DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Information Disclosure Statement
NO IDS filed.
Priority
The instant application, filed 12/08/2023, International Filing Date June/20/2019, and claims foreign priority to CN202310438906.0, filed April/20/2023.
Drawings
The drawings submitted on 12/08/2023 have been considered and accepted.
35 U.S.C. 112(f) Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following is a quotation of pre-AIA 35 U.S.C. 112, sixth paragraph:
An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier.
Such claim limitation(s) is/are: “word segmentation processing module” and “an entity recognition module,”, “part-of-speech tagging module”, “a feature scoring module”, “a word filtering module”, and “result extraction module” in Claims 8 -. These limitations are generic in the context of the art and don’t refer to any specific structure and only serve as placeholders for the structure that performs the associated function(s) without providing any information about what that structure is. MPEP 2181 I A says:
For a term to be considered a substitute for "means," and lack sufficient structure for performing the function, it must serve as a generic placeholder and thus not limit the scope of the claim to any specific manner or structure for performing the claimed function. It is important to remember that there are no absolutes in the determination of terms used as a substitute for "means" that serve as generic placeholders. The examiner must carefully consider the term in light of the specification and the commonly accepted meaning in the technological art. Every application will turn on its own facts.
PLEASE NOTE: This is NOT a rejection. Please don’t address it as a rejection. If the Applicant does not agree with the INTERPRETATION, he may argue or amend to replace the terms interpreted under 112(f) with structural terms such as “CPU” in combination with “RAM” and “ROM” as appropriately supported by the Specification. In the alternative, he may let the interpretation stand if the intent was to include a means plus function limitation in the Claim.
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph:
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific structural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Because this/these claim limitation(s) is/are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
If applicant does not intend to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-10 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding claim(s) 1, 8, 9, and 10, the limitation(s) of “acquiring”, “performing”, “performing”, “scoring”, “filtering”, and “running”, as drafted, are processes that, under broadest reasonable interpretation, covers performance of the limitation in data gathering, and human mental activity but for the recitation of generic computer components. More specifically, receiving a text is data gathering, split the text into pieces, Human can determine entity of the word/segmented text such as subject, object, predicate, which can be person, place or time representation through noun, noun phrase, verb, adverb, adjective, preposition and conjunction. human can write down a sentence on a piece of paper, split the sentence and recognize part-of-speech information and label part-of-speech information and score part-of-speech based on noun, proper and improper noun, verb, adjectives and preposition, and human can filter/remove unnecessary word, human also understand number of appearances of a single word or word -pair. If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components, then it falls within the --Mental Processes-- grouping of abstract ideas. Accordingly, the claim(s) recite(s) an abstract idea.
This judicial exception is not integrated into a practical application because the recitation of “memory, a processor and computer-readable instructions”, in claim 8, 9 and 10 and “memory” in claim 10, reads to generalized computer components, based upon the claim interpretation wherein the structure is interpreted using Paragraphs. [0099] and [0100] in the specification of US 20240354507 A1. Accordingly, these additional elements do not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. The claim(s) is/are directed to an abstract idea.
The claim(s) do(es) not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element of using generalized computer components to “acquiring”, “performing”, “performing”, “scoring”, “filtering”, and “running”, amounts to no more than mere instructions to apply the exception using a generic computer component. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept. The claim(s) is/are not patent eligible.
With respect to claim(s) 2, the claim(s) recite(s), “performing entity recognition on all improper nouns.,”, “performing entity recognition on all proper nouns …” and “determining word segmentation results …”, reads on a human write a sentence on a piece of paper and segment words and assign label as an improper noun and proper noun. No additional limitations are present. Claims directed to abstract idea.
With respect to claim(s) 3, , the claim(s) recite(s), “acquiring a preset part-of-speech list, wherein the preset part-of-speech list ...” and “performing part-of-speech tagging on all the entity recognition …” reads on a human assign/tag part-of-speech on each word according part-of-speech list. No additional limitations are present.
With respect to claim(s) 4, the claim(s) recite(s) “acquiring,” “scoring”, and “integrating”, reads on a human give score to each part-of-speech tagged word and sum that value to get total score. This claim is math algorithm involve using human mental activity. No additional limitations are present.
