U.S. patent application number 14/286770 was filed with the patent office on 2014-12-04 for learning synonymous object names from anchor texts.
This patent application is currently assigned to Google Inc.. The applicant listed for this patent is Google Inc.. Invention is credited to Jonathan Betz, Krzysztof W. Czuba, Jeffrey C. Reynar.
Application Number | 20140359409 14/286770 |
Document ID | / |
Family ID | 50736617 |
Filed Date | 2014-12-04 |
United States Patent
Application |
20140359409 |
Kind Code |
A1 |
Czuba; Krzysztof W. ; et
al. |
December 4, 2014 |
Learning Synonymous Object Names from Anchor Texts
Abstract
A repository contains objects representing entities. The objects
also include facts about the represented entities. The facts are
derived from source documents. A synonymous name of an object is
determined by identifying a source document from which one or more
facts of the entity represented by the object were derived,
identifying a plurality of linking documents that link to the
source document through hyperlinks, each hyperlink having an anchor
text, processing the anchor texts in the plurality of linking
documents to generate a collection of synonym candidates for the
entity represented by the object, and selecting a synonymous name
for the entity represented by the object from the collection of
synonym candidates.
Inventors: |
Czuba; Krzysztof W.; (New
York, NY) ; Betz; Jonathan; (Summit, NJ) ;
Reynar; Jeffrey C.; (New York, NY) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Google Inc. |
Mountain View |
CA |
US |
|
|
Assignee: |
Google Inc.
Mountain View
CA
|
Family ID: |
50736617 |
Appl. No.: |
14/286770 |
Filed: |
May 23, 2014 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
11833180 |
Aug 2, 2007 |
8738643 |
|
|
14286770 |
|
|
|
|
Current U.S.
Class: |
715/205 |
Current CPC
Class: |
G06F 40/134 20200101;
G06F 16/951 20190101 |
Class at
Publication: |
715/205 |
International
Class: |
G06F 17/22 20060101
G06F017/22 |
Claims
1. A method of determining a synonymous name for an entity
comprising: at a computer system having one or more processors and
memory storing one or more programs configured for execution by the
one or more processors: identifying a source document from which
one or more facts of an entity were derived; identifying a
plurality of linking documents having hyperlinks to the source
document, each hyperlink having an anchor text; generating a
collection of synonym candidates for the entity in accordance with
the anchor texts in the plurality of linking documents; selecting a
synonymous name for the entity from the collection of synonym
candidates; and storing the synonymous name in association with the
entity.
2. The method of claim 1, wherein identifying the source document
further comprises: identifying a list of source documents from
which one or more facts of the entity were derived; removing from
the list identifiers of source documents from which one or more
facts of another entity were derived from the list; and identifying
the source document from the list of source documents.
3. The method of claim 1, wherein generating the collection of
synonym candidates for the entity in accordance with the anchor
texts in the plurality of linking documents further comprises:
normalizing the anchor texts in the plurality of linking documents;
and generating the collection of synonym candidates for the entity
using the normalized anchor texts.
4. The method of claim 3, wherein normalizing the anchor texts in
the plurality of linking documents further comprises: identifying a
language of one of the plurality of linking documents; and applying
normalization rules for the language to normalize the anchor text
in the one of the plurality of linking documents.
5. The method of claim 1, wherein generating the collection of
synonym candidates for the entity in accordance with the anchor
texts in the plurality of linking documents further comprises:
extracting a noun phrase from one of the anchor texts in the
plurality of linking documents; and generating the collection of
synonym candidates for the entity using the extracted noun
phrase.
6. The method of claim 1, wherein generating the collection of
synonym candidates for the entity in accordance with the anchor
texts in the plurality of linking documents further comprises:
removing a prefix or a suffix from the anchor texts in the
plurality of linking documents.
7. The method of claim 1, wherein generating the collection of
synonym candidates for the entity in accordance with the anchor
texts in the plurality of linking documents further comprises:
matching an anchor text with a black list of texts; and responsive
to detecting a match of the anchor text with the black list,
removing the anchor text from the collection of synonym candidates
for the entity.
8. The method of claim 1, wherein generating the collection of
synonym candidates for the entity in accordance with the anchor
texts in the plurality of linking documents further comprises:
matching an anchor text with a white list of texts; and responsive
to detecting a match of the anchor text with the white list, adding
the anchor text into the collection of synonym candidates for the
entity.
9. The method of claim 1, wherein selecting the synonymous name for
the entity from the collection of synonym candidates further
comprises: selecting a synonym candidate occurring at a frequency
in the collection of synonym candidates no less than a minimum
threshold as the synonymous name for the entity.
10. The method of claim 9, further comprising: responsive to a
synonym candidate occurring at a frequency in the collection of
synonym candidates less than the minimum threshold, adding the
synonym candidate into a black list of texts.
11. The method of claim 1, wherein selecting the synonymous name
for the entity from the collection of synonym candidates further
comprises: selecting a synonym candidates occurring at a frequency
in the collection of synonym candidates no more than a maximum
threshold as the synonymous name for the entity.
12. The method of claim 11, further comprising: responsive to a
synonym candidate occurring at a frequency in the collection of
synonym candidates more than the maximum threshold, adding the
synonym candidate into a black list of texts.
13. The method of claim 1, wherein selecting the synonymous name
for the entity from the collection of synonym candidates further
comprises: determining quality for the plurality of linking
documents; and selecting the synonymous name for the entity from
the collection of synonym candidates based on the quality of the
linking document having the anchor text from which the synonymous
name was generated.
14. A system, for determining a synonymous name for an entity,
comprising: one or more processors; and memory storing one or more
programs to be executed by the one or more processors; the one or
more programs comprising instructions for: identifying a source
document from which one or more facts of an entity were derived;
identifying a plurality of linking documents having hyperlinks to
the source document, each hyperlink having an anchor text;
generating a collection of synonym candidates for the entity in
accordance with the anchor texts in the plurality of linking
documents; selecting a synonymous name for the entity from the
collection of synonym candidates; and storing the synonymous name
in association with the entity.
15. The system of claim 14, wherein the instructions for
identifying a source document further comprises: instructions for
identifying a list of source documents from which one or more facts
of the entity were derived; instructions for removing from the list
identifiers of source documents from which one or more facts of
another entity were derived from the list; and instructions for
identifying the source document from the list of source
documents.
