, and computer program products are described herein
that are configured to generate a referring expression in the form
of a referring noun phrase using a part-of hierarchy, a reference
model and/or a discourse model. In particular, a microplanner,
having a referring expression generation system, may be configured
to generate the referring expression based on a default descriptor
for a particular entity to be referred to (the intended referent)
and one or more salient ancestors of the intended referent.
Further, the referring expression generation system may also be
configured to determine a previously referred-to entity and, as
such, may then identify a lowest common ancestor in the hierarchy
of the intended referent and the previously referred-to entity. A
salient ancestor (e.g. a prominent or important parent in the
hierarchy) may also be determined for the intended referent. In an
instance in which the salient ancestor is higher than or equal to
the lowest common ancestor, then the default descriptor of the
intended referent may become the referring expression. However, in
an instance in which the salient ancestor of the intended referent
is lower in the hierarchy than the lowest common ancestor, the
default descriptors of one or more salient ancestors (e.g. via one
or more salient links) of an intended referent may be formed
together with a default descriptor of the intended referent to
generate a referring expression.
[0017] FIG. 1 is an example block diagram of example components of
an example natural language generation environment 100. In some
example embodiments, the natural language generation environment
100 comprises a natural language generation system 102, message
store 110, a domain model 112 and/or linguistic resources 114. The
natural language generation system 102 may take the form of, for
example, a code module, a component, circuitry and/or the like. The
components of the natural language generation environment 100 are
configured to provide various logic (e.g. code, instructions,
functions, routines and/or the like) and/or services related to the
natural language generation system, the microplanner and a
referring expression generation system.
[0018] A message store 110 or knowledge pool is configured to store
one or more messages that are accessible by the natural language
generation system 102. Messages are language independent data
structures that correspond to informational elements in a text
and/or collect together underlying data, referred to as slots,
arguments or features, which can be presented within a fragment of
natural language such as a phrase or sentence. Messages may be
represented in various ways; for example, each slot may consist of
a named attribute and its corresponding value; these values may
recursively consist of sets of named attributes and their values,
and each message may belong to one of a set of predefined types.
The concepts and relationships that make up messages may be drawn
from an ontology (e.g. a domain model 112) that formally represents
knowledge about the application scenario. In some examples, the
domain model 112 is a representation of information about a
particular domain. For example, a domain model may contain an
ontology that specifies the kinds of objects, instances, concepts
and/or the like that may exist in the domain in concrete or
abstract form, properties that may be predicated of the objects,
concepts and the like, relationships that may hold between the
objects, concepts and the like, and representations of any specific
knowledge that is required to function in the particular
domain.
[0019] In some examples, messages are created based on a
requirements analysis as to what is to be communicated for a
particular scenario (e.g. for a particular domain or genre). A
message typically corresponds to a fact about the underlying data
(for example, the existence of some observed event) that could be
expressed via a simple sentence (although it may ultimately be
realized by some other linguistic means). For example, to
linguistically describe an object, such as an engine, a user may
want to know which engine is being referred to, a status of the
engine, a condition of the engine and/or the like. In some cases,
the user may not want to know an engine temperature, but instead
want to be warned in an instance in which the engine temperature is
at a dangerous level. For example, "the right engine is too hot."
In other examples, the engine being too hot may be linked to a
resultant condition, for example "after investigating the crash, it
appears the right engine was too hot."
[0020] In some examples, a message is created in an instance in
which the raw input data warrants the construction of such a
message. For example, a wind message would only be constructed in
an instance in which wind data was present in the raw input data.
Alternatively or additionally, while messages may correspond
directly to observations taken from a raw data input, others,
however, may be derived from the observations by means of a process
of inference or based on one or more detected events. For example,
the presence of rain may be indicative of other conditions, such as
the potential for snow at some temperatures.
[0021] Messages may be instantiated based on many variations of
source data, such as but not limited to time series data, time and
space data, data from multiple data channels, an ontology, sentence
or phrase extraction from one or more texts, a text, survey
responses, structured data, unstructured data and/or the like. For
example, in some cases, messages may be generated based on text
related to multiple news articles focused on the same or similar
news story in order to generate a news story. Whereas, in other
examples, messages may be built based on survey responses and/or
event data.
[0022] Messages may be annotated with an indication of their
relative importance; this information can be used in subsequent
processing steps or by the natural language generation system 102
to make decisions about which information may be conveyed and which
information may be suppressed. Alternatively or additionally,
messages may include information on relationships between the one
or more messages.
