KR101522049B1 - Coreference resolution in an ambiguity-sensitive natural language processing system - Google Patents

Abstract

Technologies for the same directive analysis in ambiguous sensitive natural language processing systems are described herein. Techniques for incorporating instructional analysis into a natural language processing system can process documents to be indexed within an information retrieval and retrieval system. Ambiguity recognition features as well as ambiguity analysis features can work in concert with the same directive analysis. Annotations of the same indication entities as well as ambiguous interpretations can be supported by inline markup in the text content or by external entity maps. Information expressed in documents can be formally organized by facts, that is, relationships between entities in the text. The extension may support applying multiple aliases, or ambiguities, to the entity being indexed so that all possible indications or interpretations for that entity are captured in the index. Alternative stored descriptions may support the retrieval of facts by the original description or by the same directional description.

Description

[0002] COREFERENCE RESOLUTION IN ANMBIGUITY-SENSITIVE NATURAL LANGUAGE PROCESSING SYSTEM [0003]

In natural language, it is not uncommon to direct entities by different descriptions. For example, pronouns are commonly used to replace nouns. In addition, various other descriptions of the reference, or different forms, may be used to indicate an entity. As an example, consider the following parts of the text:

"Pablo Picasso was born in Malaga."

"The Spanish painter became famous for his varied styles."

"Among his paintings are the large-scale Guernica."

"He painted this disturbing masterpiece during the Spanish Civil War."

"Picasso died in 1973."

I face various language changes. For example, two different names are used, "Pablo Picasso" and "Picasso". The definitive narrative "the Spanish painter" and the two pronouns "his" and "he" are all used to direct Picasso. Two different representations are used to indicate painting: "Guernica", the name of the work, and "demonstration description", "this disturbing masterpiece."

Two language expressions can be said to be identical if they have the same referent. In other words, they are supposed to point to the same entity. The second phrase may be an anaphoric anaphor in the first phrase. Therefore, the first phrase is the antecedent of the second phrase. Knowledge of the antecedents of the antecedent may be needed to determine the subject of the antecedent. A common task of finding the same directional expressions, forward converters, and their predecessors in a document can be called coreference resolution. The same directive analysis is a process of establishing that two expressions refer to the same directive, and does not necessarily establish what the directive is. Reference resolution is the process of establishing what the referent is.

For clusters of identically indicative expressions, regardless of their adaptive relationships, the expressions can be called aliases of each other. According to the example above, the expressions "Pablo Picasso", "the Spanish painter", "his", "he", and "Picasso" form an alias cluster pointing to Picasso.

Natural language expressions often represent ambiguity. Ambiguity occurs when an expression can be interpreted in two or more meanings. For example, the sentence "The duck is ready to eat" can be interpreted as asserting that the duck has been cooked properly or that the duck needs to hunger hungry.

The same directive analysis and ambiguity analysis are two examples of natural language processing operations that can be used to mechanically support the language commonly expressed by human users. Information processing systems such as text indexing and querying that support information retrieval can benefit from increased application of natural language processing systems.

It is to be understood that these speci fi cations and other considerations are provided herein.

Technologies for the same directive analysis in ambiguous sensitive natural language processing systems are described herein. In particular, techniques for incorporating the same directive analysis functionality into a system for processing documents to be indexed within an information retrieval and retrieval system are described. This integration can improve indexing by natural language documents, by the same directive analysis, and by information that supports ambiguous semantics.

According to one aspect set forth herein, the information provided by the same directive analysis system may be integrated into a natural language processing system to improve the performance of the natural language processing system. One example of such a system is a document indexing and retrieval system.

In accordance with another aspect set forth herein, ambiguity recognition features as well as ambiguity analysis features can operate in coordination with the same directive analysis in a natural language processing system. Annotations of the same instruction entities as well as ambiguous interpretations can be supported by in-line markup in textual representations, or alternatively by external entity maps.

According to another aspect set forth herein, facts can be extracted from the text to be indexed. Information expressed in text can be formally organized by facts. When used in this sense, the fact may be any information contained in the text, and is not necessarily true. The fact can be expressed as a relationship between entities. The fact may be stored in the semantic index as a relation between the entities stored in the semantic index. In a fact based search system, a document can be searched if it includes facts that match the fact that it is determined through analysis of the query.

According to another aspect set forth herein, a process of extension may support applying a plurality of aliases, or ambiguities, to an entity being indexed. Such an extension may support additional possible instructions, or interpretations, for a given entity being captured in the semantic index. Alternative stored descriptions may support the retrieval of facts by the original description or by the same directional description.

It is to be understood that the subject matter described may also be embodied as an article of manufacture, such as a computer-controlled device, a computer process, a computing system, or a computer-readable medium. These and various other features will be apparent from a reading of the following detailed description and a review of the relevant drawings.