With respect to claim(s) 5, the claim(s) recite(s), “screening”, and “filtering” human understand all words in a sentence are part-of-speech tagged or not, after screening some insignificant words are removed. No additional limitations are present.
With respect to claim(s) 6, the claim(s) recite(s), ““running”, “““running”, and “determining”, “ which reads on a human can count the number of times word/word-pair appear in a sentence. This is mental activity of a human. Human can understand word pair co-occurrence in a word are same or not. No additional limitations are present.
With respect to claim(s) 7 the claim(s) recite(s), “acquiring”, “acquiring”, , “performing”, and “determining”, reads on a human can teach other human grammatical structure of a sentence, proper noun, nouns and nouns phrase. If human cannot answer nouns correctly in a sentence, called that can be loss. The claim mentions “training” a model, but claim is missing training step and how model is trained. No additional limitations are present.
These claims further do not remedy the judicial exception being integrated into a practical application and further fail to include additional elements that are sufficient to amount to significantly more than the judicial exception.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 10 are non-statutory under the most recent interpretation of the Interim Guidelines regarding 35 U.S.C.101 because: The specification at paragraph [0098] in US 20240354507 A1.
Claims 10 is/are drawn to a “signal” per se as recited in the preamble and as such is non-statutory subject matter. On paragraph [0098] of the Published US 20240354507 A1 Specification, the term “readable storage media" is not defined as to what the scope of the term is meant to encompass. Hence, one of ordinary skilled in the art can interpret such term to include transitory signals and non-transitory signals. It does not appear that a claim reciting a signal encoded with functional descriptive material falls within any of the categories of patentable subject matter set forth in § 101. First, a claimed signal is clearly not a "process" under § 101 because it is not a series of steps. The other three § 101 classes of machine, compositions of matter and manufactures "relate to structural entities and can be grouped as 'product' claims in order to contrast them with process claims." 1 D. Chisum, Patents § 1.02 (1994). The Applicant did not specify in the specification para [0098], computer- readable storage media excludes signal. In order to overcome the present rejection, the Applicant is advised to amend the claims by using the following terminology: "non-transitory computer- readable storage media ." Such example terminology has been also found in the Official Gazette 1351 OG 212.
Claim Rejections - 35 USC § 112
Claims 5,is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 5 recites the limitation “the alternative words”. The terms of this limitation is/are lacks antecedent basis. For examination purposes the examiner has interpreted “thealternative words” to refer to the alternative word of claim 5 in line 5.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim 1, 3, 5, and 8-10, is/are rejected under 35 U.S.C. 103 as being unpatentable over SUGIMOTO et al. US 20220027558 A1 in view of Ng Tari et al. US 20140163955 and in view of Pai et al. US 12235826 B2 and further in view of Sapugay et al. US 20210004537 A1
Regarding Claim 1, SUGIMOTO teaches:
1. A keyword extraction method, comprising: acquiring a text to be processed, and performing word segmentation on the text to be processed to obtain at least one word segmentation result; SUGIMOTO teaches (“[0069] At 501, the keywords are parsed from the text. At 502, the flow creates the co-occurrence words matrix for the text by applying keywords extraction system 102-1, dependency parser 102-2 and labeling 102-3 to the text to generate keyword candidates 102-4, dependency information 102-5, and labeling information 102-6. Then, the flow determines create co-occurrence the same way through using co-occurrence matrix creation module 105-1, 105-2 by using keyword candidates 102-4, dependency information 102-5, labeling information 102-6, and co-occurrence dictionary 104.”) By SUGIMOTO et al. US 20220027558 A1
running word co-occurrence statistics on all the target words to obtain word co-occurrence values, and SUGIMOTO teaches (“[0077] … tf-idf can be used to extract keywords …”) (“[0069] At 501, the keywords are parsed from the text. At 502, the flow creates the co-occurrence words matrix for the text by applying keywords extraction system 102-1, dependency parser 102-2 and labeling 102-3 to the text to generate keyword candidates 102-4, dependency information 102-5, and labeling information 102-6. Then, the flow determines create co-occurrence the same way through using co-occurrence matrix creation module 105-1, 105-2 by using keyword candidates 102-4, dependency information 102-5, labeling information 102-6, and co-occurrence dictionary 104.”) (“[0070] At 503, the flow calculates the confidence. In an example implementation, the formula to calculate the confidence can be sum(val_i*weights_i), wherein val_i indicates the cell value of the number i co-occurrence words and weights is the weighs of the number i co-occurrence words. Other formulas can similarly be used in accordance with the desired implementation.”) By SUGIMOTO et al. US 20220027558 A1
performing keyword extraction on all the target words based on the word co-occurrence values to obtain keyword extraction results. SUGIMOTO teaches surrounding words (i.e. all the target words). (“[0094] Depending on the desired implementation, processor(s) 2010 can be configured to determine for the selected ones of the keywords included in the co-occurrence dictionary, the surrounding words to be associated with the selected ones of the keywords in the co-occurrence dictionary based on the number of instances of co-occurrence of the surrounding words with the selected ones of the keywords by: for an extracted keyword being preceded by another extracted keyword, determining the another extracted keyword as being an instance of co-occurrence of the extracted keyword and incrementing a counter; and for the counter exceeding a threshold, determining the another extracted keyword as one of the surrounding words as illustrated at FIGS. 11-15.”)