[0023] In some example embodiments, a natural language generation
system, such as natural language generation system 102, is
configured to generate phrases, sentences, text or the like which
may take the form of a natural language text. The natural language
generation system 102 comprises a document planner 130, a
microplanner 132 and/or a realizer 134. The natural language
generation system 102 may also be in data communication with the
message store 110, the domain model 112 and/or the linguistic
resources 114. In some examples, the linguistic resources include,
but are not limited to, text schemas, aggregation rules, reference
rules, lexicalization rules and/or grammar rules that may be used
by one or more of the document planner 130, the microplanner 132
and/or the realizer 134. Other natural language generation systems
may be used in some example embodiments, such as a natural language
generation system as described in Building Natural Language
Generation Systems by Ehud Reiter and Robert Dale, Cambridge
University Press (2000), which is incorporated by reference in its
entirety herein.
[0024] The document planner 130 is configured to input the one or
more messages from the message store 110. The document planner 130
is further configured to determine how to arrange those messages in
order to describe the patterns in the one or more data channels
derived from the raw input data. The document planner 130 may
comprise a content determination process that is configured to
select the messages, such as the messages that contain a
representation of the data that is to be output via a natural
language text.
[0025] The document planner 130 may also comprise a structuring
process that determines the order of messages to be included in a
text. In some example embodiments, the document planner 130 may
access one or more text schemas for the purposes of content
determination and document structuring. A text schema is a rule set
that defines the order in which a number of messages are to be
presented in a document. For example, a medication injection
message may be described prior to a heart rate spike message. In
other examples, a steady respiration rate message may be described
after, but in relation to, the heart rate spike message.
[0026] The output of the document planner 130 may be a
tree-structured object or other data structure that is referred to
as a document plan. In an instance in which a tree-structured
object is chosen for the document plan, the leaf nodes of the tree
may contain the messages, and the intermediate nodes of the tree
structure object may be configured to indicate how the subordinate
nodes are related (e.g. elaboration, consequence, contrast,
sequence and/or the like) to each other.
[0027] The microplanner 132 is configured to construct a text
specification based on the document plan from the document planner
130, such that the document plan may be expressed in natural
language. In some example embodiments, the microplanner 132 may
perform aggregation, lexicalization and referring expression
generation. In some examples, aggregation includes, but is not
limited to, determining whether two or more messages can be
combined together linguistically to produce a more complex phrase
specification. For example, one or more messages may be aggregated
so that both of the messages can be described by a single sentence.
In some examples, lexicalization includes, but is not limited to,
choosing particular words for the expression of concepts and
relations. In some examples, referring expression generation
includes, but is not limited to, choosing how to refer to an entity
so that it can be unambiguously identified by the reader. Referring
expression generation is further described with respect to at least
FIGS. 2 and 3. The output of the microplanner 132, in some example
embodiments, is a tree-structured realization specification whose
leaf-nodes are phrase specifications, and whose internal nodes
express rhetorical relations between the leaf nodes.
[0028] A realizer 134 is configured to traverse a text
specification output by the microplanner 132 to express the text
specification in natural language. The realization process that is
applied to each phrase specification makes use of a grammar (e.g.
the grammar of the linguistic resources 114) which specifies the
valid syntactic structures in the language and further provides a
way of mapping from phrase specifications into the corresponding
natural language sentences. The output of the process is, in some
example embodiments, a natural language text. In some examples, the
natural language text may include embedded mark-up.
[0029] FIG. 2 illustrates an example referring expression
generation system 208 embodied by a microplanner 132 according to
some example embodiments described herein. The referring expression
generation system 208, in some example embodiments, is configured
to generate a referring expression (e.g. a referring noun phrase)
for an intended referent found in one or more messages within a
document plan. The referring expression may then be used to
populate an element of a phrase specification in the text
specification generated by the microplanner 132.
[0030] In some example embodiments and in order to generate the
referring expression, the referring expression generation system
208 is configured to access or otherwise be in data communication
with a data model 112 that may additionally comprise or otherwise
embodies at least a hierarchy 202 (e.g. a part-of hierarchy such as
the hierarchy shown in FIG. 4, or the like), a reference model 204
and/or the like. The reference model 204 specifies additional
information about the components in the hierarchy 202. For example,
the reference model 204 may provide a default descriptor for an
entity and may further identify a salient ancestor for that entity.