16. The system of claim 14, wherein the instructions for generating
the collection of synonym candidates for the entity in accordance
with the anchor texts in the plurality of linking documents further
comprise: instructions for normalizing the anchor texts in the
plurality of linking documents; and instructions for generating the
collection of synonym candidates for the entity using the
normalized anchor texts.
17. The system of claim 14, wherein the instructions for selecting
the synonymous name for the entity from the collection of synonym
candidates further comprises instructions for: determining quality
for the plurality of linking documents; and selecting the
synonymous name for the entity from the collection of synonym
candidates based on the quality of the linking document having the
anchor text from which the synonymous name was generated.
18. A non-transitory computer readable storage medium storing one
or more programs configured for execution by a computer, the one or
more programs comprising instructions for: identifying a source
document from which one or more facts of an entity were derived;
identifying a plurality of linking documents having hyperlinks to
the source document, each hyperlink having an anchor text;
generating a collection of synonym candidates for the entity in
accordance with the anchor texts in the plurality of linking
documents; selecting a synonymous name for the entity from the
collection of synonym candidates; and storing the synonymous name
in association with the entity.
19. The non-transitory computer readable storage medium of claim
18, wherein the instructions for identifying a source document
further includes instructions for: identifying a list of source
documents from which one or more facts of the entity were derived;
removing from the list identifiers of source documents from which
one or more facts of another entity were derived from the list; and
identifying the source document from the list of source
documents.
20. The non-transitory computer readable storage medium of claim
18, wherein the instructions for generating the collection of
synonym candidates for the entity in accordance with the anchor
texts in the plurality of linking documents further includes
instructions for: normalizing the anchor texts in the plurality of
linking documents; and generating the collection of synonym
candidates for the entity using the normalized anchor texts.
Description
RELATED APPLICATIONS
[0001] This application is a continuation of U.S. patent
application Ser. No. 11/833,180, filed Aug. 2, 2007, in which the
application is incorporated by reference herein in its
entirety.
TECHNICAL FIELD
[0002] The disclosed embodiments relate generally to fact
databases. More particularly, the disclosed embodiments relate to
determining names of entities with which objects in a repository
are associated.
BACKGROUND
[0003] As computers and networks gain popularity, web-based
computer documents ("documents") become a vast source of factual
information. Users may look to these documents to get answers to
factual questions, such as "what is the capital of Poland" or "what
is the birth date of George Washington." The factual information
included in these documents may be extracted and stored in a fact
database.
[0004] When extracting facts from documents, facts related to an
entity can be organized together in an object representing the
entity in a repository. The object can use an object name to
identify the represented entity. The object name can be a name of
the represented entity. People often use different names
(hereinafter called "synonymous names") to refer to the same
entity. For example, when a person speaks about "IBM" or "Big
Blue," the audience understands that the speaker is referring to
the International Business Machines Corporation.
[0005] When searching for answers to factual questions in objects,
it is useful to know the synonymous names of the relevant entities.
Users may conduct a search for a question about an entity using one
of its synonymous names (e.g., "IBM"). Objects containing answers
to the question may use a different synonymous name (e.g.,
"International Business Machines Corporation") to identify the same
entity. Because the name used in the objects may not match with the
name used in the search, users may end up not finding the
answers.
[0006] One conventional approach to determining synonymous names of
an object (the synonymous names of the entity represented by the
object) is to consult people familiar with the entity represented
by the object. This approach is insufficient because the vast and
rapidly increasing number of objects in the repository makes it
impractical for any human to perform the task on any meaningful
scale. This conventional approach is also expensive and vulnerable
to human errors.
[0007] For these reasons, what is needed is a way to determine
synonymous names of an object that does not suffer from the
drawbacks described above.
SUMMARY
[0008] The above and other needs are met by methods, systems, and
computer program products that determine synonymous names of an
object. Embodiments of the method comprise identifying a source
document from which one or more facts of an entity represented by
the object were derived, and identifying a plurality of linking
documents having hyperlinks to the source document, each hyperlink
having an anchor text. The method further processes the anchor
texts in the plurality of linking documents to generate a
collection of synonym candidates for the entity represented by the
object, and selects a synonymous name for the entity represented by
the object from the collection of synonym candidates. The method
stores the synonymous name in the repository in association with
the object. Embodiments of the systems and the computer program
products comprise instructions executable by a processor to
implement the methods.
BRIEF DESCRIPTION OF THE DRAWINGS
[0009] FIG. 1 shows a system architecture in accordance to one
embodiment.
[0010] FIGS. 2(a)-2(d) are block diagrams illustrating embodiments
of a data structure for facts within a repository of FIG. 1.
[0011] FIG. 2(e) is a block diagram illustrating an embodiment of
an alternate data structure for facts and objects.
[0012] FIG. 3 is a flow diagram illustrating a method for
determining synonymous names of an object in accordance to one
embodiment.
[0013] FIGS. 4(a)-(e) illustrate an example process of the method
illustrated in FIG. 3.
DESCRIPTION OF EMBODIMENTS
[0014] Embodiments are now described with reference to the figures
where like reference numbers indicate identical or functionally
similar elements.
System Architecture and Data Structure
[0015] FIG. 1 shows a system architecture 100 adapted to support
one embodiment. FIG. 1 shows components used to add facts into, and
retrieve facts from a repository 115. The system architecture 100
includes a network 104, through which any number of document hosts
102 communicate with a data processing system 106, along with any
number of object requesters 152, 154.
[0016] Document hosts 102 store documents and provide access to
documents. A document is comprised of any machine-readable data
including any combination of text, graphics, multimedia content,
etc. A document may be encoded in a markup language, such as
Hypertext Markup Language (HTML), i.e., a web page, in an
interpreted language (e.g., JavaScript) or in any other computer
readable or executable format. A document can include one or more
hyperlinks to other documents. A typical document will include one
or more facts within its content. The facts describe entities, such
as a real-world or fictional people, places, or things.
[0017] A document stored in a document host 102 may be located
and/or identified by a Uniform Resource Locator (URL), or Web
address, or any other appropriate form of identification and/or
location. A document host 102 is implemented by a computer system,
and typically includes a server adapted to communicate over the
network 104 via networking protocols (e.g., TCP/IP), as well as
application and presentation protocols (e.g., HTTP, HTML, SOAP,
D-HTML, JAVA.RTM.). The documents stored by a host 102 are
typically held in a file directory, a database, or other data
repository. A host 102 can be implemented in any computing device
(e.g., from a PDA or personal computer, a workstation,
mini-computer, or mainframe, to a cluster or grid of computers), as
well as in any processor architecture or operating system.