This summary is provided to introduce a selection of the concepts further described in the following detailed description in a simplified form. This summary is not intended to be exhaustive or to limit the scope of the claimed subject matter to the essential features or essential features of the claimed subject matter. Furthermore, the claimed subject matter is not limited to implementations that solve any or all disadvantages pointed out in any part of this specification.

1 is a network architecture diagram illustrating an information retrieval system in accordance with aspects of the embodiments presented herein.
2 is a functional block diagram illustrating various components of a natural language index and query system in accordance with aspects of the embodiments presented herein.
Figure 3 is a functional block diagram illustrating the same directive analysis and ambiguity analysis in a natural language processing system in accordance with aspects of the embodiments presented herein.
4 is a logic flow diagram illustrating aspects of processes for ambiguous sensitive indexing by the same directive analysis in accordance with aspects of the embodiments presented herein.
5 is a computer architecture diagram illustrating exemplary computer hardware and software architectures for a computing system capable of implementing aspects of the embodiments presented herein.

The following detailed description relates to techniques for analyzing the same directive in a ambiguous sensitive natural language processing system. Through the use of the techniques and concepts presented herein, the same directive analysis function can be integrated into a natural language processing system that processes documents to be indexed for use in an information retrieval and retrieval system. This integration can improve the index by information that supports the same directive analysis for natural language documents being indexed.

While the teachings herein are presented in the general context of program modules executing in connection with the execution of an operating system and application programs on a computer system, those skilled in the art will appreciate that other implementations, along with other types of program modules, I will recognize. Generally, program modules include routines, programs, components, data structures, and other types of structures that perform particular tasks or implement particular abstract data types. It will also be appreciated by those skilled in the art that the teachings herein may be practiced with other computer system configurations, including hand held devices, microprocessor systems, microprocessor based or programmable consumer electronics, minicomputers, You will know.

In the following detailed description, reference is made to the accompanying drawings, which form a part hereof, and in which is shown by way of illustration specific embodiments or examples. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Reference will now be made to the drawings, in which like reference numerals represent like elements throughout the several views, to describe aspects of a computing system and method for the same directive analysis in a ambiguous sensitive natural language processing system.

Referring now to Figure 1, an exemplary operating environment for the implementations presented herein will be described in detail. In particular, network architecture diagram 100 illustrates an information retrieval system in accordance with aspects of the embodiments presented herein. The client computers 110A-110D may interface to the server 120 through the network 140 to obtain information related to the natural language engine 130. Although four client computers 110A-110D are illustrated, it should be understood that any number of client computers 110A-110D may be utilized. Client computers 110A-110D may be geographically dispersed across a network 140, placed in one location, or in any combination thereof. It should be appreciated that although only one server 120 is illustrated, the functionality of the server 120 may be distributed across any number of multiple servers 120. [ Such multiple servers 120 may be located in one location, geographically distributed across the network 140, or any combination thereof.

According to one or more embodiments, the natural language engine 130 may support a search engine function. In a search engine scenario, a user query may be issued from the client computers 110A-110D to the server 120 via the network 140. [ The user query can be a natural language format. At the server, the natural language engine 130 may process the natural language query to support the search based on the syntax and semantics extracted from the natural language query. The results of such a search may be provided from the server 120 to the client computers 110A-110D via the network 140. [

One or more search indices may be stored in, or related to, the server 120. The information in the search index may be populated from a set of source information, or from a corpus. For example, in a web search implementation, content can be collected and indexed from various web sites on various web servers (not shown) across the network 140. [ Such collection and indexing may be performed on the server 120 or by software running on another computer (not shown). The collection may be performed by web crawlers or spider applications. The natural language engine 130 may be applied to the collected information so that the natural language content collected from the corpus can be indexed based on the syntax and semantics extracted by the natural language engine 130. The indexing and retrieval is discussed in more detail with respect to FIG.

The client computers 110A-110D may function as a terminal client, a hypertext browser client, a graphical display client, or other networked clients to the server 120. [ For example, a web browser application on client computers 110A-110D may support interfacing with a web server application on server 120. [ Such browsers may use controls, plug-ins, or applets to support interfacing to the server 120. [ The client computers 110A-110D may also utilize other customized programs, applications, or modules to interface with the server 120. [ The client computers 110A-110D may be a desktop computer, laptop, handheld, mobile terminal, mobile phone, television set-top box, kiosk, server, terminal, thin- .