In some examples, a default descriptor is the name used for a
particular entity in a text in an instance in which no additional
information about higher level entities in the hierarchy 202 is
provided. For example, the default descriptor of Eurocopter AS332
Super Puma Helicopter shown in FIG. 4 may be "Super Puma". "Super
Puma" may provide enough information for a reader in a helicopter
or aircraft genre or domain to identify that the intended referent
of "Super Puma" is a Eurocopter AS332. As such, the default
descriptor typically used for a given entity in a genre may not be
as rich or as complicated as the full name or nomenclature used in
an underlying hierarchy.
[0031] In some examples, a salient ancestor is an ancestor of an
intended referent in the hierarchy 202 that may be added to a
referring expression in an instance in which it is insufficient to
use the default descriptor of the intended referent alone. In some
examples, the salient ancestor is the intended referent's parent in
the hierarchy 202; however, in other examples one or more levels
within the hierarchy 202 may be skipped or marked as to be ignored.
For example and with reference to FIG. 4, the reference model 204
may indicate in the helicopter genre that the salient ancestor of
the left engine is the Super Puma and not its parent, the power
system. Such a rule may be present because the genre may specify
that an engine is to be referred to as "the Super Puma's left
engine", not "the power system's left engine" or "the Super Puma's
power system's left engine". Alternatively or additionally, the
reference model 204 may be configured to include multiple Boolean
flags for each entity, such that the Boolean flags are configured
to indicate whether an entity is always salient or if an entity
should be skipped when looking for salient ancestors.
[0032] In an instance in which the salient ancestor is null or the
intended referent is otherwise indicated as always salient, then
the intended referent may be described without reference to a
salient ancestor. For example, in a geographic ontology or
hierarchy, Chicago and Springfield would both be beneath the state
of Illinois; however Chicago would likely be marked as always
salient (because in most contexts "Chicago" by itself is
sufficient). On the other hand, Springfield would not be marked as
always salient and would have Illinois as a salient ancestor,
because "Springfield" without "Illinois" would not be meaningful.
Further, in a sports ontology, a player such as David Beckham or
Ronaldo may be marked as always salient, whereas Freddi Montero may
require a salient ancestor of Seattle Sounders FC in a textual
output. In some examples, Seattle Sounders may need a further
salient ancestor of Major League Soccer, in other example textual
outputs.
[0033] In some example embodiments a discourse model 206 is
embodied by or may be accessed by the microplanner 132. The
discourse model 206 is configured to record the entities previously
referred to in the present text (e.g. entities mentioned in
previous phrase specifications), along with the referring
expressions that were used to refer to them. For example, and with
reference to FIG. 4, if a previous referring expression referred to
"right engine casings", the next referring expression may only need
to refer to "the turbine" since the right engine is already
salient. As such, the referring expression generation system 208
may access the discourse model 206 in order to determine the
previous entities referred to.
[0034] In some example embodiments, the referring expression
generation system 208 may determine the lowest common ancestor in
the hierarchy 202 between an intended referent and the previous
referent. Using the lowest common ancestor, the intended referent
and the salient ancestors, a referring expression, for example in
the form of a referring noun phrase, is generated by the referring
expression generation system 208 by including the default
descriptors of the intended referent and its ancestors (e.g. by
following the salient ancestor links) until the lowest common
ancestor, or one of its ancestors that is salient, is reached. The
generation of the referring expression is further described with
reference to FIG. 3.
[0035] FIG. 3 illustrates an example flow diagram that may be
performed by the referring expression generation system 208 in
accordance with some example embodiments of the present invention.
As is shown in block 302, the referring expression generation
system 208 may identify an intended referent. The referring
expression generation system 208 may be activated or otherwise
executed by the microplanner 132, such as in an instance in which
the microplanner is applying one or more rules (e.g. lexicalization
rules) to a slot in a message that specifies an entity. In other
embodiments, the referring expression generation system 208 may
access or otherwise receive the intended referent via the
microplanner 134, the natural language generation system 108, a
lexicalization rule and/or the like.
[0036] In block 304, the referring expression generation system 208
may access the discourse model 206 to obtain the previous entity
referred to. The previous entity referred to is the last or prior
entity that was the intended referent in the generation of a
referring expression. The previous entity referred to may then be
identified or otherwise located in the hierarchy by the referring
expression generation system 208. Using the previous entity
referred to, the referring expression generation system 208 may be
configured to set a lowest common ancestor to be the lowest entity
within the hierarchy that is an ancestor of both the intended
referent and the previous entity referred to in block 306. In an
instance in which the previous entity referred to is null, then the
lowest common ancestor may be set to a root entity of the hierarchy
(e.g. Super Puma in FIG. 4).