[0018] FIG. 1 shows components used to manage facts in a fact
repository 115. The data processing system 106 includes one or more
importers 108, one or more janitors 110, a build engine 112, a
service engine 114, and a fact repository 115 (also called simply a
"repository"). Each of the foregoing are implemented, in one
embodiment, as software modules (or programs) executed by the
processor 116. Importers 108 operate to process documents received
from the document hosts, read the data content of documents, and
extract facts (as operationally and programmatically defined within
the data processing system 106) from such documents. The importers
108 also determine the subject or subjects (i.e., the entity or
entities) with which the facts are associated, and extract such
facts into individual items of data, for storage in the repository
115. In one embodiment, there are different types of importers 108
for different types of documents, for example, dependent on the
format or document type.
[0019] Janitors 110 operate to process facts extracted by the
importer 108. This processing can include but is not limited to,
data cleansing, object merging, and fact induction. In one
embodiment, there are a number of different janitors 110 that
perform different types of data management operations on the facts.
For example, one janitor 110 may traverse some set of facts in the
repository 115 to find duplicate facts (that is, facts that convey
the same factual information) and merge them. Another janitor 110
may also normalize facts into standard formats. Another janitor 110
may also remove unwanted facts from the repository 115, such as
facts related to pornographic content. Other types of janitors 110
may be implemented, depending on the types of data management
functions desired, such as translation, compression, spelling or
grammar correction, and the like.
[0020] Various janitors 110 act on facts to normalize attribute
names, and values and delete duplicate and near-duplicate facts so
an object does not have redundant information. For example, we
might find on one page that Britney Spears' birthday is "Dec. 2,
1981" while on another page that her date of birth is "Dec. 2,
1981." Birthday and Date of Birth might both be rewritten as
"Birthdate" by one janitor and then another janitor might notice
that Dec. 2, 1981 and Dec. 2, 1981 are different forms of the same
date. It would choose the preferred form, remove the other fact and
combine the source lists for the two facts. As a result, one source
page for this fact will contain an exact match of the fact while
another source page will contain text that is considered synonymous
with the fact.
[0021] The build engine 112 builds and manages the repository 115.
The service engine 114 is an interface for querying the repository
115. The service engine 114's main function is to process queries,
score matching objects, and return them to the caller but it is
also used by the janitor 110.
[0022] The repository 115 stores factual information about
entities. The information is extracted from a plurality of
documents that are located on document hosts 102. A document from
which a particular fact may be extracted is a source document (or
"source") of that particular fact. In other words, a source of a
fact includes that fact (or a synonymous fact) within its
contents.
[0023] The repository 115 contains one or more facts. In one
embodiment, the facts are logically organized into "objects," and
each object contains a collection of facts associated with a single
entity (i.e., real-world or fictional person, place, or thing).
Each fact is associated with exactly one object. One implementation
for this association includes in each fact an object ID that
uniquely identifies the associated object. In this manner, any
number of facts may be associated with an individual object, by
including the object ID for that object in the facts. In one
embodiment, objects themselves are not physically stored in the
repository 115, but rather are defined by the set or group of facts
with the same associated object ID, as described below. Further
details about facts in the repository 115 are described below, in
relation to FIGS. 2(a)-2(d).
[0024] Some embodiments operate on the facts and/or objects in
different orders than described above. For example, in one
embodiment the importer 108 provides facts directly to the build
engine 112 and/or repository 115. The janitors 110, in turn,
operate on the facts and/or objects in the repository 115. It
should also be appreciated that in practice at least some of the
components of the data processing system 106 will be distributed
over multiple computers, communicating over a network. For example,
the repository 115 may be deployed over multiple servers. As
another example, the janitors 110 may be located on any number of
different computers. For convenience of explanation, however, the
components of the data processing system 106 are discussed as
though they were implemented on a single computer.
[0025] In another embodiment, some or all of document hosts 102 are
located on the data processing system 106 instead of being coupled
to the data processing system 106 by a network. For example, the
importer 108 may import facts from a database that is a part of or
associated with the data processing system 106.
[0026] FIG. 1 also includes components to access the repository 115
on behalf of one or more object requesters 152, 154. Object
requesters are entities that request objects from the repository
115. Object requesters 152, 154 may be understood as clients of the
system 106, and can be implemented in any computer device or
architecture. As shown in FIG. 1, a first object requester 152 is
located remotely from system 106, while a second object requester
154 is located in the data processing system 106. For example, in a
computer system hosting a blog, the blog may include a reference to
an object whose facts are in the repository 115. An object
requester 152, such as a browser displaying the blog, will access
data processing system 106 so that the information of the facts
associated with the object can be displayed as part of the blog web
page. As a second example, a janitor 110 or other entity considered
to be part of data processing system 106 can function as an object
requester 154, requesting the facts of objects from the repository
115.
[0027] FIG. 1 shows that the data processing system 106 includes a
memory 107 and one or more processors 116. The memory 107 includes
the importers 108, janitors 110, build engine 112, service engine
114, and requester 154, each of which is preferably implemented as
instructions stored in memory 107 and executable by processor 116.
Memory 107 also includes the repository 115. The repository 115 can
be stored in a memory of one or more computer systems or in a type
of memory such as a disk. FIG. 1 also includes a computer readable
storage medium 118 containing, for example, at least one of
importers 108, janitors 110, the build engine 112, the service
engine 114, the requester 154, and at least some portions of the
repository 115. FIG. 1 also includes one or more input/output
devices 120 that allow data to be input and output to and from the
data processing system 106. It will be understood that embodiments
of the data processing system 106 also include standard software
components such as operating systems and the like and further
include standard hardware components not shown in the figure for
clarity of example.
[0028] FIG. 2(a) shows an example format of a data structure for
facts within the repository 115, according to some embodiments. As
described above, the repository 115 includes facts 204 describing
entities such as real-world and fictional people, places, and
things. Each fact 204 includes a unique identifier for that fact,
such as a fact ID 210. Each fact 204 includes at least an attribute
212 and a value 214. For example, a fact associated with the entity
George Washington may include an attribute of "date of birth" and a
value of "Feb. 22, 1732." In one embodiment, all facts are stored
as alphanumeric characters since they are extracted from web pages.
