Patent Application
20140280013
Web pages, data feeds, electronic articles, electronic books, or other forms of electronic content can include a mix of facts about different people, places, objects, events, or other subjects. For example, an encyclopedia article about Steve Jobs may include information about corporations (e.g., Apple Computer, NeXT computer, Pixar, Pepsi-Cola, and Disney), places (e.g., Reno, Nev.), other people (e.g., Bill Gates, Dmitry Medvedev, Paul McCartney, and Steve Wozniak), events (e.g., personal computer revolution), and objects (e.g., iPhone, iPod, iPad, etc.).
This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.
The present technology is directed to analyzing content having facts to determine identities of people, places, objects, events, or other types of subjects with an associated date/time, season, period of time, and/or other chronological values. The facts in the content may then be associated with, organized, and/or stored in a network server, a client device, and/or other suitable storage locations, based on the identified subjects and corresponding chronology values. Upon receiving from a user a request for information about a subject, in certain embodiments, the server may query the stored facts to identify and retrieve facts corresponding to the requested subject and transmit the retrieved facts and also chronology values associated with the facts to the user. In other embodiments, the client device may query the stored facts with respect to the requested subject and present the retrieved facts with corresponding chronology values to the user.
Various embodiments of systems, devices, components, modules, routines, and processes for chronology based content processing are described below. In the following description, example software codes, values, and other specific details are included to provide a thorough understanding of various embodiments of the present technology. A person skilled in the relevant art will also understand that the technology may have additional embodiments. The technology may also be practiced without several of the details of the embodiments described below with reference to
As used herein, the term a “subject” generally refers to a person, place, object, event, or other types of topic for which a user may desire to obtain information. Also used herein, the term “content” generally refers to web pages, data feeds, electronic articles, electronic books, and/or other information containing facts. For example, web pages can include electronic encyclopedia pages (e.g., Wikipedia). Data feeds can include news feeds, information updates, emails, and/or other suitable types of mechanisms for users to receive updated data from data sources. Electronic articles or books can include publications in digital form produced on, published through, and/or readable on computers, tablets, smartphones, and/or other suitable computing devices.
As discussed above, a succinct list of facts about people, places, objects, events, or other types of subject matters may not be readily available to users. Several embodiments of the present technology are directed to automatically detecting facts associated with respect to people, places, objects, events, or other types of subjects and chronology values associated with the facts contained in a content. The facts may then be associated with, arranged, and stored based on the detected subjects and chronology values. As a result, users may readily obtain a time line of facts about a particular subject from the stored facts.
As shown in
In certain embodiments, the client device 110 can include a desktop, a laptop, a tablet, a smartphone, and/or other suitable types of computing device. As shown in
The server 102 can be configured analyze content 150 to extract a list of fact records 144 about a subject from the content 150. Upon receiving a request from the client device 110 or the optional client devices 110′, the server 102 can also be configured to transmit a list of facts about the requested subject to the client device 110 or the optional client devices 110′. Even though the server 102 is illustrated in
As shown in
The processor 104 can be configured to execute instructions of software components. For example, as shown in
In operation, the server 102 collects, receives, or otherwise obtains the content 150 with the input component 132. Based on the detection rules 142, the process component 136 analyzes the content 150 to identify subjects and corresponding chronology values in the content 150. The process component 136 then associates and arranges facts in the content 150 with the identified subjects and corresponding chronology values. The database component 134 may then facilitate storing the arranged facts as fact records 144 in the database 109. In certain embodiments, upon receiving a request for information about a subject, the database component 134 facilitates query and retrieval of a list of fact records 144 for the subject. In other embodiments, the request can also include the subject and a date. The database component 134 then facilitates query and retrieval of a list of fact records 144 based on both the subject and the date in the request. In further embodiments, an additional hardware and/or software component (e.g., an HDFS sub-server, not shown) may be configured to perform the foregoing operations in addition to or in lieu of the database component 134 of the server 102. The output component 138 then transmits the retrieved list of fact records 144 to the client device 110 or other suitable client devices. As such, a user of the client device 110 may readily obtain a succinct list of facts about the subject arranged in a chronological order.
The detection module 160 is configured to identify subjects and corresponding chronology values in the content 150. In the illustrated embodiment shown in
In certain embodiments, the time detector 161a can be configured to detect a date, time, and/or other suitable reference point in a calendar system based on the detection rules 142. Examples of the detection rules 142 can include the following:
The time detector 161a can also be configured to parse and/or otherwise transform the identified date. For instance, in the previous example, the time detector 161a can be configured to determine that the expression of “2013-02-13” includes a year of “2013,” a month of February, and a day of the 13th. Thus, the time detector 161a may transform the identified date into “Feb. 13, 2013,” “Feb. 13, 2013,” 2013.02.13,” and/or other suitable format.