The network 140 may be any communication network capable of supporting communication between the client computers 110A-110D and the server 120. [ The network 140 may be wired, wireless, optical, radio, packet switched, circuit switched, or any combination thereof. The network 140 may utilize any topology and the links of the network 140 may be implemented in a variety of networks such as Ethernet, DSL, cable modem, ATM, SONET, MPLS, PSTN, POTS modem, PONS, HFC, satellite, ISDN, WiFi Protocols, or bandwidths, such as, for example, WiMAX, mobile cellular, any combination thereof, or any other data interconnection or networking mechanism. The network 140 may be any other network for interconnecting intranets, the Internet, the Internet, the World Wide Web, LAN, WAN, MAN, or computer systems.

In addition to the illustrated network environment, it should be appreciated that the natural language engine 130 may be operated locally. For example, server 120 and client computers 110A-110D may be combined into a single computing device. Such a combined system may support locally or remotely stored search indexes.

Referring now to FIG. 2, a functional block diagram illustrates various components of natural language engine 130 in accordance with one exemplary embodiment. As discussed above, the natural language engine 130 may support information searches. To support such searches, a content acquisition process 200 is performed. The operations related to the content acquisition 200 extract information from the documents provided as the text content 210. This information may be stored in a semantic index 250 that may be used for searching. Operations related to user search 205 may support processing of user input search queries. The user query can take the form of a natural language query (260). The natural language engine 130 may analyze the user input to convert the query into a representation to be compared with the information represented in the semantic index 250. The content and structure of the information in the semantic index 250 may support rapid matching and retrieval of documents, or portions of documents, that are related to the meaning of the query or natural language query 260.

The textual content 210 may include documents in a very general sense. Examples of such documents may include web pages, text documents, scanned documents, databases, information listings, other Internet content, or any other information source. This text content 210 may provide a corpus of information to be searched. Processing textual content 210 may take place in two stages as syntactic parsing 215 and semantic mapping 225. Preliminary language processing steps may occur before parsing 215 or at the beginning of parsing 215. For example, text content 210 may be separated at sentence boundaries. Appropriate nouns can be identified as names of specific people, places, objects or events. In addition, the grammatical properties of meaningful word ends can be determined. For example, in English, a noun ending in "s" may be plural nouns, while a verb ending in "s" may be a verb in the third person singular.

The parsing 215 is provided herein as a generic example only, but may be performed by a parsing system, such as the Xerox Linguistic Environment (XLE), rather than limiting the possible implementations of such descriptions. The parser 215 can convert sentences into expressions that clarify syntactic relationships between words. The parser 215 may apply the grammar 220 associated with the particular language being used. For example, the parser 215 may apply the grammar 220 for English. Grammar 220 may be used in conjunction with a lexical functional grammar (LFG) or a Head-Driven Phrase Structure Grammar (CCG), a Combinatory Categorial Grammar (CCG), a Probabilistic Context-free Grammar (PCFG) Lt; / RTI > can be formatted as other suitable parsing mechanisms, such as those based. Grammar 220 can specify possible ways to construct meaningful sentences in a given language. Parser 215 may apply rules of grammar 220 to strings of text content 210.

The grammar 220 may be provided for various languages. For example, LFG grammars were generated for English, French, German, Chinese, and Japanese. Other grammars may be provided. Grammar 220 can be developed by manual acquisition in which grammar rules are defined by a linguist or dictionary author. Alternatively, machine learning acquisition may involve automated observation and analysis of many examples of text from large corpus to automatically determine grammar rules. A combination of manual definition and machine learning may be used to obtain the rules of grammar 220.

The parser 215 may apply the grammar 220 to the text content 210 to determine the syntax structure. In the case of LFG-based parsing, syntactic structures consist of constituent structures (c-structures) and functional structures (f-structures). The c-structure may represent a hierarchy of component phrases and words. The f-structure can encode the roles and relationships between the various components of the c-structure. The f-structure can also represent information derived from the forms of words. For example, the plural of the noun or the tense of the verb can be specified in the f-structure.

During the semantic mapping process 225 following the parsing process 215, information from the syntax constructs may be extracted and combined with information about the semantics of the words in the sentence. A semantic map of the sentence or a semantic representation may be provided as content semantics 240. [ The semantic mapping 225 may augment the conceptual properties of individual words in the syntactic relations provided by the parser 215. The results may be converted into representations of the semantics of the sentences from the text content 210. Semantic mapping 225 can determine the roles performed by the words in the sentence. For example, a subject performing an action, something used to perform an action, or something affected by an action. For search indexing, they may be stored in the semantic index 250 along with their role in the words. Thus, the search from the semantic index 250 may depend only on the meaning within the sentences in which the word appears in the text content 210, rather than depending only on the separated word. Semantic mapping 225 can support disambiguation of terms, determination of antecedent relationships, and expansion of terms by synonyms, hypernyms, or hyponyms.