In another embodiment, facts also can store binary data values.
Other embodiments, however, may store fact values as mixed types,
or in encoded formats.
[0029] As described above, each fact is associated with an object
ID 209 that identifies the object with which the fact is
associated. Thus, each fact that describes the same entity (such as
George Washington), will have the same object ID 209. In one
embodiment, the objects are logical concepts that exist as a
collection of facts having the same object ID. In another
embodiment, objects are stored as units of data in memory, and
include references (for example, pointers or IDs) to the facts
associated with the object. The logical data structure of a fact
can take various forms; in general, a fact is represented by a
tuple that includes a fact ID, an attribute, a value, and an object
ID. The storage implementation of a fact can be in any underlying
physical data structure.
[0030] FIG. 2(b) shows an example of facts having respective fact
IDs of 10, 20, and 30 in the repository 115. Facts 10 and 20 are
associated with an object identified by object ID "1." Fact 10 has
an attribute of "Name" and a value of "China." Fact 20 has an
attribute of "Category" and a value of "Country." Thus, the object
identified by object ID "1" has a name fact 205 with a value of
"China" and a category fact 206 with a value of "Country." Fact 30
208 has an attribute of "Property" and a value of" "Bill Clinton
was the 42nd President of the United States from 1993 to 2001."
Thus, the object identified by object ID "2" has a property fact
with a fact ID of 30 and a value of "Bill Clinton was the 42nd
President of the United States from 1993 to 2001." In the
illustrated embodiment, each fact has one attribute and one value.
The number of facts associated with an object is not limited; thus
while only two facts are shown for the "China" object, in practice
there may be dozens, even hundreds of facts associated with a given
object. Also, the value fields of a fact need not be limited in
size or content. For example, a fact about the economy of "China"
with an attribute of "Economy" would have a value including several
paragraphs of text, numbers, and perhaps even tables of figures.
This content can be formatted, for example, in a markup language.
For example, a fact having an attribute "original html" might have
a value of the original html text taken from the source web
page.
[0031] Also, while the illustration of FIG. 2(b) shows the explicit
coding of object ID, fact ID, attribute, and value, in practice the
content of the fact can be implicitly coded as well (e.g., the
first field being the object ID, the second field being the fact
ID, the third field being the attribute, and the fourth field being
the value). Other fields include but are not limited to: the
language used to state the fact (English, etc.), how important the
fact is, the source of the fact, a confidence value for the fact,
and so on.
[0032] FIG. 2(c) shows an example object reference table 210 that
is used in some embodiments. Not all embodiments include an object
reference table. The object reference table 210 functions to
efficiently maintain the associations between object IDs and fact
IDs. In the absence of an object reference table 210, it is also
possible to find all facts for a given object ID by querying the
repository 115 to find all facts with a particular object ID. While
FIGS. 2(b) and 2(c) illustrate the object reference table 210 with
explicit coding of object and fact IDs, the table also may contain
just the ID values themselves in column or pair-wise
arrangements.
[0033] FIG. 2(d) shows an example of a data structure for facts
within the repository 115, according to some embodiments, showing
an extended format of facts. In this example, the fields include an
object reference link 216 to another object. The object reference
link 216 can be an object ID of another object in the repository
115, or a reference to the location (e.g., table row) for the
object in the object reference table 210. The object reference link
216 allows facts to have as values other objects. For example, for
an object associated with the entity "United States," there may be
a fact with the attribute of "president" and the value of "George
W. Bush," with "George W. Bush" being an object having its own
facts in the repository 115. In some embodiments, the value field
214 stores the name of the linked object and the link 216 stores
the object identifier of the linked object. Thus, this "president"
fact would include the value 214 of "George W. Bush", and an object
reference link 216 that contains the object ID for the "George W.
Bush" object. In some other embodiments, facts 204 do not include a
link field 216 because the value 214 of a fact 204 may store a link
to another object.
[0034] Each fact 204 also may include one or more metrics 218. A
metric provides an indication of the quality of the fact. In some
embodiments, the metrics include a confidence level and an
importance level. The confidence level indicates the likelihood
that the fact is correct. The importance level indicates the
relevance of the fact to the object, compared to other facts for
the same object. The importance level may optionally be viewed as a
measure of how vital a fact is to an understanding of the entity
associated with the object.
[0035] Each fact 204 includes a list of one or more sources 220
that include the fact and from which the fact was extracted. Each
source may be identified by a URL, or Web address, or any other
appropriate form of identification and/or location, such as a
unique document identifier.
[0036] The facts illustrated in FIG. 2(d) include an agent field
222 that identifies the importer 108 that extracted the fact. For
example, the importer 108 may be a specialized importer that
extracts facts from a specific source (e.g., the pages of a
particular web site, or family of web sites) or type of source
(e.g., web pages that present factual information in tabular form),
or an importer 108 that extracts facts from free text in documents
throughout the Web, and so forth.
[0037] Some embodiments include one or more specialized facts, such
as a name fact 207 and a property fact 208. A name fact 207 is a
fact that conveys a name for the entity associated with the object
in which the fact is included. A name fact 207 includes an
attribute 224 of "name" and a value, which is the name of the
associated entity. For example, for an object associated with
country Spain, a name fact would have the value "Spain." A name
fact 207, being a special instance of a general fact 204, includes
the same fields as any other fact 204; it has an attribute, a
value, a fact ID, metrics, sources, etc. The attribute 224 of a
name fact 207 indicates that the fact is a name fact, and the value
is the actual name. The name may be a string of characters. An
object may have one or more associated name facts, as many entities
can have more than one name. For example, an object associated with
Spain may have associated name facts conveying the country's common
name "Spain" and the official name "Kingdom of Spain." As another
example, an object associated with the U.S. Patent and Trademark
Office may have associated name facts conveying the agency's
acronyms "PTO" and "USPTO" as well as the official name "United
States Patent and Trademark Office." If an object does have more
than one associated name fact, one of the name facts may be
designated as a primary name and other name facts may be designated
as secondary names, either implicitly or explicitly. The name facts
associated with an object are also called synonymous names of the
object.