[0037] In block 308, the default descriptor of the intended
referent in the reference model is added to a descriptor queue. In
some example embodiments, the descriptor queue is initialized as an
empty queue prior to the first instance of block 308. Alternatively
or additionally, the process shown in FIG. 3 may be configured to
end in an instance in which the intended referent is identified,
such as by the reference model 204, a flag or the like, as always
salient. In such cases, the referring expression (e.g. referring
noun phrase) may then take the form of the default descriptor of
the intended referent. As is shown in block 310, the target
referent is set to be the intended referent.
[0038] As shown in block 312, a salient ancestor of the intended
referent in the reference model may be identified, such as via the
reference model 204. At decision block 314, the referring
expression generation system 208 is configured to determine whether
the salient ancestor is lower in the hierarchy than the lowest
common ancestor. If so, then at block 316, the current salient
ancestor is set as a new target referent. In block 318, the default
descriptor for the new target referent is added to the descriptor
queue and a new salient ancestor for the new target referent is
determined in block 312 The process of blocks 312-318 continues
until, at decision block 314, the salient ancestor of the current
target referent is higher than or equal to the lowest common
ancestor. As noted, during each iteration through blocks 312-318,
the default descriptor for the target referent is added to the
descriptor queue. As such, the hierarchy is traversed using salient
ancestor links. For example salient ancestor links may be
represented as:
[0039] (1) start with TARGET
[0040] (2) go to Salient Ancestor (TARGET)
[0041] (3) go to Salient Ancestor (Salient Ancestor (TARGET))
[0042] (4) go to Salient Ancestor (Salient Ancestor (Salient
Ancestor (TARGET))
[0043] Alternatively or additionally, in some examples other
methods of traversal may be used such as, but not limited to,
traversing each hierarchy and skipping those entities marked as to
be ignored, traversing the hierarchy and including parent entities,
and/or the like.
[0044] At block 320, the first element of the descriptor queue is
removed and designated as the head noun of a referring noun phrase.
In some example embodiments, the head noun may be assigned a
determiner of "the". For example if the intended referent was
"Super Puma", the default descriptor may be stored in the referring
noun phrase as "the Super Puma".
[0045] At decision block 322, in an instance in which the
descriptor queue is not empty, it is determined whether a
predetermined premodifier count as been reached. In some examples,
a premodifier count is predetermined and indicates the maximum
number of premodifiers that may be placed before the head noun in
the referring noun phrase. In an instance in which the premodifier
count has not been reach or satisfied, then, at block 324, the
default descriptor of the first element of the descriptor queue is
set as a premodifier to the head noun in the referring noun phrase.
For example, in an instance in which a blade is the intended
referent, "turbine" may be added as a premodifier resulting in a
referring expression "the turbine blades". The premodifier count is
also incremented in block 324. Such a process continues until the
premodifier count is reached or the descriptor queue is empty.
[0046] In an instance in which the maximum premodifier count is
reached or the descriptor queue is empty, then at decision block
326, it is determined whether the descriptor queue is empty. If the
descriptor queue is not empty, then at block 328, the first element
of the descriptor queue is set as a postmodifier to the head noun
in the referring noun phrase. In some examples the first element is
added to a prepositional phrase having the proposition "of" and is
then added to the referring noun phrase. For example, in an
instance in which a blade is the intended referent and "turbine" is
the premodifier, "right engine" may be added as the postmodifier
resulting in a referring expression "the turbine blade of the right
engine". Such a process continues until the descriptor queue is
empty.
[0047] At block 330, the referring noun phrase is returned as the
referring expression for use in a phrase specification and
eventually the textual output of the natural language generation
system 102. Alternatively or additionally, the discourse model 206
is updated to reflect the most recently identified intended
referent.
[0048] Alternatively or additionally, one or more entities may not
have a default descriptor in a reference model. For example, in the
hierarchy of FIG. 4, Gear 1 through to Gear 8 may not be assigned
individual default descriptors and, in such example cases, the
entities may be referred to by the name of a class, a type or the
like (e.g. "gear").