(“[0038] FIG. 2 illustrates an example of the co-occurrence dictionary creation module 102, in accordance with an example implementation. In the example implementation of FIG. 2, the text 101 is processed by three modules: keywords extraction system 102-1, dependency parser 102-2 and labeling 102-3. The keyword extraction system 102-1 extracts keyword candidates from all texts. The result may include words which are not actually keywords for all texts. The keyword extraction system 102-1 can be any type of system such as tf-idf, topic modeling or as known to a person of ordinary skill in the art in accordance with the desired implementation. The result is the keyword candidates 102-4.”) (“[0069] At 501, the keywords are parsed from the text. At 502, the flow creates the co-occurrence words matrix for the text by applying keywords extraction system 102-1, dependency parser 102-2 and labeling 102-3 to the text to generate keyword candidates 102-4, dependency information 102-5, and labeling information 102-6. Then, the flow determines create co-occurrence the same way through using co-occurrence matrix creation module 105-1, 105-2 by using keyword candidates 102-4, dependency information 102-5, labeling information 102-6, and co-occurrence dictionary 104.”) (“[0077] …, tf-idf can be used to extract keywords and select sentences which contain many keywords over a threshold.”) By SUGIMOTO et al. US 20220027558 A1
SUGIMOTO does not explicitly teach performing entity recognition on all the word segmentation results through a preset entity recognition model to obtain at least one entity recognition result.
Ng Tari teaches:
performing entity recognition on all the word segmentation results through a preset entity recognition model to obtain at least one entity recognition result; FIG. 5, ELEMENT 510. Ng Tari teaches (“[0013] FIG. 5 is a flow chart of a method for extracting ontological information from a body of text shown in FIGS. 2 and 4.”) (“[0049] Further, from subset of noun phrases 412, concept extraction system 400 classifies noun phrases as either an entity 430, such as "inlet plenum" (shown in FIG. 3 as first noun 326 and second noun 328), or a property 432, such as proper noun 316 "water". … … Named-entity-recognition module 402 stores classification information for entity 430 and property 432 in parse tree database 212. …”) by Ng Tari et al. US 20140163955 A1 (“[0053] In the exemplary embodiment, named-entity-recognition module 402 and concept-extraction module 420 submit PTQL queries 220 (shown in FIG. 2) to parse tree database 212 when performing execution tasks such as, without limitation, identifying subset of sentences 410 with verb phrase 406, identifying subset of noun phrases 412 related to verb phrase 406 within the subset of sentences 410, classifying noun phrases as either an entity or a property, and identifying conceptual relationship 434 between entity 430 and property 432.”) [0054] FIG. 5 is a flow chart of an exemplary method 500 for extracting ontological information from body of text 202 (shown in FIGS. 2 and 4). More specifically, method 500 extracts conceptual relationship 434 (shown in FIG. 4) from body of text 202 (shown in FIG. 2). Body of text 202 is converted 502 into parse tree format. Verb phrase 406 (shown in FIG. 4) is identified 504. Subset of sentences 410 (shown in FIG. 4) is identified 506 from body of text 202 using verb phrase 406. From subset of sentences 410, subset of noun phrases 412 (shown in FIG. 4) is identified 508. From subset of noun phrases 412, entity 430 (shown in FIG. 4) and property 432 (shown in FIG. 4) are classified 510. Conceptual relationship 434 (shown in FIG. 4) between entity 430 and property 432 is identified 512, and is then output 514. In some embodiments, conceptual relationship 434 may be output 514 to user 156 (shown in FIG. 1) using computing system 120 (shown in FIG. 1). …”) by Ng Tari et al. US 20140163955 A1
Ng Tari is considered to be analogous to the claimed invention because it relates to generally to language processing systems and, more particularly, to techniques for extracting ontological information from a body of text.