Semantic mapping 225 may apply knowledge resources 230 as techniques and rules for extracting semantics from sentences. The knowledge resources may be obtained through both manual definition and machine learning, as discussed with respect to the acquisition of grammars 220. Semantic mapping 225 processing may provide content semantics 240 in a semantic extensible markup language (semantic XML or semxml) representation. Any suitable representation language, such as expressions written in PROLOG, LISP, JSON, YAML, etc., may also be used. Content semantics 240 may specify the roles performed by the words in the sentences of text content 210. Content semantics 240 may be provided to the indexing process 245. [

The index may support representing a large corpus of information so that the positions of words and phrases can be quickly identified in the index. Traditional search engines can use keywords as search terms to map indexes to articles or documents where the keywords appear from the keywords specified by the user. The semantic index 250 may represent semantic meanings of words in addition to the words themselves. Semantic relationships can be assigned to words during content acquisition 200 and user search 205. [ Queries for the semantic index 250 may be based on words of a particular role, not just words. The roles are those performed by the word in the sentence or phrase stored in the semantic index 250. [ The semantic index 250 can be regarded as an inverted index whose entries are quickly searchable databases that are semantic words (i.e., words in a given role) and documents in which the words appear, or pointers to web pages. The semantic index 250 may support hybrid indexing. Such hybrid indexing can combine features and functions of both keyword indexing and semantic indexing.

User input of queries may be supported in the form of natural language questions 260. The query may be analyzed similar to that used in content acquisition 200, or through the same natural language pipeline. That is, the natural language query 260 can be processed by the parser 265 to extract the syntax structure. Following the syntax parsing 265, the natural language query 260 may be processed for semantic mapping 270. Semantic mapping 270 may provide query semantics 275 to be used in search process 280 for semantic index 250 as discussed above. The search process 280 may support hybrid index queries, in which both keyword index search and semantic index search may be provided alone or in combination.

In response to the user query, the results of the search 280 from the semantic index 250 along with the query semantics 275 may inform the ranking process 285. Ranking can use both keyword and semantic information. During ranking 285, the results obtained by search 280 may be ordered by various metrics in an attempt to place the most desirable results closer to the top of the retrieved information to be presented to the user as a result presentation 290. [ .

Referring now to FIG. 3, a functional block diagram illustrates the same directive analysis and ambiguity analysis in natural language processing system 300 in accordance with aspects of the embodiments presented herein. As an example of an application, natural language processing system 300 may support an information search engine for document indexing and searching. Such a natural language capable search engine can extend the information stored in its index based on language analysis. The system can also support the discovery of intent within the user query by analyzing the query verbally. The same directive analysis and ambiguity analysis features discussed herein may operate in conjunction with syntax parsing 215, semantic mapping 225, and semantic indexing 245, as discussed with respect to FIG. The same directive analysis may be performed directly on the text content 210, or may use information from parsing 215 or semantic mapping 225 operations.

As illustrated, the same directive analysis (320, 370) can be performed directly on the segmented document and as part of the semantic mapping (225). These two occurrences of the same directive analysis 320, 370 can be merged, or their information outputs can be merged. It should be appreciated that the same directive parsing may also occur between the syntax parsing 215 and the semantic mapping 225. The same directive analysis can also occur at any other stage in the natural language processing pipeline. There may be one, two, or more identical directive analysis components, or stages, at various locations within a natural language processing system. The text content 210 may be analyzed for information to be stored in the semantic index 250. The search may include querying the semantic index 250 for the desired information.

The content segment 310 may be performed on documents that make up the text content 210. The documents can be segmented for more efficient and potentially more accurate equivalent directive analysis 320. [ The same directive analysis 320 may consider potential reference relationships across the entire document. In the case of long documents, it can take a lot of time to compare distant expressions. When the speed of processing is considered, the content segmentation 310 of documents prior to the same directive analysis 320 may substantially reduce the time used for processing. Content segmentation 310 may effectively reduce the amount of content text 210 being explored in attempts of the same directive analysis 320. [

The content segmentation 310 may provide information to the semantic same directive analysis 370 indicating when new document segmentation is to begin. Such information may be provided as a segmentation signal 312 or by inserting a mark-up in the content document segment. External files or other mechanisms containing meta information may also be used.

The structure of the document may be used to identify segment boundaries where the indicative relationships are unlikely to intersect. The document structure can be inferred from explicit markup such as paragraph boundaries, chapters, or section headings. The document structure can also be found through language processing. Segments that exceed the specified length may be further subdivided. The desired subdivision length can be expressed, for example, by the number of sentences or the number of words.

If valid document structuring is not available, heuristic or statistical criteria can be applied. Such criteria may be specified to have a tendency to keep the same directives together while limiting the size of the segment to a predetermined maximum value. Various other approaches for segmenting textual content 210 documents may also be applied. Content segmentation 310 may also specify the entire document as one segment.