[0038] A property fact 208 is a fact that conveys a statement about
the entity associated with the object. Property facts are generally
used for summary information about an object. A property fact 208,
being a special instance of a general fact 204, also includes the
same fields (such as attribute, value, fact ID, etc.) as other
facts 204. The attribute field 226 of a property fact 208 indicates
that the fact is a property fact (e.g., attribute is "property")
and the value is a string of text that conveys the statement of
interest. For example, for the object associated with Bill Clinton,
the value of a property fact may be the text string "Bill Clinton
was the 42nd President of the United States from 1993 to 2001."
Some objects may have one or more associated property facts while
other objects may have no associated property facts. It should be
appreciated that the data structures shown in FIGS. 2(a)-2(d) and
described above are merely exemplary. The data structure of the
repository 115 may take on other forms. Other fields may be
included in facts and some of the fields described above may be
omitted. Additionally, each object may have additional special
facts aside from name facts and property facts, such as facts
conveying a type or category (for example, person, place, movie,
actor, organization, etc.) for categorizing the entity associated
with the object. In some embodiments, an object's name(s) and/or
properties may be represented by special records that have a
different format than the general fact records 204.
[0039] As described previously, a collection of facts is associated
with an object ID of an object. An object may become a null or
empty object when facts are disassociated from the object. A null
object can arise in a number of different ways. One type of null
object is an object that has had all of its facts (including name
facts) removed, leaving no facts associated with its object ID.
Another type of null object is an object that has all of its
associated facts other than name facts removed, leaving only its
name fact(s). Alternatively, the object may be a null object only
if all of its associated name facts are removed. A null object
represents an entity or concept for which the data processing
system 106 has no factual information and, as far as the data
processing system 106 is concerned, does not exist. In some
embodiments, facts of a null object may be left in the repository
115, but have their object ID values cleared (or have their
importance set to a negative value). However, the facts of the null
object are treated as if they were removed from the repository 115.
In some other embodiments, facts of null objects are physically
removed from the repository 115.
[0040] FIG. 2(e) is a block diagram illustrating an alternate data
structure 290 for facts and objects in accordance with embodiments
of the invention. In this data structure, an object 290 contains an
object ID 292 and references or points to facts 294. Each fact
includes a fact ID 295, an attribute 297, and a value 299. In this
embodiment, an object 290 actually exists in memory 107.
[0041] As described above, an object may explicitly exist in the
repository 115, or it may exist merely as a collection of facts
with a common object ID. Reference is made to particular objects
for the purposes of illustration; one of skill in the art will
recognize that the systems and methods described herein are
applicable to a variety of implementations and that such references
are not limiting. When reference is made to a fact being associated
with an object, it should be understood that in at least one
embodiment a fact is associated with an object by sharing a common
object ID with other facts. For example, a fact could be associated
with an object of a given type by sharing a common object ID at
least with a type fact indicating the given type (or as another
example, with a category fact indicating a particular category of
object). Furthermore, in various embodiments, facts and objects can
be stored in a variety of structures, such as fact and/or object
repositories. When reference is made herein to the repository 115,
it should be understood that various embodiments may store facts
and/or objects in a variety of data structures.
Overview of Methodology
[0042] Referring now to FIG. 3, there is shown a flow diagram
illustrating a method 300 for determining synonymous names of an
object in accordance with one embodiment. Other embodiments perform
steps of the method 300 in different orders and/or perform
different or additional steps than the ones shown in FIG. 3. The
steps of the method 300 may be implemented in software, hardware,
or a combination of hardware and software.
[0043] In one embodiment, the steps of the method 300 may be
performed by the data processing system ("system") 106 as shown in
FIG. 1, although one skilled in the art will recognize that the
method 300 could be performed by systems having different
architectures as well. The system 106 can perform multiple
instances of the steps of the method 300 concurrently and/or
perform steps in parallel.
[0044] The method 300 will now be described in detail. An object
representing (or describing) an entity is identified 308 from the
repository 115. As described above, each object can be identified
by a unique object ID and is defined by the collection of facts
associated with the object ID.
[0045] The system 106 identifies 310 a list of source documents
associated with the object. A source document associated with an
object is a document from which one or more facts of the object was
extracted (or derived). A source document can be located and/or
identified by a unique identifier such as a URL. In one embodiment,
each fact in the repository 115 includes a field for unique
identifiers of associated source documents (hereinafter called the
source field). The system 106 can retrieve the facts associated
with the object from the repository 115 using the object reference
table 210 as described above with reference to FIG. 2(c). After
retrieving the facts associated with the object, the system 106 can
identify 310 the list of source documents associated with the
object based on the source fields of the retrieved facts. A fact
can have multiple source documents.
[0046] The subject of a source document tends to be the entity
represented by the object associated with the source document. This
is because the content of the source document includes at least one
fact about the entity. However, a source document may have more
than one subject. For example, a source document may be a BLOG
covering a broad range of topics, the entity represented by the
associated object being one of them. In one embodiment, the system
106 removes source documents having multiple subjects from the
identified list of source documents. The system 106 may construct a
list of source documents and their associated objects. If a source
document associates with objects representing different entities,
the system 106 can remove the source document from the identified
list of source documents.
[0047] For each of the source documents in the identified list, the
system identifies 320 linking documents containing hyperlinks to
the source document. As described above, a document can include one
or more hyperlinks to other documents. Therefore, a linking
document may also include hyperlinks to documents other than to the
source document. However, the hyperlink that is of interest to the
system 106 is the hyperlink to the source document. As used herein,
a linking document containing a hyperlink to a source document is
called a linking document for the source document.
[0048] A hyperlink includes a starting anchor tag, which includes
one or more parameters (or markup attributes), and an ending anchor
tag. The starting and ending anchor tags define the hyperlink. A
hypertext reference attribute (e.g., "HREF") is one type of markup
attribute. The hypertext reference attribute indicates that the
associated value is the address of the destination of the
hyperlink. The text between the starting anchor tag and the ending
anchor tag is called the anchor text of the hyperlink. For example,
in the following hyperlink,
[0049] <a href="http://www.cnn.com/">CNN</a>
"<a href="http://www.cnn.com/">" is the starting anchor tag,
"CNN" is the anchor text, and "</a>" is the ending anchor
tag. In the starting anchor tag, "href" is the hypertext reference
attribute and "http://www.cnn.com/" is the associated value and the
address of the destination of the hyperlink (the destination
address). The anchor text, if clicked by a user, triggers a request
(e.g., an HTTP request) for a document located at the destination
address (the destination document). Because an anchor text is
rendered for presentation for the destination document to a user,
it tends to reflect the subject of the destination document.