Therefore, it would have been obvious for someone of ordinary skill in the art before the effective filing date of the claimed invention to modify SUGIMOTO, to incorporate the teachings of Ng Tari in order to including part-of-speech tagging on entity recognition.
One could have been motivated to do because system can detect duplicate entities. (“[0050] … Alternatively, normalizing module 450 may use any methods of normalizing like nouns and noun phrases. Methods of normalizing help to reduce what may be duplicate entities. …‘’) by Ng Tari et al. US 20140163955 A1
The combination does not explicitly teach performing part-of-speech tagging on all the entity recognition results to obtain part-of-speech tagging results.
Pai teaches:
performing part-of-speech tagging on all the entity recognition results to obtain part-of-speech tagging results; Pai teaches (“(127) According to an embodiment of the present invention, the task of clustering of articles corresponding to similar events may involve the following sub-tasks: (a) Clean the articles in the corpus in a manner similar to de-duplication; (b) Tokenize the articles and apply Named Entity Recognition (NER) to extract and remove all named entities from the corpus; (c) Perform position (POS) tagging of tokens and keep only nouns, adjectives, adverbs and verbs for each article (e.g., union set of leftover tokens forms the vocabulary for clustering); (d) Create various matrices for trying clustering like tf-idf, count, binary, etc.; (e) Use clustering algorithms on each matrix and get cluster labels; …” col. 18, lines 5-30) by Pai et al. US 12235826 B2
Pai is considered to be analogous to the claimed invention because it relates to system performing machine driven analysis and determination of integrity due diligence risk associated with third party entities.
Therefore, it would have been obvious for someone of ordinary skill in the art before the effective filing date of the claimed invention to modify SUGIMOTO, and Ng Tari, to incorporate the teachings of Pai in order to including part-of-speech tagging on entity recognition.
One could have been motivated to do because system can improve on keyword classification. (“(81) … Machine learning models may be used along with the keywords to improve the classification. In addition, one article may belong to more than one section. …‘’) by Pai et al. US 12235826 B2
The combination does not explicitly teach filtering all the part-of-speech tagging results based on all the score values to obtain at least one target word.
Sapugay teaches:
filtering all the part-of-speech tagging results based on all the score values to obtain at least one target word; and Sapugay teaches modify or remove potential POS taggings generated by the POS plug-in 318 that are known to be erroneous. Also teaches scoring part-of-speech and comparing with threshold value. If threshold is below then entry is not included. (“[0097] For the embodiment illustrated in FIG. 11, the set of plug-ins 314 of the structure subsystem 172 include: a part of speech (POS) plug-in 318, one or more correction plug-in(s) 320, a variation filter (VF) plug-in 322, a parser plug-in 354, and a final scoring and filtering (FSF) plug-in 326. The functions of the plug-ins 314 are discussed in greater detail with respect to FIG. 12. In general, the POS plug-in 318 includes a ML-based component (e.g., a feedforward artificial neural network) that is trained to perform POS tagging of each token of an utterance with an associated part of speech (e.g., verb, noun, adjective, pronoun, adverb). The POS plug-in 318 is designed to output multiple potential POS taggings of an utterance, as well as corresponding confidence scores for each potential POS tagging of the utterance. The correction plug-in(s) 320 include a POS correction plug-in that applies ML-based techniques or applies rules (e.g., stored in the database 106) to modify or remove potential POS taggings generated by the POS plug-in 318 that are known to be erroneous. The VF plug-in 322 applies a mathematical comparison of potential POS taggings generated by the POS plug-in 318, and removes POS taggings that are not sufficiently different from one another. The parser plug-in 324 may include a rules-based or ML-based component (e.g., a feedforward artificial neural network) that is designed and/or trained to generate a respective meaning representation for each of the remaining candidate POS taggings, as well as corresponding confidence scores for the parsing operation. The correction plug-in(s) 320 also include a parser correction plug-in that applies ML-based techniques or applies rules (e.g., stored in the database 106) to modify or remove potential meaning representations generated by the parser plug-in 324 that are known to be erroneous. The FSF plug-in 326 determines a final confidence score