Unified semantics-focused language processing and zero base knowledge building system

10496,749

14/825171

August 13, 2015

A method and a language processing and knowledge building system (LPKBS) for processing textual data, receives textual data and a language object; segments the textual data into sentences and each sentence into words; generates a list of one or more natural language phrase objects (NLPOs) for each word by identifying vocabulary classes and vocabulary class features for each word based on vocabulary class feature differentiators; creates sentence phrase lists, each including a combination of one NLPO selected per word from each list of NLPOs; groups two or more NLPOs in each sentence phrase list based on word to word association rules, the vocabulary classes, the vocabulary class features, and a position of each NLPO; replaces each such group of NLPOs with a consolidated NLPO; maps each segmented sentence to a sentence type; identifies a semantic item for each mapped NLPO; and identifies and stores associated attributes and relations.

Krishnamurthy, Satyanarayana (Bangalore, IN)

1. A method for processing textual data, said method employing a language processing and knowledge building system comprising at least one processor configured to execute computer program instructions for performing said method, said method comprising: receiving by said language processing and knowledge building system at least two portions of said textual data and, for each of the two portions of said textual data, a different corresponding language object each said language object corresponding to a text language of said corresponding portion of textual data and comprising an alphabet, a vocabulary, a vocabulary class for each word, punctuation, sentence types, sentence part types, word association rules, semantic item creation rules and semantic consequence rules for the text language; for each said portion of textual data, without user intervention: a) segmenting said received portion of textual data into one or more sentences by said language processing and knowledge building system based on a plurality of sentence terminators predefined in said language object; b) segmenting each of said one or more sentences into a plurality of words by said language processing and knowledge building system based on a plurality of word separators predefined in said language object; c) generating a list of one or more natural language phrase objects for each of said words by said language processing and knowledge building system by identifying vocabulary classes and vocabulary class features for said each of said words based on vocabulary class feature differentiators predefined in said language object, said vocabulary class comprising a multivalued mapping from the vocabulary to a vocabulary class set of a language model predefined by the language object; d) creating one or more sentence phrase lists by said language processing and knowledge building system using each said generated list of one or more natural language phrase objects, wherein each of said created one or more sentence phrase lists comprises a combination of one natural language phrase object selected for said each of said words from said each said generated list of one or more natural language phrase objects; e) grouping two or more natural language phrase objects in said each of said created one or more sentence phrase lists by said language processing and knowledge building system based on word to word association rules predefined in said language object, said identified vocabulary classes, said identified vocabulary class features, and a position of each natural language phrase object in said each of said created one or more sentence phrase lists, and replacing each said grouped two or more natural language phrase objects in said each of said created one or more sentence phrase lists with a consolidated natural language phrase object; f) mapping said segmented each of said one or more sentences to a sentence type by: mapping each natural language phrase object present in said each of said created one or more sentence phrase lists at a current point in said processing of said received portion of textual data to a sentence part type in a sentence type selected iteratively from a plurality of sentence types predefined in said language object by said language processing and knowledge building system, based on word to sentence part type association rules predefined in said language object, using said identified vocabulary classes, said identified vocabulary class features, and said position of said each natural language phrase object in said each of said created one or more sentence phrase lists at said current point in said processing of said received portion of textual data, wherein said each natural language phrase object at said current point in said processing of said received portion of textual data is one of: one from said generated list of one or more natural language phrase objects and said consolidated natural language phrase object; and identifying said sentence type of said segmented each of said one or more sentences by said language processing and knowledge building system from a sentence type with a highest number of successfully mapped sentence part types; g) verifying, by said language processing and knowledge building system, for said mapped each natural language phrase object in said each of said created one or more sentence phrase lists mapped successfully to said identified sentence type, that a root word of said mapped each natural language phrase object in said each of said created one or more sentence phrase lists mapped successfully to said identified sentence type was not encountered earlier in a discourse and that a discourse context does not already have an earlier created matching semantic item whose expressions attribute contains the root word; h) creating, by said language processing and knowledge building system, for said mapped each natural language phrase object in said each of said created one or more sentence phrase lists mapped successfully to said identified sentence type, one or more semantic items corresponding to the root word of said mapped each natural language phrase object in said each of said created one or more sentence phrase lists mapped successfully to said identified sentence type and adding the root word to the expressions attribute of the created semantic item; i) creating attributes of one or more of created semantic items, and further creating relations between said created semantic items, based on said identified sentence type and semantic consequence rules of said identified sentence type, predefined in said language object, and adding said created attributes and created relations to said one of said created semantic items and further storing said created semantic items, attributes and said created relations in said discourse context and optionally in a system knowledge by said language processing and knowledge building system; and j) transforming said portion of textual data into data structures and making the data structures available for use by software applications, said data structures comprising for each said sentence a corresponding said sentence type and said sentence phrase lists associated with said sentence, each comprising an array of one of said natural language phrase objects associated with each of said words of said sentence, said natural language phrase objects in said array thereof each comprising an array of said associated word and grammatical attributes of said associated word; whereby textual data is processed for ease of analysis and use in computer application software; said method further comprising identifying the sentence part type for each of unmapped natural language phrase objects by said language processing and knowledge building system based on a semantic compounding mechanism predefined in said language object, wherein said semantic compounding mechanism comprises rules for associating sentence part types with natural language phrase objects based on one or more of a plurality of semantic compounding types, wherein said semantic compounding types comprise a concurrency semantic compounding type, a cause-effect relationship semantic compounding type, a condition semantic compounding type, and a shared object semantic compounding type between two sentences, wherein semantic compounding is combining multiple propositions into a single sentence to achieve overall economy of expression, wherein the concurrency semantic compounding type specifies that the multiple propositions describe actions or events that are happening or happened or can happen at the same time, wherein the cause-effect relationship semantic compounding type specifies that one or some of the multiple propositions describe or depict things that serve as the cause for things that are described or depicted by the other or others among the multiple propositions, wherein the condition semantic compounding type specifies that one or some of the multiple propositions describe or depict conditions that need to be satisfied, or state that needs to be realized, or actions or events that should occur, before what is described or depicted as a promise or a possibility by the other or others among the multiple propositions can actually become a reality, wherein the shared object semantic compounding type specifies that an entity or some entities participating in one or more of the multiple propositions also participates or participate in another or many others among the multiple propositions.