Therefore, the anchor text of a hyperlink in a linking document for
a source document tends to reflect the subject of the source
document. As used herein, the anchor text of a hyperlink in a
linking document for a source document is called the anchor text
for the source document in the linking document. It is noted that
methods of linking documents other than using hyperlinks can also
be used and the described process can be readily applied to these
other methods.
[0050] The system 106 processes 330 the anchor texts in the linking
documents to generate a collection of synonym candidates (also
known as the "anchor synset") for the object name. This step is
designed to remove those anchor texts that are not related to the
subject of the associated source document (e.g. "Click here!") and
to clean up the remaining anchor texts (e.g., removing portions of
an anchor text unrelated to the subject of the associated source
document). The results of the processing 330 are the collection of
synonym candidates, each of which is intended to describe the
entity represented by the object.
[0051] As discussed above, anchor texts for a source document tend
to reflect the subject of the source document, which in turn tends
to describe the entity represented by the associated object. It
follows that the anchor texts for a source document tend to
describe the entity represented by the object associated with the
source document. Thus, the system 106 can generate synonym
candidates of the object name from anchor texts for the associated
source documents.
[0052] For example, authors of linking documents can associate the
anchor text "Big Blue" or "IBM" with a hyperlink to a source
document about the International Business Machines Corporation,
which is a source document of an object representing the
International Business Machines Corporation. Because the subject of
the source document (the International Business Machines
Corporation) correctly describes the entity represented by the
object, and the anchor texts for the source document reflect its
subject, the anchor texts ("Big Blue" and "IBM") tend to be valid
synonymous names of the object.
[0053] Some anchor texts may be invalid synonymous names of the
object. Authors of linking documents can use generic language as
the anchor text (e.g., "click here," "see wikipedia article") or
include in the anchor texts language that is not related to the
subject of the associated linking document (e.g., "click here for
an excellent article about IBM"). An anchor text including both
information describing the subject of the associated source
document and unrelated information (e.g., "click here for an
excellent article about") is called a partially-related anchor
text. An anchor text that includes only unrelated information
(e.g., "Click here!") is called an unrelated anchor text. The
system 106 processes 330 anchor texts in the identified linking
documents to remove unrelated anchor texts and unrelated
information from partially-related anchor texts.
[0054] In one embodiment, the system 106 removes unrelated
information from a partially-related anchor text by extracting a
noun phrase (or a noun) from the anchor text. Because the synonym
candidates are intended to describe the object, they are either
nouns (e.g., "Canada") or noun phrases (e.g., "the International
Business Machines Corporation"). By extracting a noun phrase from
the anchor text, the system 106 identifies a portion of the anchor
text that tends to be relevant to the object. For example, the noun
phrase of an anchor text "Learn about Google Inc." is "Google Inc."
By extracting a noun phrase from an anchor text, the system 106
removes text unrelated to the subject of the associated source
document (e.g., "Learn about" in the above example). The system 106
can then add the extracted noun phrase into the anchor synset.
[0055] In one embodiment, the system 106 has a collection of texts
that are commonly used in anchor texts and unrelated to subjects of
the associated destination documents (hereinafter called the "black
list"). This black list can be compiled by domain experts (e.g.,
administrators of the system 106) or automatically generated by the
system 106. The black list can include standard anchor texts such
as "here," "click here," "download," and the like. The system 106
can remove an unrelated anchor text by matching it with texts in
the black list and not adding it into the anchor synset if a match
is detected.
[0056] The black list may also contain prefix and suffix texts. The
system 106 may remove unrelated information from a
partially-related anchor text by matching it with the prefix and/or
suffix texts and remove the matched prefix and/or suffix from it.
For example, "Wikipedia article about" can be a frequently used
prefix and included in the black list. The system 106 identifies
that an anchor text "Wikipedia article about the King" includes the
prefix and removes it from the anchor text. Other popular prefix
texts include "See Wikipedia for," "article for," and the like. The
system 106 can then add the remainder into the anchor synset.
[0057] In one embodiment, the system 106 applies normalization
rules to an anchor text to standardize its format before processing
330. Examples of the normalization rules include removal of
punctuation, such as removing commas in a string, conversion of
uppercase characters in a string to corresponding lowercase
characters, such as from "America" to "america," and stop word
removal, such as removing stop words such as "the," "a," and "of"
from a string. For example, after applying the above normalization
rules, an anchor text "Click here!" becomes "click here."
Subsequently, the system 106 detects a match in the black list for
the normalized anchor text and does not add it to the anchor
synset. In one embodiment, the system 106 applies
language-dependent normalization rules based on the language of the
linking document. For example, the system 106 can identify Spanish
as the language of a linking document, and apply a set of Spanish
stop word removal rules to the anchor texts in the linking
document.
[0058] Similarly, the system 106 can have a collection of texts
that tend to be valid synonymous names (hereinafter called "white
list"). This white list can be compiled by authorized personnel or
imported from one or more information sources. For example, the
white list can contain company names extracted from a
business-related website or peoples' names from a telephone
directory. The system 106 can process 330 the anchor texts by
matching them with the texts in the white list and add those that
match into the anchor synset. By adding the anchor texts that match
with an entry in the white list, the system 106 generates synonym
candidates that tend to be valid synonymous names.
[0059] The system 106 selects 340 synonymous names of the object
from the collection of synonym candidates (the anchor synset). The
synonym candidates generated may contain identification language
that does not qualify as synonymous names. For example, some of the
synonym candidates can reflect the author's personal opinion (e.g.,
"my favorite movie star"), while some others can be descriptive
names used exclusively within a small group of people (e.g., "Party
Ed"). The system selects 340 synonymous names by filtering out
these invalid synonym candidates.
[0060] In one embodiment, the system 106 selects 340 synonymous
names based on the frequency of occurrence of the synonym
candidates within the anchor synset. For example, the system 106
can be configured to select 340 synonym candidates that occur at a
frequency above a minimum threshold. The threshold can be
user-defined or dynamically modified by the system 106. Rarely
occurring synonym candidates tend to be incorrect synonymous names
(e.g., containing spelling errors). Further, even if the rarely
occurring synonym candidates contain legitimate synonym names,
these legitimate synonym names are rarely used and can be omitted
with minimal harm. Examples of synonym candidates that occur
infrequently include authors' personal opinions and names used only
by a small group of people.