for each generated meaning representation, and then outputs the final set 304 of meaning representations having a corresponding final confidence score that is greater than a predefined threshold.”) (“[0104] … For example, the difference value assigned to a verb-verb comparison, a noun-noun comparison, an adjective-adjective, etc., may be zero; the difference value assigned to a command form verb-verb comparison may be slightly greater than zero (e.g., 0.1); the difference value assigned to a verb-noun comparison, a verb-adjective, a noun-adjective, etc., may be one, and so forth, within the database 106. In certain embodiments, the database 106 may further store weighting values for different POS tags, such that certain POS tags (e.g., verbs) have a greater contribution to the output of the variability function than other POS tags (e.g., nouns, pronouns). For this example, the weights of the POS tags are equivalent. As such, the variability function may calculate a variability score between the first and second entries (e.g., (0.1 for the difference between first tags+0 for the difference between the second tags+0 for the difference between the third tags)/(3 tags compared)=0.03), and then compare this variability score to the variation threshold value 364 (e.g., 0.3). Since the variability score is below the variation threshold value 364, the second entry is not included in the final nominee set 362 of POS taggings. This process continues with the third entry in the corrected set 358 of potential POS taggings being compared to the first entry (e.g., (1 for the difference between first tags+0 for the difference between the second tags+0 for the difference between the third tags)/(3 tags compared)=0.3, which is at the variation threshold value of 0.3), and the third entry is included in the final nominee set 362 of potential POS taggings. Subsequently, the fourth entry in the corrected set 358 of potential POS taggings is compared to the first entry (e.g., (0.1 for the difference between first tags+1 for the difference between the second tags+0 for the difference between the third tags)/(3 tags compared)=0.33, which is greater than the variation threshold value of 0.3), and also compared to the third entry (e.g., (1 for the difference between first tags+1 for the difference between the second tags+0 for the difference between the third tags)/(3 tags compared)=0.66, which is greater than the variation threshold value of 0.3), and is also included in the final nominee set 362 of potential POS taggings that are carried forward in the process 340.”) by Sapugay et al. US 20210004537 A1
scoring all the part-of-speech tagging results through a preset scoring metric to obtain score values; Sapugay teaches (“[0097] For the embodiment illustrated in FIG. 11, the set of plug-ins 314 of the structure subsystem 172 include: a part of speech (POS) plug-in 318, one or more correction plug-in(s) 320, a variation filter (VF) plug-in 322, a parser plug-in 354, and a final scoring and filtering (FSF) plug-in 326. The functions of the plug-ins 314 are discussed in greater detail with respect to FIG. 12. In general, the POS plug-in 318 includes a ML-based component (e.g., a feedforward artificial neural network) that is trained to perform POS tagging of each token of an utterance with an associated part of speech (e.g., verb, noun, adjective, pronoun, adverb). The POS plug-in 318 is designed to output multiple potential POS taggings of an utterance, as well as corresponding confidence scores for each potential POS tagging of the utterance. … … The parser plug-in 324 may include a rules-based or ML-based component (e.g., a feedforward artificial neural network) that is designed and/or trained to generate a respective meaning representation for each of the remaining candidate POS taggings, as well as corresponding confidence scores for the parsing operation. The correction plug-in(s) 320 also include a parser correction plug-in that applies ML-based techniques or applies rules (e.g., stored in the database 106) to modify or remove potential meaning representations generated by the parser plug-in 324 that are known to be erroneous. The FSF plug-in 326 determines a final confidence score for each generated meaning representation, and then outputs the final set 304 of meaning representations having a corresponding final confidence score that is greater than a predefined threshold.”) (“[0104] … Next, the VF plug-in 322 may consider the second entry in the corrected set 358 of potential POS taggings by comparing it to the first entry using the variability function. An example variability function may be a weighted average. For this example, when the first and second entries are compared, the first tag (e.g., command form of verb) of the first entry and the first tag (e.g., verb) of the second entry are