2. The method of claim 1 , wherein said discourse context is configured as a mere template to store data at multiple meta-levels, wherein said data can be semantic items, and also to store additional information such as the last occurring question, a dereferencing-aid-map which is a map with a defined, fixed set of keys, wherein each key is a specific combination of vocabulary class feature values, and corresponding values would be initially null and would be repeatedly updated to hold the last occurring phrase objects with the vocabulary class feature values.

3. The method of claim 1 , wherein said system knowledge is configured as a mere template to store data at multiple meta-levels, wherein said data can be semantic items.

4. The method of claim 1 , wherein said semantic item is one of a semantic classification, a relation between two semantic classifications, an action category, a specific object, a link between two specific objects as an instance of the relation between the semantic classifications to which the specific objects belong, a symbolic object, and an attribute of a semantic classification.

5. The method of claim 1 , further comprising creating semantic items, by said language processing and knowledge building system, and storing said created semantic items, in said discourse context and optionally in said system knowledge, wherein said creation of said semantic items is based on predefined semantic item creation rules.

6. The method of claim 5 , further comprising recursive application of rules for creation of semantic items, in respect of consolidated language phrase objects, on each language phrase object of the consolidated language phrase objects.

7. The method of claim 1 , further comprising determining whether to create one of a semantic classification, a class attribute, and a symbolic object for a mapped natural language phrase object with a common noun root word by said language processing and knowledge building system using heuristic rules on context and by processing lexical meaning definition text data for said common noun root word obtained from a vocabulary object predefined in said language object.

8. The method of claim 1 , wherein said language object is predefined for a natural language used in said received textual data.

9. The method of claim 1 , wherein said vocabulary classes comprise parts of speech comprising a noun, a pronoun, a verb, an adjective, and an adverb.

10. The method of claim 1 , wherein said vocabulary class features comprise a case, a gender, a number, and a tense of said each of said words.

11. The method of claim 1 , further comprising identifying or deriving one or more root words and semantic variations for said each of said words by said language processing and knowledge building system based on connector word rules and morphed word rules predefined in said language object for vocabulary class feature differentiation and sense differentiation.

12. The method of claim 11 , further comprising validating said identified or derived vocabulary classes and each of said identified or derived one or more root words for said each of said words by said language processing and knowledge building system by querying the vocabulary object predefined in said language object.

13. The method of claim 1 , wherein each of said one or more natural language phrase objects comprises said identified vocabulary class, said identified vocabulary class features, and the root word of said each of said words, wherein the root word is derived by either reverse application of morphological rules defined in the language object or is identified by lookup in a table defined in the language object.