In further embodiments, the time detector 161a can also be configured to recognize relative chronology terms in the content 150. As used herein, a “relative chronology term” general refers to a chronology value expressed in relation to a referenced point in a calendar system. Examples of relative chronology terms can include a number of seconds, minutes, hours, weeks, months, year, decade, or century, before or after a reference time point. In other examples, the relative chronology terms can also include ages, time lapses, and/or other suitable statements from which a date/time may be inferred. In yet other embodiments, the time detector 161a can be configured to detect other suitable chronological information via natural language processing, compound term processing, deep linguistic processing, semantic indexing, and/or other suitable techniques. Example operations of the time detector 161a are described in more detail later with reference to
The item detector 161b can be configured to detect subjects in the content 150. For example, the item detector 161b may be configured to detect names of people (e.g., Steve Jobs), places (e.g., Reno, Nev.), objects (e.g., iPad), or events (e.g., personal computer revolution). In another example, the item detector 161b may also be configured to detect a title (e.g., Dr.), nickname, abbreviation (e.g., JFK), modification (e.g., Sr., Jr., etc.), and/or other suitable identifiers of subjects. In certain embodiments, the item detector 161b may be configured to detect the subjects based on the detection rules 142. For example, names, nicknames, abbreviations, modification, and/or other identifiers may be stored as detection rules 142 in the database 109. In other embodiments, the item detector 161b may be configured to detect the subjects via natural language processing and/or other suitable techniques. Example operations of the item detector 161a are described in more detail later with reference to
The analysis module 162 may be configured to determine identities of the detected subjects based on chronological and/or other suitable context in the content 150. For example, the name “Steve Jobs” may be associated with multiple people. However, if the content 150 includes references to Apple Inc., the personal computer revolution, iPad, iPod, or computers in general, then the name “Steve Jobs” is likely referring to the co-founder and ex-CEO of Apple Inc. On the other hand, if the content 150 includes references to gold mining in the 1800's, then the name “Steve Jobs” is not likely referring to the co-founder and ex-CEO of Apple Inc. In another example, the name “Bill Gates” may be associated with the ex-CEO of Microsoft or the frontiersman and fortune hunter of the Klondike Gold Rush. If the content 150 includes references to Microsoft, computer software, or computer in general, the name “Bill Gates” is likely referring to the ex-CEO of Microsoft. If the content 150 includes references to gold, mining, gold mining, or Alaska, the name “Bill Gates” is probably referring to the frontiersman.
The calculation module 166 can include counters, timers, summers, subtractors, and/or other suitable accumulation routines configured to perform various types of calculations to facilitate operation of other modules. In one embodiment, the calculation module 166 may include a summer or subtractor that calculates a date value for a relative chronology term in the content 150. For example, if the content 150 includes an expression of “two days ago” from a reference date, then the expression “two days ago” may be converted into a date by subtracting two days from the reference date. In other embodiments, the calculation module 166 may include routines for performing time averaging, window averaging, filtering, and/or other suitable operations.
The control module 164 is configured to associate facts of the content 150 individually with the detected subjects with corresponding chronology value. In certain embodiments, the control module 164 may be configured to identify a chronology range for the content 150. For example, the control module 164 may determine a chronology range of Feb. 24, 1955 to Oct. 5, 2011, in the sentence “Steven Paul Jobs (Feb. 24, 1955-Oct. 5, 2011) was an American entrepreneur and inventor, best known as the co-founder, chairman, and CEO of Apple Inc.” In other embodiments, the control module 164 may not be configured to identify a chronology range because a range is not available and/or other suitable reasons.
The control module 164 may also be configured to divide the content 150 into chapters, paragraphs, sentences, fragments, and/or other suitable sections, and associate the individual sections with one or more detected subjects and corresponding chronology value based on at least one of grammar, syntax, and/or other suitable criteria. For instance, if a first section includes a detected subject (e.g., “Steve Jobs”) and a chronology value (e.g., “1955”), the control module 164 may associate the first section with the detected subject and chronology value. In one example, if a second section immediately following the first section does not contain any subject or chronology value, then the control module 164 may associate the second section to the subject (e.g., “Steve Jobs”) and chronology value (e.g., “1955”) of the first section. In another example, if the second section does not contain any subject but contains different chronology value (e.g., “1960”), the control module 164 may associate the second section with the subject (e.g., “Steve Jobs”) of the first section but with the new chronology value (i.e., “1960”). In a further example, if the second section contains a new subject (e.g., “Bill Gates”) but does not contain any new chronology value, the control module 164 may associate the second section with the new subject (i.e., “Bill Gates”) with the chronology value (i.e., “1955”) of the first section. In yet a further example, if the second section contains a first detected subject (e.g., “Steve Jobs”) and a second subject (e.g., “Bill Gates”), the control module 164 may infer that the second section is related to both the first and second subject with the chronology value of the first section.