The same directive analysis (320, 370) can be used to identify the same directives and aliases in the content text 210. For example, when indexing the sentence "He painted Guernica ", it may be judicious to determine that" he " This is especially true if fact-based searches are being used. Analyzing the synonyms for Picasso can help index the fact that Picasso painted Guernica, rather than the less informative fact that "male" he "Guernica" was painting. Without this ability to identify and index the pronoun's referent, it may be difficult to retrieve the document in response to the query "Picasso painted" using the fact-based retrieval method. The recall of the system can be improved when documents related to queries that might otherwise not be returned are returned.

An annotation 330 may be applied to the textual content 210 to assist in tracking entities and possible same indications. The confidence values of the analysis decisions can also be annotated or marked up in the text content 210. The analysis decisions can be recorded by adding explicit comment marks to the text. For example, the text, "John visited Mary. He met her in 2003." The annotation 330 can be applied as "[E1: 0.9 John] visited [E2: 0.8 Mary]. [E1: 0.9 He] met [E2: 0.8 her] in 2003." Where the words "John" and "He" may be related as entity 1 (E1) with a confidence value of 0.9. Similarly, the words "Mary" and "her" may be related as entity 2 (E2) having a confidence value of 0.8. The confidence value may represent a measure of confidence in the same directive analysis 320 decision. The annotations can directly encode the same directive verdicts or the annotations can function as identifiers that link the relevant terms in the annotated text to additional information in the stand-by annotation 325. [

The same directive analysis 320 determinations can be used as part of the process of constructing the semantic mapping 225. The reference expressions used by the same directive analysis 320 may be incorporated into the input representation for the semantic mapping 225 by inline annotations in the text content 210. The instructions may also be provided separately in the outer stand-by-side entity map 325.

Within a large collection of documents of the text content 210, such as the World Wide Web, the same sentence can appear multiple times in different contexts. These different contexts may provide different candidates for the same directive analysis 320. Because the syntax parsing 215 may be computationally expensive, it may be advantageous to store the parsing results for the sentences in a cache. Such a caching mechanism 350 can quickly assist in retrieving parsing information in the future when a sentence is encountered.

If the same directive analysis 320 is applied to a single sentence appearing in different contexts, it may identify different identical referential relations for the same reference expressions since the same directive may be context-dependent. Thus, different entity identifiers may be inserted inline in the text. For example, the text "He is smart" appearing in two different documents may contain two different identifiers, "[E21 He] is smart." And "[E78 He] is smart." Here, the word "He" in the first document refers to a person other than the word "He" in the second document.

There may be different sources of information for the shallow equal directive analysis 320. For example, in addition to expression detection performed during the same directive analysis 320, there may be a system dedicated to finding the appropriate names in the textual content 210. These different sources can identify conflicting analysis information. For example, conflicting analyzes can occur where the boundaries intersect. For example, two systems may have identified the following conflicting reference expressions:

"[John] told [George Washington] [Irving] was a great writer."

"[John] told [George] [Washington Irving] was a great writer."

Consider the following conflicts of crossed boundaries: [George Washington] in the first string conflicts with [George] in the second string. Also, [George Washington] in the first string conflicts with [Washington Irving] in the second string. Based on trust information or contextual factors, different strategies can be applied repeatedly to analyze and preserve these conflicts. In a "drop" strategy, two or more conflicting boundaries can be resolved by dropping having the lowest confidence. In a "merge" strategy, where two or more boundaries are equally plausible in compatible contexts, the boundaries may be moved accordingly. For example, "[Mr. John] Smith" and "Mr. [John Smith]" can be merged to provide "[Mr. John Smith]". In a "preserve" strategy, multiple boundaries can be preserved by keeping the configuration of their boundaries and their trust values as ambiguous outputs if they do not support merging or dropping. For example, "[Alexander the Great]" and "[Alexander] [the Great]" can be provided as alternative ambiguous analyzes.

The parsing component 215 may be a ambiguity recognition parser that supports direct parsing of ambiguous input in which a syntactic parse 355 may preserve ambiguity. Alternatively, ambiguous input analyzes may need to be parsed separately, and multiple output structures may be communicated separately to the semantic component 225. The semantic processing 225 may be applied multiple times to each output of the syntax parser 215, as discussed in more detail below. Whereby different semantic outputs can be generated for different syntax inputs. Alternatively, semantic mapping 225 can combine the various inputs and process them all together.