14. The method of claim 1 , wherein said word to word association rules comprise rules for associating noun collocations, a word of article vocabulary class with a word of noun vocabulary class, a word of an adjective vocabulary class with a word of a noun vocabulary class, a word of an adverb vocabulary class with a word of a verb vocabulary class, and a word of an adverb vocabulary class with a word of an adjective vocabulary class.

15. The method of claim 1 , further comprising grouping one of said grouped two or more natural language phrase objects and ungrouped natural language phrase objects in said each of said created one or more sentence phrase lists based on said identified vocabulary classes, said identified vocabulary class features, list item separators, and list terminators predefined in said language object.

16. The method claim 2 , further comprising, for each language phrase object in the sentence phrase list, a value, for the key that matches the vocabulary class features of the language phrase object, in the dereferencing-aid-map in said discourse context with said language phrase object to facilitate dereferencing of a word of a pronoun vocabulary class in subsequent sentences.

17. The method of claim 1 , wherein said sentence part type comprises an action denoter, a doer denoter, a class denoter, a behaviour denoter, and a part denoter.

18. The method of claim 1 , wherein said word to sentence part type association rules comprise syntactic feature-based non-semantic rules specifying a mapping of a natural language phrase object to the sentence part type in the sentence type based on vocabulary classes, vocabulary class features in said natural language phrase object, and a position of said natural language phrase object in said sentence phrase list, independent of the specific meanings of the word(s) making up the language phrase object.

19. The method of claim 1 , further comprising identifying, for said each of said one or more sentences, preliminarily, a sentence type based on sentence terminators and question terminators predefined in said language object by language processing and knowledge building system, wherein said preliminary sentence type comprises one of a proposition and a question and finally, a refined sentence type (or proposition type) or a refined query type, based on the matching of the language phrase objects in the sentence phrase list to the sentence part types defined for the candidate refined sentence type or query type, wherein said refined sentence type or proposition type comprise of ObjectClassDescriptor, ClassStructureDescriptor, ClassBehaviourDescriptor, ObjectStructureDescriptor, ActionDescriptor and wherein said refined query type comprise of ClassStructureQuery, ClassBehaviourQuery, ObjectStructureQuery, ObjecyBehaviourQuery, ActionQuery, wherein the refined sentence types and refined query types constitute a communication ontology valid for the language and are independent of the domain-specific semantics or contextual semantics or meanings of the individual words in the sentence or query.

20. The method of claim 19 , further comprising updating a last question attribute in said discourse context by said language processing and knowledge building system when said identified sentence type is said question, to facilitate processing of a possible answer in a subsequent sentence even if said subsequent sentence omits or lacks words for one or more sentence part types normally expected for sentence types.

21. The method of claim 1 , further comprising identifying natural language phrase objects for each unmapped sentence part type by said language processing and knowledge building system using a value for the last question attribute in the discourse context and valid-in-interim association rules.

22. The method of claim 1 , wherein the textual data may be input by way of any of: i) text meant to be read and understood by the system for building up the knowledge to be used in other computer applications such as question-answering-systems, ii) text meant to be read and understood by the system for building an abstractive summary of the text in a summarization-system, iii) user's question, meant to be answered by the system, in a question-answering-system, or iv) user's observations, questions, requests, commands or comments in the course of a conversation in computer applications such as conversation-systems.

23. The method of claim 1 , wherein the textual data may be a single word or a short phrase or a full sentence or a question or a paragraph consisting of multiple sentences and/or questions or a chunk of text consisting of multiple paragraphs or may be the content of an entire text file.

24. The method of claim 1 , wherein the textual data is received in one shot and constitutes the entire first input, apart from the language object, for one complete operation of the method and system, on which the system is able to complete and completes all the steps of the method without any further interaction with the user or intervention from the user or any further input from the user.

25. The method of claim 1 , wherein the textual data may be any of: A) textual data interactively input by a user of the system using a computer system's keyboard, B) transcript or transformed form of what is spoken by a user of the system using the computer system's voice input device, C) transcript, howsoever prepared, of the conversation between two or more humans, D) textual data present in a text file on non-transitory storage device of the computer system, or E) textual data received on a communication link to which the computer system is connected.