[0061] The system 106 can also be configured to select 340 the
synonym candidates that occur at a frequency below a maximum
threshold. Synonym candidates that occur extremely frequently also
tend not to be synonymous names because they can be general phrases
unrelated to the subject of the associated source document.
Examples of synonym candidates with extremely high occurrence rates
include "the company," "home page," and "click here." In some
embodiments, the system 106 can add a synonym candidate that occurs
frequently into the black list. Alternatively, the system 106 can
output the synonym candidates occurring at a frequency exceeding
the maximum threshold so that an administrator can review them and
select 340 them if they are removed by mistake.
[0062] In one embodiment, instead of selecting 340 synonymous names
based on the frequency of occurrence of the synonym candidates, the
system 106 selects 340 synonymous names based on the proportion of
the synonym candidates in the collection of synonym candidates. For
example, the system 106 can be configured to select 340 only the
synonym candidates that constitute more than 5% of the total anchor
synset.
[0063] In one embodiment, the system 106 selects 340 synonymous
names based on the quality of the associated linking documents. The
quality of a linking document can be user defined or machine
generated. For example, the system 106 can determine the quality of
the linking document based on a page rank of the linking document.
A page rank is a numerical weight for a document determined by a
link analysis algorithm such as the algorithm described in U.S.
Pat. No. 6,285,999. Because high quality linking documents tend to
have high quality anchor texts (e.g., fewer spelling errors),
synonym candidates generated from these documents are more likely
to be proper synonymous names of the object. In some embodiments,
the system 106 adds synonymous names generated from high-quality
documents into the white list.
[0064] In one embodiment, the system 106 assigns a score for each
synonym candidate in the anchor synset, and selects 340 synonymous
names based on the assigned scores. The score for a synonym
candidate may be determined based on a score function taking into
account one or more of the following factors: the frequency of
occurrence or the proportion of the synonym candidate within the
anchor synset, the quality of the associated linking documents,
whether the synonym candidate has a match in the white list or the
black list, and whether the synonym candidate is properly
capitalized (e.g., whether the first character of each word in the
synonym candidate and only these characters are capitalized). The
system 106 may select 340 a synonym candidate as a synonymous name
if its score above a minimum threshold and/or below a maximum
threshold.
[0065] The system 106 can process 330 the anchor texts associated
with different source documents separately and generate a
collection of synonym candidates for each of the source documents.
The system 106 can then select 340 synonymous names from each of
the collections. Alternatively, the system 106 can process 330 the
anchor texts associated with different source documents together
and generate one single anchor synset, and subsequently select 340
synonymous names from the anchor synset.
[0066] After selecting 340 the synonymous names, the system 106
adds 350 facts including the selected synonymous names to the
object in the repository 115. For example, the system 106 may
create a name fact for each of the selected synonymous names,
assign the synonymous names as the values of the name facts, and
associate the created facts with the object.
[0067] After the system 106 determines the synonymous names of the
object, it can process the other objects in the repository 115 and
determine their synonymous names. As noted above, the system can
determine the synonymous names of multiple objects concurrently
and/or in parallel.
[0068] It is noted that the process described above is
language-neutral and can be used to determine synonymous names in
any language.
Example Process
[0069] FIGS. 4(a) through 4(e) illustrate an example process of the
method 300 described above with respect to FIG. 3. Initially, as
illustrated in FIG. 4(a), the system 106 identifies 308 an object
402 representing the rock singer Elvis Presley. The object 402 has
three associated facts 410, 412, and 414. As further illustrated in
FIG. 4(b), the fact 410 has an attribute of "Name," a value of
"Elvis Aaron Presley," and a source of
"http://www.elvis.com/elvisology/bio/elvis_overview.asp." The fact
412 has an attribute of "Date of Birth," a value of "Jan. 8, 1935,"
and a source of "http://en.wikipedia.org/wiki/Elvis." The fact 414
has an attribute of "Origin," a value of "East Tupelo, Miss.,
United States," and a source of
"http://www.history-of-rock.com/elvis.sub.--presley.htm." The name
of the object 402 is the value of the fact 410, "Elvis Aaron
Presley."
[0070] The system 106 identifies 310 source documents associated
with the object 402 by identifying facts associated with the object
402, the facts 410, 412, and 414, retrieving these facts from the
repository 115, and identifying 310 the associated source documents
by accessing source fields of the retrieved facts. The fact 410 is
associated with a source document 420. The fact 412 is associated
with a source document 422. The fact 414 is associated with a
source document 424. Therefore, the system 106 identifies 310
source documents associated with the object 402 as the source
documents 420, 422, and 424. As illustrated in FIG. 4(b), the URLs
of the source documents 420, 422, and 424 are
"http://www.elvis.com/elvisology/bio/elvis_overview.asp,"
"http://en.wikipedia.org/wiki/Elvis," and
"http://www.history-of-rock.com/elvis_presley.htm,"
respectively.
[0071] The system 106 identifies 320 linking documents for the
source documents 420, 422, and 424. The system 106 identifies 320
two linking documents 430 and 432 for the source document 420,
three linking documents 432, 434, and 436 for the source document
422, and two linking documents 438 and 440 for the source document
424. It is noted that the linking document 432 is identified as a
linking document for both the source documents 420 and 422.
[0072] Referring now to FIG. 4(c), the column labeled "Linking
Document ID" (linking document column) contains the document
identifiers of linking documents identified 320 by the system 106.
The column labeled "Source Document ID" (source document column)
contains the document identifier of the source document to which
the linking document as identified in the linking document column
links. The column "Anchor Text" contains the anchor text for the
associated source document in the associated linking document. As
shown in FIG. 4(c), the anchor text for the source document 420 in
the linking document 430 is "An article about the King." The anchor
text for the source document 420 in the linking document 432 is
"Homepage of The King." The anchor text for the source document 422
in the linking document 432 is "Learn about Elvis." The anchor text
for the source document 422 in the linking document 434 is
"Wikipedia article about the King." The anchor text for the source
document 422 in the linking document 436 is "Click here!" The
anchor text for the source document 424 in the linking document 438
is "The Best Rock Singer Ever." The anchor text for the source
document 424 in the linking document 440 is "Find out more about
Elvis."