[0049] By way of example, in order to generate a referring
expression for an entity that does not have a default descriptor,
the referring expression generation system 208 may, in some example
embodiments, determine whether the intended referent has been
previously referred to. In an instance in which the intended
referent has not been previously referred to, then the referring
noun phrase may include "a" and the class or type of the intended
referent (e.g. a gear). In an instance in which the intended
referent has been previously referred to, then the referring
expression generation system 208 may determine the previous
references in the one or more phrase specifications to a same class
or type via the discourse model. In an instance in which the
intended referent is the most recently referred-to entity in the
previously referred-to entities then the referring noun phrase may
include the determiner "the" and the intended referent class or
type name. Otherwise the referring noun phrase is generated to
include "one of the" and the entity is set to plural (e.g. "one of
the gears"). Alternatively or additionally, sets may be referred to
in a same or similar manner. For example, if referring to multiple
unnamed entities of the same type and in the same hierarchy
position (e.g. the gears), phrases such as "four of the gears" or
"all of the gears" may be returned.
[0050] FIG. 4 illustrates an example part-of hierarchy that may be
accessed by the referring expression generation system in
accordance with some example embodiments of the present invention.
In some examples, the part-of hierarchy shown with respect to FIG.
4 was generated based on the official air accident report
referenced below; however in other examples, the part-of hierarchy
may be provided by an equipment manufacturer, an equipment manager
and/or the like. The example part-of hierarchy includes a power
system and a transmission system of a Eurocopter AS332 L2 Super
Puma Helicopter. As is described herein, the hierarchy of FIG. 4
may be used to generate referring expressions, such as those
underlined expressions in the following example text: [0051]
Initial examination of the engines revealed significant damage to
their external casings. The free turbine case of the right engine
was found to have been breached and the turbine blades were found
severely damaged. The epicyclic reduction gearbox had suffered
significant damage. The epicyclic module case and ring gear had
split vertically and had separated from the main module, (see FIG.
20). The first stage planet carrier was found lying on the remains
of the main module and the remains of all eight first stage planet
gears, together with pieces of a second stage planet gear, were
recovered from within the MGB main module and its surrounding
area.
[0052] Text was taken from Section 1.12.2 of an official air
accident report
(http://www.aaib.gov.uk/cms_resources/2-2011%20G-REDL.pdf).
[0053] For example in order to describe the blades 402, the
referring expression generation system 210 is configured to
determine a previous entity referred to in the text, such as via
the discourse model 206. In this example, the previous entity
referred to in the text was the free turbine case 404. Therefore,
in this example, the lowest common ancestor in the hierarchy,
between the free turbine case 404 and the blades 402, is the right
engine 406. A reference model, such as reference model 204, may
indicate that the turbine 408 is the salient ancestor of the blades
402. As such, the blades 402 are designated to be referred to by
the head noun of a referring noun phrase. Because the turbine 408
is located beneath the lowest common ancestor in the hierarchy, the
turbine 408 is then set as the target referent and its salient
ancestor, the right engine 406, is determined. In this example, the
right engine 406 is the salient ancestor of the turbine 408 and is
also the lowest common ancestor and therefore the process ends. The
referring noun phrase may be then generated by the referring
expression generation system 210 having "blades" as the head noun
and having "turbine" as a premodifier or postmodifier. For example,
the referring noun phrase may be: "the turbine blades" or "the
blades of the turbine".
[0054] FIG. 5 is an example block diagram of an example computing
device for practicing embodiments of an example configurable
microplanner. In particular, FIG. 5 shows a computing system 500
that may be utilized to implement a natural language generation
environment 100 having a natural language generation system 102
including, in some examples, a document planner 130, a microplanner
132 and/or a realizer 134 and/or a user interface 510. One or more
general purpose or special purpose computing systems/devices may be
used to implement the natural language generation system 102 and/or
the user interface 510. In addition, the computing system 500 may
comprise one or more distinct computing systems/devices and may
span distributed locations. In some example embodiments, the
natural language generation system 102 may be configured to operate
remotely via the network 550. In other example embodiments, a
pre-processing module or other module that requires heavy
computational load may be configured to perform that computational
load and thus may be on a remote device or server. For example, the
realizer 134 may be accessed remotely. Furthermore, each block
shown may represent one or more such blocks as appropriate to a
specific example embodiment. In some cases one or more of the
blocks may be combined with other blocks. Also, the natural
language generation system 102 may be implemented in software,
hardware, firmware, or in some combination to achieve the
capabilities described herein.