26. A language processing and knowledge building system for processing textual data, said language processing and knowledge building system comprising: a non-transitory computer readable storage medium configured to store computer program instructions defined by modules of said language processing and knowledge building system; at least one processor communicatively coupled to said non-transitory computer readable storage medium, said at least one processor configured to execute said defined computer program instructions; and said modules of said language processing and knowledge building system comprising: a data reception module configured to receive at least two portions of said textual data and, for each of the two portions of said textual data, a different corresponding language object, each said language object corresponding to a text language of said corresponding portion of textual data and comprising an alphabet, a vocabulary, a vocabulary class for each word, punctuation, sentence types, sentence part types, word association rules for the language of the text, semantic item creation rules and semantic consequence rules for the text language; a data segmentation module configured to segment each said received portion of textual data into one or more sentences based on a plurality of sentence terminators predefined in said language object; said data segmentation module further configured to segment each of said one or more sentences into a plurality of words based on a plurality of word separators predefined in said language object; a phrase object processing module configured to generate a list of one or more natural language phrase objects for each of said words by identifying vocabulary classes and vocabulary class features for said each of said words based on vocabulary class feature differentiators predefined in said language object, said vocabulary class comprising a multivalued mapping from the vocabulary to a vocabulary class set of a language model predefined by the language object; said phrase object processing module further configured to create one or more sentence phrase lists using each said generated list of one or more natural language phrase objects, wherein each of said created one or more sentence phrase lists comprises a combination of one natural language phrase object selected for said each of said words from said each said generated list of one or more natural language phrase objects; said phrase object processing module further configured to group two or more natural language phrase objects in said each of said created one or more sentence phrase lists based on word to word association rules predefined in said language object, said identified vocabulary classes, said identified vocabulary class features, and a position of each natural language phrase object in said each of said created one or more sentence phrase lists, and to replace each said grouped two or more natural language phrase objects in said each of said created one or more sentence phrase lists with a consolidated natural language phrase object, to reduce the sentence phrase lists; a mapping module configured to map said segmented each of said one or more sentences to a sentence type by: mapping each natural language phrase object present in said each of said created one or more sentence phrase lists at a current point in said processing of each said received portion of textual data to a sentence part type in a sentence type selected iteratively from a plurality of sentence types predefined in said language object, based on word to sentence part type association rules predefined in said language object, using said identified vocabulary classes, said identified vocabulary class features, and said position of said each natural language phrase object in said each of said created one or more sentence phrase lists at said current point in said processing of each said received portion of textual data, wherein said each natural language phrase object at said current point in said processing of each said received portion of textual data is one of: one from said generated list of one or more natural language phrase objects and said consolidated natural language phrase objects; and identifying said sentence type of said segmented each of said one or more sentences from a sentence type with a highest number of successfully mapped sentence part types; and a semantic item creation module configured to verify, for said mapped each natural language phrase object in said each of said created one or more sentence phrase lists mapped successfully to said identified sentence type, that a root word of said mapped each natural language phrase object in said each of said created one or more sentence phrase lists mapped successfully to said identified sentence type was not encountered earlier in a discourse and that a discourse context does not already have an earlier created matching semantic item whose expressions attribute contains the root word and to create, for said mapped each natural language phrase object in said each of said created one or more sentence phrase lists mapped successfully to said identified sentence type, a semantic item corresponding to a root word of said mapped each natural language phrase object in said each of said created one or more sentence phrase lists mapped successfully to said identified sentence type and to add the root word to the expressions attribute of the created semantic item; said semantic item creation module further configured to create attributes of one or more of created semantic items, and further to create relations between said created semantic items, based on said identified sentence type and semantic consequence rules of said identified sentence type, predefined in said language object, and to add said created attributes and created relations to said created semantic items and further to store said created semantic items, created attributes and said created relations in said discourse context; whereby said language processing and knowledge building system is configured to process textual data and transform said textual data without user intervention into data structures and make the data structures available for use by software applications, said data structures comprising for each said sentence a corresponding said sentence type and said sentence phrase lists associated with said sentence, each comprising an array of one of said natural language phrase objects associated with each of said words of said sentence, said natural language phrase objects in said array thereof each comprising an array of said associated word and grammatical attributes of said associated word; wherein said phrase object processing module is further configured to identify the sentence part type for each of unmapped natural language phrase objects based on a semantic compounding mechanism predefined in said language object, wherein said semantic compounding mechanism comprises rules for associating sentence part types with natural language phrase objects based on one or more of a plurality of semantic compounding types, wherein said semantic compounding types comprise a concurrency semantic compounding type, a cause-effect relationship semantic compounding type, a condition semantic compounding type, and a shared object semantic compounding type between two sentences, wherein semantic compounding is combining multiple propositions into a single sentence to achieve overall economy of expression, wherein the concurrency semantic compounding type specifies that the multiple propositions describe actions or events that are happening or happened or can happen at the same time, wherein the cause-effect relationship semantic compounding type specifies that one or some of the multiple propositions describe or depict things that serve as the cause for things that are described or depicted by the other or others among the multiple propositions, wherein the condition semantic compounding type specifies that one or some of the multiple propositions describe or depict conditions that need to be satisfied, or state that needs to be realized, or actions or events that should occur, before what is described or depicted as a promise or a possibility by the other or others among the multiple propositions can actually become a reality, wherein the shared object semantic compounding type specifies that an entity or some entities participating in one or more of the multiple propositions also participates or participate in another or many others among the multiple propositions.