The semantic mapping 225 may be present with semantic normalization 360. Multiple ambiguous phrase pars (355) outputs of a sentence can share meaning with different forms. For example, this can happen in the standardization of passive language. Considering "John gave Mary a present", the word "John" is the subject and "Mary" is the indirect object. Considering "a present was given to Mary by John", the subject is "Mary" and "John" is the object. Normalization 360 may provide outputs indicating that these two examples are the same as "John" is a semantic-subject and "Mary" is a semantic-indirect-object. Alternatively, "John" may be identified as a behavior agent and "Mary" may be identified as a recipient. Likewise, the same expressions can be provided for "Rome's destruction of Carthage" and "Rome destroyed Carthage".

Semantic normalization can also add information about different words in a parsed sentence. For example, words may be identified in a lexicon and associated with their synonyms, parentheticals, possible aliases, and other lexical information.

The semantic-based same directive analysis 370 can analyze expressions based on syntax and semantic information. For example, "John saw Bill. He greeted him." Can analyze "he" as "John" and "him" as "Bill". This analysis gives both "he" and "John", and "him" and "Bill" can both be specified because they are objects.

Shallow identical directive parser 320 can function by closely examining the document segment in which the terms appear. In contrast, semantic same directive analysis (370) or deep same-directive analysis can process one sentence at a time. Possible antecedents of the sentences may be placed in the antecedent store 375 so that semantic same directive analysis 370 of later sentences can access the earlier introduced elements. The antecedents can be stored along with their grammatical functions and roles in sentences, information about their distance in the text, information about their relationships with other antecedents, and various other information.

Expression merging 380 may combine the information from representations from the shallow in same directive analysis 320, the stand-by side annotation 325, and the semantic same directive analysis 370. Information about the terms to be combined may be identified using string alignment or annotations 330. [ Other mechanisms for combining the two annotations for the same text may be used.

Syntax parsing 215 may be a natural point of integration for optionally detected reference expressions. A parser can support inferring grammatical relationships such as constructs within sentences such as components, or subject and object. Ambiguity enable syntax parser 215 may identify a number of alternative structural representations of the sentence. In one example, the information from the same directive analysis 320 is used to filter the output of the syntax parser 215 by keeping only the expressions whose left boundaries of each reference expression match the beginning of the parts that are compatible from the pars . For example, the same directive analysis can establish the same directives as in "[E0 John] told [E1 George] [E2 Washington Irving] was a great writer." The syntax parser 215 may separately provide four parsing possibilities:

1. [John] and [George] and [Washington Irving]

2. [John] and [George] and [Washington] and [Irving]

3. [John] and [George Washington] and [Irving]

4. [John] and [George Washington Irving]

Parser possibilities Nos. 3 and 4 may be filtered because they are incompatible with the left border of the entity E2 "Washington Irving " provided by the instruction analysis 320.

The process of extension 385 may add additional information to the representation. For example, for "John sold a car from Bill", the extension 385 may additionally output a representation for "Bill bought a car from John". Similarly, for "John killed Bill ", extension 385 may further output a representation for" Bill died ".

Traditional search engines can search for documents in response to user queries based on matching keywords or terms. Documents can be ranked according to factors such as how many of the terms from the queries appear in the documents, how often they appear, or how closely the terms appear together, in these traditional systems have.

Consider the example query "Picasso painted" in a first example document that includes "Picasso's friend Matisse painted prolifically." With a second example document containing "Picasso was born in Malaga. He painted Guernica." If everything else is the same, the traditional system can rank the second document higher than the first document because the words "Picasso" and "painted" are closer together in the second document. . In contrast, a system that can analyze that the word "He" in the first document points to Picasso can accurately rank the first document higher on the basis of this knowledge. Assuming the query "Picasso painted" reflects the intent of the user to figure out what Picasso has drawn, the first document is obviously more relevant.

The natural language processing system 300 may have different architectures. In one embodiment, a pipeline may be provided in which information from one stage of language processing is passed as input to later stages. It is to be appreciated that these approaches may be implemented with any other architecture that can operate to extract facts to be indexed from the natural language text content 210.

Referring now to FIG. 4, additional details regarding the embodiments presented herein for the same directive analysis in a ambiguous sensitive natural language processing system will be provided. In particular, FIG. 4 is a flow chart illustrating aspects of processes 400 for ambiguous sensitive indexing by the same directive analysis in accordance with aspects of the embodiments presented herein.

It is to be understood that the logic operations described herein are implemented as (1) program modules executing in a computing system or as a sequence of computer-implemented acts and / or (2) as interconnected machine logic circuits or circuit modules in a computing system . This implementation is a matter of choice that depends on the performance and other requirements of the computing system. Accordingly, the logical operations described herein are variously referred to as state operations, structural devices, steps, or modules. These operations, structures, acts and modules may be implemented in software, in hardware, in special purpose digital logic, and any combination thereof. It should also be understood that more or fewer operations may be performed than those illustrated in the Figures and described herein. These operations may also be performed sequentially, in parallel, or in a different order than those described herein.