[0055] In the example embodiment shown, computing system 500
comprises a computer memory ("memory") 501, a display 502, one or
more processors 503, input/output devices 504 (e.g., keyboard,
mouse, CRT or LCD display, touch screen, gesture sensing device
and/or the like), other computer-readable media 505, and
communications interface 506. The processor 503 may, for example,
be embodied as various means including one or more microprocessors
with accompanying digital signal processor(s), one or more
processor(s) without an accompanying digital signal processor, one
or more coprocessors, one or more multi-core processors, one or
more controllers, processing circuitry, one or more computers,
various other processing elements including integrated circuits
such as, for example, an application-specific integrated circuit
(ASIC) or field-programmable gate array (FPGA), or some combination
thereof. Accordingly, although illustrated in FIG. 5 as a single
processor, in some embodiments the processor 503 comprises a
plurality of processors. The plurality of processors may be in
operative communication with each other and may be collectively
configured to perform one or more functionalities of the reference
system as described herein.
[0056] The natural language generation system 102 is shown residing
in memory 501. The memory 501 may comprise, for example, transitory
and/or non-transitory memory, such as volatile memory, non-volatile
memory, or some combination thereof. Although illustrated in FIG. 5
as a single memory, the memory 501 may comprise a plurality of
memories. The plurality of memories may be embodied on a single
computing device or may be distributed across a plurality of
computing devices collectively configured to function as the
natural language system, the microplanner and/or the reference
system. In various example embodiments, the memory 501 may
comprise, for example, a hard disk, random access memory, cache
memory, flash memory, a compact disc read only memory (CD-ROM),
digital versatile disc read only memory (DVD-ROM), an optical disc,
circuitry configured to store information, or some combination
thereof.
[0057] In other embodiments, some portion of the contents, some or
all of the components of the natural language generation system 102
may be stored on and/or transmitted over the other
computer-readable media 505. The components of the natural language
generation system 102 preferably execute on one or more processors
503 and are configured to enable operation of a configurable
microplanner, as described herein.
[0058] Alternatively or additionally, other code or programs 530
(e.g., an administrative interface, a Web server, and the like) and
potentially other data repositories, such as other data sources
540, also reside in the memory 501, and preferably execute on one
or more processors 503. Of note, one or more of the components in
FIG. 5 may not be present in any specific implementation. For
example, some embodiments may not provide other computer readable
media 505 or a display 502.
[0059] The natural language generation system 102 is further
configured to provide functions such as those described with
reference to FIG. 1. The natural language generation system 102 may
interact with the network 550, via the communications interface
506, with remote data sources 556 (e.g. remote reference data,
remote performance data, remote aggregation data, remote knowledge
pools and/or the like), third-party content providers 554 and/or
client devices 558. The network 550 may be any combination of media
(e.g., twisted pair, coaxial, fiber optic, radio frequency),
hardware (e.g., routers, switches, repeaters, transceivers), and
protocols (e.g., TCP/IP, UDP, Ethernet, Wi-Fi, WiMAX, Bluetooth)
that facilitate communication between remotely situated humans
and/or devices. In some instance the network 550 may take the form
of the internet or may be embodied by a cellular network such as an
LTE based network. In this regard, the communications interface 506
may be capable of operating with one or more air interface
standards, communication protocols, modulation types, access types,
and/or the like. The client devices 558 include desktop computing
systems, notebook computers, mobile phones, smart phones, personal
digital assistants, tablets and/or the like.
[0060] In an example embodiment, components/modules of the natural
language generation system 102 are implemented using standard
programming techniques. For example, the natural language
generation system 102 may be implemented as a "native" executable
running on the processor 503, along with one or more static or
dynamic libraries. In other embodiments, the natural language
generation system 102 may be implemented as instructions processed
by a virtual machine that executes as one of the other programs
530. In general, a range of programming languages known in the art
may be employed for implementing such example embodiments,
including representative implementations of various programming
language paradigms, including but not limited to, object-oriented
(e.g., Java, C++, C#, Visual Basic.NET, Smalltalk, and the like),
functional (e.g., ML, Lisp, Scheme, and the like), procedural
(e.g., C, Pascal, Ada, Modula, and the like), scripting (e.g.,
Perl, Ruby, Python, JavaScript, VBScript, and the like), and
declarative (e.g., SQL, Prolog, and the like).
[0061] The embodiments described above may also use synchronous or
asynchronous client-server computing techniques. Also, the various
components may be implemented using more monolithic programming
techniques, for example, as an executable running on a single
processor computer system, or alternatively decomposed using a
variety of structuring techniques, including but not limited to,
multiprogramming, multithreading, client-server, or peer-to-peer,
running on one or more computer systems each having one or more
processors. Some embodiments may execute concurrently and
asynchronously, and communicate using message passing techniques.