27. The language processing and knowledge building system of claim 26 , wherein said discourse context is configured as mere template to store data at multiple meta-levels, wherein said data can be semantic items, and also to store additional information such as the last occurring question, a dereferencing-aid-map which is a map with a defined, fixed set of keys, wherein each key is a specific combination of vocabulary class feature values, and values would be initially null and would be updated repeatedly to hold the last occurring phrase objects with the vocabulary class feature values.

28. The language processing and knowledge building system of claim 26 , wherein said system knowledge is configured as mere template to store data at multiple meta-levels, wherein said data can be semantic items.

29. The language processing and knowledge building system of claim 26 , wherein said semantic item is one of a semantic classification, a relation between two semantic classifications, an action category, a specific object, a link between two specific objects as an instance of the relation between the semantic classifications to which the specific objects belong, a symbolic object, and an attribute of a semantic classification.

30. The language processing and knowledge building system of claim 26 , wherein said semantic item creation module is further configured to create semantic items, and store said created semantic items, in said discourse context and optionally in said system knowledge, wherein said creation of said semantic items is based on predefined semantic item creation rules.

31. The language processing and knowledge building system of claim 30 is further configured for recursive application of rules for creation of semantic items, in respect of consolidated language phrase objects, on each language phrase object of the consolidated language phrase objects.

32. The language processing and knowledge building system of claim 26 , wherein said semantic item processing module is further configured to determine whether to create one of a semantic classification, a class attribute, and a symbolic object for a mapped natural language phrase object with a common noun root word using heuristic rules on context and by processing lexical meaning definition text data for said common noun root word obtained from a vocabulary object predefined in said language object.

33. The language processing and knowledge building system of claim 26 , wherein said language object is predefined for a natural language used in said received textual data.

34. The language processing and knowledge building system of claim 26 , wherein said vocabulary classes comprise parts of speech comprising a noun, a pronoun, a verb, an adjective, and an adverb.

35. The language processing and knowledge building system of claim 26 , wherein said vocabulary class features comprise a case, a gender, a number, and a tense of said each of said words.

36. The language processing and knowledge building system of claim 26 , wherein said phrase object processing module is further configured to identify or derive one or more root words and semantic variations for said each of said words based on connector word rules and morphed word rules predefined in said language object for vocabulary class feature differentiation and sense differentiation.

37. The language processing and knowledge building system of claim 36 , wherein said phrase object processing module is further configured to validate said identified or derived vocabulary classes and each of said identified or derived one or more root words for said each of said words by querying the vocabulary object predefined in said language object.

38. The language processing and knowledge building system of claim 26 , wherein each of said one or more natural language phrase objects comprises said identified vocabulary class, said identified vocabulary class features, and the root word of said each of said words, wherein the root word is derived by either reverse application of morphological rules defined in the language object or is identified by lookup in a table defined in the language object.

39. The language processing and knowledge building system of claim 26 , wherein said word to word association rules comprise rules for associating noun collocations, a word of article vocabulary class with a word of noun vocabulary class, a word of an adjective vocabulary class with a word of a noun vocabulary class, a word of an adverb vocabulary class with a word of a verb vocabulary class, and a word of an adverb vocabulary class with a word of an adjective vocabulary class.

40. The language processing and knowledge building system of claim 26 , wherein said phrase object processing module is further configured to group one of said grouped two or more natural language phrase objects and ungrouped natural language phrase objects in said each of said created one or more sentence phrase lists based on said identified vocabulary classes, said identified vocabulary class features, list item separators, and list terminators predefined in said language object.