The routine 400 begins at operation 410, where a portion of the textual content 210 may be retrieved for analysis and indexing. At operation 420, the text content 210 is segmented for the boundaries of the regions of text that the analysis process searches and analyzes much. This segmentation may be based on a structure within the text, such as sentences, paragraphs, pages, chapters, or sections. Segmentation may also be based on the number of words, the number of sentences, or other metrics of space or complexity.

In operation 430, the same directives may be analyzed within the text content 210. In conjunction with the boundaries established within operation 430, the same directives can be identified and matched. Alias groups can be established. Surface structures can be used to provide "shallow" analysis. The ambiguities that occur during the same directive analysis can be annotated. Such an annotation 340 may be provided as markup in the text content 210 or through the use of an external entity map. A similar annotation may also be used to label the indications and the indication entities with entity numbers. An annotation may also be provided to indicate the confidence level of the established directive assays.

In operation 440, syntax parsing may convert sentences into expressions that clarify syntactic relationships between words. Parser 215 may apply grammar 220 associated with a particular language to provide syntax parsing 355 information.

In operation 450, semantic representations may be extracted from the text content 210. The information represented in the document in textual content 210 may be formally organized by representations of relationships between entities within the text. These relationships can be referred to as facts in a general sense.

At operation 455, the syntax pars 355 information output from syntax parser 215 may be used to support deep same-direct parser 370. The semantic representations generated during operation 450 may also be used.

In operation 460, representations from the shallow in same directive analysis operation 430 may be integrated with information from the deep same directive analysis operation 455. Ambiguity enable syntax parser 215 may identify a number of alternative structural representations of the sentence. The information from the same directive analysis can be used to filter the output of the syntax parser 215.

At operation 470, the semantics of textual content 210 may be expanded to include selected implicit representations. At operation 475, facts may be extracted from the semantic representations representing relationships between the states of the entities, events and events within the content text. In operation 480, facts and entities may be stored in the semantic index 250.

Routine 400 may terminate after operation 480. [ It should be understood, however, that the routine 400 can be applied repeatedly or sequentially to retrieve portions of the textual content 210 to be applied to the semantic index 250. [

Referring now to FIG. 5, an exemplary computer architecture 500 may execute the software components described herein for the same directive analysis in a ambiguous sensitive natural language processing system. The computer architecture shown in FIG. 5 illustrates a conventional desktop, laptop, or server computer and may be used to execute any aspects of the software components presented herein. It should be understood, however, that the described software components may also be implemented in other exemplary computing environments, such as mobile devices, televisions, set-top boxes, kiosks, vehicle information systems, mobile phones, embedded systems, Any one or more of client computers 110A-110D or server computers 120 may be implemented as computer system 500 in accordance with embodiments.

The computer architecture illustrated in Figure 5 includes a system memory 13 that includes a central processing unit 10 (CPU), a random access memory 14 (RAM) and a read only memory 16 (ROM) 13) to the CPU 10 via the system bus 11. [ A basic input / output system may be stored in the ROM 16, including basic routines that help to transfer information between elements within the computer 500, such as when running. The computer 500 may further include a mass storage device 15 for storing various program modules, such as those associated with an operating system 18, software, data, and natural language engine 130. The natural language engine 130 may execute portions of the software components described herein. The semantic index 250 associated with the natural language engine 130 may be stored in the mass storage device 15.

The mass storage device 15 may be connected to the CPU 10 via a mass storage controller (not shown) connected to the bus 11. [ The mass storage device 15 and its associated computer-readable media may provide non-volatile storage for the computer 500. Although the description of computer-readable media contained herein refers to a mass storage device such as a hard disk or a CD-ROM drive, those skilled in the art will appreciate that any use of computer readable media, Lt; RTI ID = 0.0 > computer storage media. ≪ / RTI >

By way of example, and not limitation, computer readable media can comprise volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data Media. For example, the computer-readable medium may be RAM, ROM, EPROM, EEPROM, flash memory or other solid state memory technology, CD-ROM, digital versatile disk (DVD), HD-DVD, BLU- But are not limited to, a magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage device, or any other medium that can be used to store the desired information and which can be accessed by the computer 500 no.

According to various embodiments, the computer 500 may operate in a networked environment using logical connections to remote computers via a network, such as the network 140. The computer 500 may be connected to the network 140 via a network interface unit 19 connected to the bus 11. It should be appreciated that the network interface unit 19 may also be used to connect to other types of networks and remote computer systems. The computer 500 may also include an input / output controller 12 for receiving and processing inputs from a number of other devices, including a keyboard, a mouse, or an electronic stylus (not shown). Similarly, the input / output controller 12 may provide output to a video display, printer, or other type of output device (also not shown).