The control module 164 may also be configured to assemble facts associated a particular subject (e.g., “Steve Jobs”) with corresponding chronology values. As used herein, “assembling facts” generally refers to creating and/or modifying a collection of facts regarding a particular subject. Assembling facts may include arranging facts, filtering facts, accepting user editing of certain facts, and/or other suitable operations. In one embodiment, the control module 164 is configured to assemble facts only associated with the particular subject. In other embodiments, the control module 164 is configured to assemble facts associated at least with the particular subject. In further embodiments, the control module 164 may be configured to assemble facts in other suitable manners. The facts from the searches may be arranged as records in chronological and/or other suitable orders. The control module 164 is also configured to establish dependencies, associations, or relationships of various subjects in the assembled facts.
In operation, the detection module 160 receives the content 150 and detects subjects with corresponding chronology value with the time detector 161a and the item detector 161b. During detection of the chronology value, the calculation module 166 may facilitate converting relative chronology terms into a date and/or other suitable reference point in a calendar system. The analysis module 162 then analyzes the content 150 to determine identities of the detected subjects based on chronological and/or other suitable context in the content 150. Based on the determined identities and the detected and/or converted chronology value, the control module 164 associates facts contained in portions of the content 150 with at least one identified subject and at least one chronology value. The control module 164 then assembles the facts based on the determined identities and corresponding chronology values.
As shown in
The process 200 then includes detecting one or more chronological values (e.g., with the time detector 161a in
In certain embodiments, detecting chronological values can include identifying chronological tokens and relative chronology terms. The term “chronological token” generally refers to numbers, words, phrases, or other representations of a date and/or time. Chronological tokens can individually include at least one of a year, month, day, time, and/or other suitable reference points in a calendar system. For example, in the sentence “Steven Paul Jobs (Feb. 24, 1955-Oct. 5, 2011) was an American entrepreneur and inventor, best known as the co-founder, chairman, and CEO of Apple Inc.” both Feb. 24, 1955 and Oct. 5, 2011 can be identified as chronological tokens. The term “relative chronology term” generally refers to numbers, words, phrases, or other representations of a time period relative to a reference date and/or time. For example, in the sentence “[the] Jobs family moved from San Francisco to Mountain View, Calif. when Steve was five years old,” the phrase “when Steve was five years old” can be a relative chronology term. The detected relative chronology terms may then be converted into chronological tokens based on context in the content 150. In the previous example, the process 200 can determine that the Jobs family moved to Mountain View, Calif. in 1960 because Steve was born in 1955.
In certain embodiments, detecting subjects can include detecting references to subjects and identifying the subjects associated with the references. The references to subjects can include names, nicknames, abbreviations, modifications, and/or other suitable designations. Such references may be detected by parsing, natural language processing, and/or other suitable techniques. The references can then be analyzed to determine the identities of the subjects the references are directed to. For example, the name “George Washington” can be directed to the first President of the United States, or an American jazz trombonist who had played in Louis Armstrong's orchestra.
In one embodiment, the identities of the subjects may be determined based on at least one keyword or phrase in the content 150. In the previous example, if the content 150 includes keywords such as “American Revolution,” “military career,” the name “George Washington” is probably directed to the first President of the United States. On the other hand, if the content 150 includes keywords such as “trombone,” “orchestra,” “ensembles,” the name “George Washington” is probably directed to the American jazz trombonist. In certain embodiments, all of the keywords may have the same weight for determining the identities of the subjects. In other embodiments, at least one of the keywords may have a different weight than other keywords.
In another embodiment, the identities of the subjects may be determined by comparing one or more corresponding chronology tokens to reference values or value ranges. For instance, in the previous example, if the content 150 includes chronology tokens such as 1907, 1932, and 1933, the name “George Washington” is probably not directed to the first President of the United States because the chronology tokens falls outside of his lifetime from 1732 to 1799. On the other hand, if the content 150 includes chronology tokens such as 1732, 1776, and 1795, the name “George Washington” is probably not directed to the American jazz trombonist.
In yet another embodiment, the identities of the subjects may be determined based on one or more references to other people in the content 150. For instance, in the previous example, if the content 150 includes references to “Thomas Jefferson,” “John Adams,” or “Friedrich von Steuben,” the name “George Washington” is probably directed to the first President of the United States. On the other hand, if the content 150 includes references to “Louis Armstrong,” “Red Allen,” or “Fletcher Henderson,” the name “George Washington” is probably directed to the American jazz trombonist. In further embodiments, the identities of the subjects may be determined based on a combination of at least some of the foregoing techniques with equal or different weights for different techniques.
The process 200 then includes assembling facts for individual subjects at stage 206. In certain embodiments, sentences, paragraphs, and/or other components of the content 150 may be associated with and sorted based on the detected subjects and chronological tokens, as discussed in more detail above with reference to
The process 200 can then optionally include storing the assembled facts for the individual subjects in a database (e.g., the database 109 of
Specific embodiments of the technology have been described above for purposes of illustration. However, various modifications may be made without deviating from the foregoing disclosure. In addition, many of the elements of one embodiment may be combined with other embodiments in addition to or in lieu of the elements of the other embodiments. Accordingly, the technology is not limited except as by the appended claims.