Equivalent synchronous embodiments are also supported. Also, other
functions could be implemented and/or performed by each
component/module, and in different orders, and by different
components/modules, yet still achieve the described functions.
[0062] In addition, programming interfaces to the data stored as
part of the natural language generation system 102, such as by
using one or more application programming interfaces can be made
available by mechanisms such as through application programming
interfaces (API) (e.g. C, C++, C#, and Java); libraries for
accessing files, databases, or other data repositories; through
scripting languages such as XML; or through Web servers, FTP
servers, or other types of servers providing access to stored data.
The message store 110, the domain model 112 and/or the linguistic
resources 114 may be implemented as one or more database systems,
file systems, or any other technique for storing such information,
or any combination of the above, including implementations using
distributed computing techniques. Alternatively or additionally,
the message store 110, the domain model 112 and/or the linguistic
resources 114 may be local data stores but may also be configured
to access data from the remote data sources 556.
[0063] Different configurations and locations of programs and data
are contemplated for use with techniques described herein. A
variety of distributed computing techniques are appropriate for
implementing the components of the illustrated embodiments in a
distributed manner including but not limited to TCP/IP sockets,
RPC, RMI, HTTP, Web Services (XML-RPC, JAX-RPC, SOAP, and the
like). Other variations are possible. Also, other functionality
could be provided by each component/module, or existing
functionality could be distributed amongst the components/modules
in different ways, yet still achieve the functions described
herein.
[0064] Furthermore, in some embodiments, some or all of the
components of the natural language generation system 102 may be
implemented or provided in other manners, such as at least
partially in firmware and/or hardware, including, but not limited
to one or more ASICs, standard integrated circuits, controllers
executing appropriate instructions, and including microcontrollers
and/or embedded controllers, FPGAs, complex programmable logic
devices ("CPLDs"), and the like. Some or all of the system
components and/or data structures may also be stored as contents
(e.g., as executable or other machine-readable software
instructions or structured data) on a computer-readable medium so
as to enable or configure the computer-readable medium and/or one
or more associated computing systems or devices to execute or
otherwise use or provide the contents to perform at least some of
the described techniques. Some or all of the system components and
data structures may also be stored as data signals (e.g., by being
encoded as part of a carrier wave or included as part of an analog
or digital propagated signal) on a variety of computer-readable
transmission mediums, which are then transmitted, including across
wireless-based and wired/cable-based mediums, and may take a
variety of forms (e.g., as part of a single or multiplexed analog
signal, or as multiple discrete digital packets or frames). Such
computer program products may also take other forms in other
embodiments. Accordingly, embodiments of this disclosure may be
practiced with other computer system configurations.
[0065] FIG. 6 is a flowchart illustrating an example method
performed by a reference system in accordance with some example
embodiments described herein. As is shown in operation 602, an
apparatus may include means, such as the microplanner 132, the
reference model 204, the referring expression generation system
208, the processor 503, or the like, for identifying an intended
referent to be referred to in a textual output. In some example
embodiments, the default descriptor of the intended referent is set
as a head noun of a referring noun phrase. Alternatively or
additionally, the default descriptor of an entity further comprises
at least one of a class name or a type name.
[0066] As is shown in operation 604, an apparatus may include
means, such as the microplanner 132, the hierarchy 202, the
discourse model 206, the reference system 208, the processor 503,
or the like, for determining a lowest common ancestor for the
intended referent and a previously referred-to entity within a
part-of hierarchy, such as an equipment part-of hierarchy. As is
shown in operation 606, an apparatus may include means, such as the
microplanner 132, the hierarchy 202, the discourse model 206, the
referring expression generation system 208, the processor 503, or
the like, for determining the previously referred-to entity based
on a last entity mentioned in a discourse model. In some example
embodiments, the previously referred-to entity is set to a root
component of the part-of hierarchy in an instance in which the
previous reference is set to null.
[0067] As is shown in operation 608, an apparatus may include
means, such as the microplanner 132, the hierarchy 202, the
referring expression generation system 208, the processor 503, or
the like, for determining that a salient ancestor of the intended
referent is higher than or equal to the lowest common ancestor in
the part-of hierarchy, such that the referring noun phrase
comprises the default descriptor of the intended referent. As is
shown in operation 610, an apparatus may include means, such as the
microplanner 132, the hierarchy 202, the referring expression
generation system 208, the processor 503, or the like, for
determining that a salient ancestor of the intended referent is
lower in the part-of hierarchy than a lowest common ancestor in an
instance in which the intended referent is marked as not
salient.