As mentioned briefly above, a number of program modules and data files may be stored on the computer 500, including an operating system 18 suitable for controlling the operation of networked desktops, laptops, server computers, or other computing environments. The mass storage device 15 and the RAM 14. [ The mass storage device 15, ROM 16, and RAM 14 may also store one or more program modules. In particular, mass storage device 15, ROM 16, and RAM 14 may store a natural language engine 130 for execution by CPU 10. The natural language engine 130 may comprise software components for implementing portions of the processes discussed in detail with respect to Figures 2-4. The mass storage device 15, ROM 16, and RAM 14 may also store other types of program modules. The mass storage device 15, the ROM 16, and the RAM 14 may also store the semantic index 250 associated with the natural language engine 130.

Based on the foregoing, it should be appreciated that techniques have been provided for the same directive analysis in ambiguous sensitive natural language processing systems. Although the teachings herein have been described in language specific to computer architecture features, methodological acts, and computer-readable media, it is to be understood that the invention defined in the appended claims is not necessarily limited to the specific features, acts, Or < / RTI > media. Rather, the specific features, steps, and media are disclosed as exemplary forms of implementing the claims.

The foregoing is provided by way of illustration only and should not be construed as limiting. Various modifications and changes may be made thereto without departing from the exemplary embodiments and applications illustrated and described in the following claims and without departing from the true spirit and scope of the present invention as set forth in the following claims.

Claims (20)

CLAIMS What is claimed is: 1. A method for integrating coreference resolution mechanisms,
Retrieving a portion of text using a natural language engine of the server;
Using the natural language engine of the server to identify the same core within the portion of the text;
Extracting a fact from a portion of the text using the natural language engine of the server, the fact having meaning; And
Using the natural language engine of the server to identify an ambiguity within a portion of the text;
Expanding the fact to an expanded fact using the natural language engine of the server, wherein the expanded fact has the same directive meaning that is different from the meaning and based on the identified same directive, Includes ambiguous meaning -
≪ / RTI > 2. The method of claim 1, wherein identifying the same directive within a portion of the text includes identifying the same directive within a portion of the text using at least some of syntactic parsing. The method of claim 1, wherein identifying the same directive within a portion of the text includes identifying the same directive within a portion of the text using at least some semantic mapping. 2. The method of claim 1, wherein identifying the same directive comprises identifying the same directive having ambiguity. delete delete 2. The method of claim 1, further comprising using the natural language engine of the server to store the expanded facts in an index operable to support information retrieval. 8. The method of claim 7 further comprising using the natural language engine of the server to retrieve the expanded fact from the index in response to a search query. 2. The method of claim 1, further comprising annotating the same directives identified in the portion of the text using the natural language engine of the server. 3. The method of claim 2, further comprising: caching information from the syntax parsing using the natural language engine of the server. A computer storage medium having computer executable instructions stored thereon, the instructions, when executed by a computer,
Search for parts of text,
Identifying the same directive within a portion of the text,
Extracting facts from a portion of the text, the facts having meaning,
Identifying ambiguities within the portion of the text,
And having the same directive meaning different from said fact based on said identified same directive, to expand said fact to have ambiguous meaning based on said identified ambiguity. 12. The computer storage medium of claim 11, wherein identifying the same directive comprises identifying the same directive within a portion of the text using at least some of the syntax parsing. 12. The computer storage medium of claim 11, wherein identifying the same directive comprises identifying the same directive within a portion of the text using at least some semantic mapping. 12. The computer storage medium of claim 11, wherein identifying the same directive comprises identifying the same directive having ambiguity. delete delete 12. The computer storage medium of claim 11, further comprising instructions for causing the computer to store the expanded facts in an index operable to support information retrieval. 18. The computer storage medium of claim 17, further comprising instructions for causing the computer to retrieve the expanded fact from the index in response to a search query. 12. The computer storage medium of claim 11, further comprising instructions for causing the computer to annotate the same directives identified within the portion of the text. As a method for integrating the same directive analysis mechanisms,
Using a natural language engine of the server computer to search for a portion of text;
Identifying the same directive within a portion of the text using the natural language engine of the server computer;
Using the natural language engine of the server computer to identify ambiguities within a portion of the text;
Extracting facts from a portion of the text using the natural language engine of the server computer, the facts having significance;
Expanding the fact using the natural language engine of the server computer, wherein the expanded fact has the same directive meaning that is different from the meaning and based on the identified same directive, and an ambiguous meaning based on the identified ambiguity (including ambiguous meaning);
Storing the extended facts in an index operable to support information retrieval; And
Retrieving the expanded fact from the index in response to a search query
≪ / RTI >

Images