[0073] The system 106 processes 330 the anchor texts in the linking
documents 430, 432, 434, 436, 438, and 440 and generates a
collection of synonym candidates for the object 402. Referring to
FIG. 4(d), the system 106 processes 330 the anchor text in the
linking document 430 ("An article about the King") by removing the
common prefix "An article about" and generates a synonym candidate
"the King;" processes 330 the anchor text for the source document
420 in the linking document 432 ("Homepage of The King") by
removing the common prefix "Homepage of and generates a synonym
candidate "The King;" processes 330 the anchor text for the source
document 422 in the linking document 432 ("Learn about Elvis") by
removing the common prefix "Learn about" and generates a synonym
candidate "Elvis;" processes 330 the anchor text in the linking
document 434 ("Wikipedia article about the King") by removing the
common prefix "Wikipedia article about" and generates a synonym
candidate "the King;" processes 330 the anchor text in the linking
document 440 ("Find out more about Elvis") by removing the common
prefix "Find out more about" and generates a synonym candidate
"Elvis." The system 106 generates a synonym candidate based on the
anchor text in the linking document 438 ("The Best Rock Singer
Ever"). The system 106 detects a match in the black list for the
anchor text in the linking document 436 ("Click here!") and does
not generate any synonym candidates based on it.
[0074] The system 106 selects 340 synonymous names from the
collection of synonym candidates. Referring to FIG. 4(d), the
collection of the synonym candidates includes "the King," "The
King," "Elvis," "the King," "The Best Rock Singer Ever," and
"Elvis." The system 106 selects 340 the synonym candidates
occurring no less than twice (minimum threshold) and no more than
one hundred times (maximum threshold). Assuming the system 106 is
case insensitive, it selects 340 the synonym candidates "the King,"
which occurs three times, and "Elvis," which occurs twice. The
synonym candidate "The Best Rock Singer Ever" has only one
occurrence, smaller than the minimum threshold, and thus is not
selected 340. Therefore, the system 106 correctly identifies the
synonymous names "The King" and "Elvis" for the object 402
representing Elvis Presley.
[0075] The system 106 adds 350 two facts to the object 402. As
illustrated in FIG. 4(e), the system 106 creates a fact 416 for the
synonymous name "The King" and a fact 418 for the synonymous name
"Elvis." The source field of the facts 416, 418 shown in FIG. 4(e)
is empty. However, the system 106 may list the URLs of the linking
documents from which the synonymous names are derived in the
corresponding source field. For example, the source field of the
name fact 416 may include URLs for the linking documents 430, 432,
and 434.
[0076] Reference in the specification to "one embodiment" or to "an
embodiment" means that a particular feature, structure, or
characteristic described in connection with the embodiments is
included in at least one embodiment of the invention. The
appearances of the phrase "in one embodiment" in various places in
the specification are not necessarily all referring to the same
embodiment.
[0077] Some portions of the above are presented in terms of
algorithms and symbolic representations of operations on data bits
within a computer memory. These algorithmic descriptions and
representations are the means used by those skilled in the data
processing arts to most effectively convey the substance of their
work to others skilled in the art. An algorithm is here, and
generally, conceived to be a self-consistent sequence of steps
(instructions) leading to a desired result. The steps are those
requiring physical manipulations of physical quantities. Usually,
though not necessarily, these quantities take the form of
electrical, magnetic or optical signals capable of being stored,
transferred, combined, compared and otherwise manipulated. It is
convenient at times, principally for reasons of common usage, to
refer to these signals as bits, values, elements, symbols,
characters, terms, numbers, or the like. Furthermore, it is also
convenient at times, to refer to certain arrangements of steps
requiring physical manipulations of physical quantities as modules
or code devices, without loss of generality.
[0078] It should be borne in mind, however, that all of these and
similar terms are to be associated with the appropriate physical
quantities and are merely convenient labels applied to these
quantities. Unless specifically stated otherwise as apparent from
the following discussion, it is appreciated that throughout the
description, discussions utilizing terms such as "processing" or
"computing" or "calculating" or "determining" or "displaying" or
"determining" or the like, refer to the action and processes of a
computer system, or similar electronic computing device, that
manipulates and transforms data represented as physical
(electronic) quantities within the computer system memories or
registers or other such information storage, transmission or
display devices.
[0079] Certain aspects of the present invention include process
steps and instructions described herein in the form of an
algorithm. It should be noted that the process steps and
instructions of the present invention can be embodied in software,
firmware or hardware, and when embodied in software, can be
downloaded to reside on and be operated from different platforms
used by a variety of operating systems.
[0080] The present invention also relates to an apparatus for
performing the operations herein. This apparatus may be specially
constructed for the required purposes, or it may comprise a
general-purpose computer selectively activated or reconfigured by a
computer program stored in the computer. Such a computer program
may be stored in a computer readable storage medium, such as, but
is not limited to, any type of disk including floppy disks, optical
disks, CD-ROMs, magnetic-optical disks, read-only memories (ROMs),
random access memories (RAMs), EPROMs, EEPROMs, magnetic or optical
cards, application specific integrated circuits (ASICs), or any
type of media suitable for storing electronic instructions, and
each coupled to a computer system bus. Furthermore, the computers
referred to in the specification may include a single processor or
may be architectures employing multiple processor designs for
increased computing capability.
[0081] The algorithms and displays presented herein are not
inherently related to any particular computer or other apparatus.
Various general-purpose systems may also be used with programs in
accordance with the teachings herein, or it may prove convenient to
construct more specialized apparatus to perform the required method
steps. The required structure for a variety of these systems will
appear from the description below. In addition, the present
invention is not described with reference to any particular
programming language. It will be appreciated that a variety of
programming languages may be used to implement the teachings of the
present invention as described herein, and any references below to
specific languages are provided for disclosure of enablement and
best mode of the present invention.
[0082] While the invention has been particularly shown and
described with reference to a preferred embodiment and several
alternate embodiments, it will be understood by persons skilled in
the relevant art that various changes in form and details can be
made therein without departing from the spirit and scope of the
invention.
[0083] Finally, it should be noted that the language used in the
specification has been principally selected for readability and
instructional purposes, and may not have been selected to delineate
or circumscribe the inventive subject matter. Accordingly, the
disclosure of the present invention is intended to be illustrative,
but not limiting, of the scope of the invention, which is set forth
in the following claims.
* * * * *
References