[0068] As is shown in operation 612, an apparatus may include
means, such as the microplanner 132, the hierarchy 202, the
reference model 204, the referring expression generation system
208, the processor 503, or the like, for causing the salient
ancestor to be set as a current target referent and a new salient
ancestor to be determined for the current target referent, wherein
the default descriptor of each current target referent is added to
the referring noun phrase and the part-of hierarchy is traversed
via salient ancestor links until the new salient ancestor of the
current target referent is higher than or equal to the lowest
common ancestor.
[0069] In some example embodiments, the referring noun phrase
comprises a predetermined maximum number of premodifiers of the
default descriptor of the intended referent, wherein the
premodifiers comprise one or more default descriptors of the one or
more parts of the part-of hierarchy traversed. In additional
example embodiments, the referring noun phrase comprises a number
of postmodifiers of the default descriptor of the intended
referent, wherein the postmodifiers comprise the remaining one or
more default descriptors of the one or more parts of the part-of
hierarchy traversed not included as premodifiers.
[0070] FIGS. 3 and 6 illustrate example flowcharts of the
operations performed by an apparatus, such as computing system 500
of FIG. 5, in accordance with example embodiments of the present
invention. It will be understood that each block of the flowcharts,
and combinations of blocks in the flowcharts, may be implemented by
various means, such as hardware, firmware, one or more processors,
circuitry and/or other devices associated with execution of
software including one or more computer program instructions. For
example, one or more of the procedures described above may be
embodied by computer program instructions. In this regard, the
computer program instructions which embody the procedures described
above may be stored by a memory 501 of an apparatus employing an
embodiment of the present invention and executed by a processor 503
in the apparatus. As will be appreciated, any such computer program
instructions may be loaded onto a computer or other programmable
apparatus (e.g., hardware) to produce a machine, such that the
resulting computer or other programmable apparatus provides for
implementation of the functions specified in the flowcharts'
block(s). These computer program instructions may also be stored in
a non-transitory computer-readable storage memory that may direct a
computer or other programmable apparatus to function in a
particular manner, such that the instructions stored in the
computer-readable storage memory produce an article of manufacture,
the execution of which implements the function specified in the
flowcharts' block(s). The computer program instructions may also be
loaded onto a computer or other programmable apparatus to cause a
series of operations to be performed on the computer or other
programmable apparatus to produce a computer-implemented process
such that the instructions which execute on the computer or other
programmable apparatus provide operations for implementing the
functions specified in the flowcharts' block(s). As such, the
operations of FIGS. 3 and 6, when executed, convert a computer or
processing circuitry into a particular machine configured to
perform an example embodiment of the present invention.
Accordingly, the operations of FIGS. 3 and 6 define an algorithm
for configuring a computer or processor, to perform an example
embodiment. In some cases, a general purpose computer may be
provided with an instance of the processor which performs the
algorithm of FIGS. 3 and 6 to transform the general purpose
computer into a particular machine configured to perform an example
embodiment.
[0071] Accordingly, blocks of the flowchart support combinations of
means for performing the specified functions and combinations of
operations for performing the specified functions. It will also be
understood that one or more blocks of the flowcharts', and
combinations of blocks in the flowchart, can be implemented by
special purpose hardware-based computer systems which perform the
specified functions, or combinations of special purpose hardware
and computer instructions.
[0072] In some example embodiments, certain ones of the operations
herein may be modified or further amplified as described below.
Moreover, in some embodiments additional optional operations may
also be included (some examples of which are shown in dashed lines
in FIG. 6). It should be appreciated that each of the
modifications, optional additions or amplifications described
herein may be included with the operations herein either alone or
in combination with any others among the features described
herein.
[0073] Many modifications and other embodiments of the inventions
set forth herein will come to mind to one skilled in the art to
which these inventions pertain having the benefit of the teachings
presented in the foregoing descriptions and the associated
drawings. Therefore, it is to be understood that the inventions are
not to be limited to the specific embodiments disclosed and that
modifications and other embodiments are intended to be included
within the scope of the appended claims. Moreover, although the
foregoing descriptions and the associated drawings describe example
embodiments in the context of certain example combinations of
elements and/or functions, it should be appreciated that different
combinations of elements and/or functions may be provided by
alternative embodiments without departing from the scope of the
appended claims. In this regard, for example, different
combinations of elements and/or functions than those explicitly
described above are also contemplated as may be set forth in some
of the appended claims. Although specific terms are employed
herein, they are used in a generic and descriptive sense only and
not for purposes of limitation.
* * * * *
References