41. The language processing and knowledge building system of claim 27 , wherein said phrase object processing module is further configured to update, for each language phrase object in the sentence phrase list, a value, for the key that matches the vocabulary class features of the language phrase object, in the dereferencing-aid-map in said discourse context with said language phrase object to facilitate dereferencing of a word of a pronoun vocabulary class in subsequent sentences.

42. The language processing and knowledge building system of claim 26 , wherein said sentence part type comprises an action denoter, a doer denoter, a class denoter, a behaviour denoter, and a part denoter.

43. The language processing and knowledge building system of claim 26 , wherein said word to sentence part type association rules comprise syntactic, feature-based non-semantic rules specifying a mapping of a natural language phrase object to the sentence part type in the sentence type based on said vocabulary classes, said vocabulary class features in said natural language phrase object, and a position of said natural language phrase object in said sentence phrase lists, independent of the specific meanings of the word(s) making up the language phrase object.

44. The language processing and knowledge building system of claim 26 , wherein said phrase object processing module is further configured to identify, for said each of said one or more sentences, preliminarily, a sentence type based on sentence terminators and question terminators predefined in said language object, wherein said sentence type comprises one of a proposition and a question and finally, a refined sentence type (or proposition type) or a refined query type, based on the matching of the language phrase objects in the sentence phrase list to the sentence part types defined for the candidate refined sentence type or query type, wherein said refined sentence type or proposition type comprise of ObjectClassDescriptor, ClassStructureDescriptor, ClassBehaviourDescriptor, ObjectStructureDescriptor, ActionDescriptor and wherein said refined query type comprise of ClassStructureQuery, ClassBehaviourQuery, ObjectStructureQuery, ObjecyBehaviourQuery, ActionQuery, wherein the refined sentence types and refined query types constitute a communication ontology valid for the language and are independent of the domain-specific semantics or contextual semantics or meanings of the individual words in the sentence or query.

45. The language processing and knowledge building system of claim 25 , wherein said phrase object processing module is further configured to update a last question attribute in said discourse context when said identified sentence type is said question, to facilitate processing of a possible answer in a subsequent sentence even if said subsequent sentence omits or lacks words for one or more sentence part types normally expected for sentence types.

46. The language processing and knowledge building system of claim 26 , wherein said phrase object processing module is further configured to identify natural language phrase objects for each unmapped sentence part type using a value for the last question attribute in the discourse context and valid-in-interim association rules.

47. The system of claim 26 , wherein the textual data may be input by way of any of: i) text meant to be read and understood by the system for building up the knowledge to be used in other computer applications such as question-answering-systems, ii) text meant to be read and understood by the system for building an abstractive summary of the text in a summarization-system, iii) user's question, meant to be answered by the system, in a question-answering-system, or iv) user's observations, questions, requests, commands or comments in the course of a conversation in computer applications such as conversation-systems.

48. The system of claim 26 , wherein the textual data may be a single word or a short phrase or a full sentence or a question or a paragraph consisting of multiple sentences and or questions or a chunk of text consisting of multiple paragraphs or may be the content of an entire text file.

49. The system of claim 26 , wherein the textual data is received in one shot and constitutes the entire first input apart, from the language object, for the complete operation of the method or system, on which the system is able to complete and completes all the steps of the method without any further interaction with the user or intervention from the user or any further input from the user.

4887212 December 1989 Zamora
5083268 January 1992 Hemphill et al.
5386556 January 1995 Hedin et al.
6278967 August 2001 Akers
7464026 December 2008 Calcagno et al.
7613666 November 2009 Baisley
7739102 June 2010 Bender
7823065 October 2010 Hintz
7865358 January 2011 Green
8078450 December 2011 Anisimovich et al.
8150677 April 2012 Menezes
8234106 July 2012 Marcu et al.
8332394 December 2012 Fan
8346563 January 2013 Hjelm et al.
8666730 March 2014 Todhunter et al.
2003/0004706 January 2003 Yale
2007/0106493 May 2007 Sanfilippo
2008/0091408 April 2008 Roulland
2010/0235165 September 2010 Todhunter
2012/0016678 January 2012 Gruber
2012/0253793 October 2012 Ghannam et al.
2012/0303356 November 2012 Boyle
2014/0039879 February 2014 Vadim Berman
2016/0147736 May 2016 Danielyan
1589440 October 2005

edspgr.10496749