compared. Difference values for different tag comparisons may be stored as part of the rules 114 in the database 106. For example, the difference value assigned to a verb-verb comparison, a noun-noun comparison, an adjective-adjective, etc., may be zero; the difference value assigned to a command form verb-verb comparison may be slightly greater than zero (e.g., 0.1); the difference value assigned to a verb-noun comparison, a verb-adjective, a noun-adjective, etc., may be one, and so forth, within the database 106. In certain embodiments, the database 106 may further store weighting values for different POS tags, such that certain POS tags (e.g., verbs) have a greater contribution to the output of the variability function than other POS tags (e.g., nouns, pronouns). For this example, the weights of the POS tags are equivalent. As such, the variability function may calculate a variability score between the first and second entries (e.g., (0.1 for the difference between first tags+0 for the difference between the second tags+0 for the difference between the third tags)/(3 tags compared)=0.03), and then compare this variability score to the variation threshold value 364 (e.g., 0.3). Since the variability score is below the variation threshold value 364, the second entry is not included in the final nominee set 362 of POS taggings….”) by Sapugay et al. US 20210004537 A1
Sapugay is considered to be analogous to the claimed invention because it relates to generally to the fields of natural language understanding (NLU) and artificial intelligence (AI), and more specifically, to a hybrid learning system for NLU.
Therefore, it would have been obvious for someone of ordinary skill in the art before the effective filing date of the claimed invention to modify SUGIMOTO, Ng tari and Pai to incorporate the teachings of Sapugay in order to scoring all the part-of-speech tagging results through a preset scoring metric to obtain score values.
One could have been motivated to do because system can achieve accurate similarity scores. (“[0123] … After each comparison, the meaning search subsystem 152 may refine set 430 of similarity scores generated via a previous iteration, or alternatively, replace each previously-generated similarity score of the set 430 with its more accurate counterpart. Indeed, because more processing resources are utilized during application of subsequent comparison functions 410, the set 430 of similarity scores is generally improved in accuracy and/or precision as additional comparison functions 410 are applied. ...”) by Sapugay et al. US 20210004537 A1
Claim 8 is a device claim with a limitation similar to the limitation of method Claim 1 and is rejected under similar rationale. Additionally,
A keyword extraction device, comprising: SUGIMOTO teaches (“[0038] FIG. 2 illustrates an example of the co-occurrence dictionary creation module 102, in accordance with an example implementation. In the example implementation of FIG. 2, the text 101 is processed by three modules: keywords extraction system 102-1, dependency parser 102-2 and labeling 102-3. The keyword extraction system 102-1 extracts keyword candidates from all texts. The result may include words which are not actually keywords for all texts. The keyword extraction system 102-1 can be any type of system such as tf-idf, topic modeling or as known to a person of ordinary skill in the art in accordance with the desired implementation. The result is the keyword candidates 102-4.”) by SUGIMOTO et al. US 20220027558 A1
a word segmentation processing module, SUGIMOTO teaches (“[0038] … , the text 101 is processed by three modules: keywords extraction system 102-1, dependency parser 102-2 and labeling 102-3. …”) by SUGIMOTO et al. US 20220027558 A1
an entity recognition module SUGIMOTO teaches (“[0038] … , the text 101 is processed by three modules: keywords extraction system 102-1, dependency parser 102-2 and labeling 102-3. …”) by SUGIMOTO et al. US 20220027558 A1
a part-of-speech tagging module SUGIMOTO teaches (“[0038] … , the text 101 is processed by three modules: keywords extraction system 102-1, dependency parser 102-2 and labeling 102-3. …”) by SUGIMOTO et al. US 20220027558 A1
a feature scoring module SUGIMOTO teaches (“[0037] The confidence calculation module 109 calculates the confidence for each keyword 110 from co-occurrence matrix 106-2 and the weights for each co-occurrence words 108.”) by SUGIMOTO et al. US 20220027558 A1
a result extraction module, SUGIMOTO teaches (“[0035] In the first aspect involving the calculation of weights, the input texts are collected at 101. Given the input texts 101, the co-occurrence dictionary creation module 102 creates a co-occurrence dictionary 104 from the texts 101 by using dependency rules 103. The co-occurrence matrix creation module 105-1 creates a co-occurrence matrix 106-1 from the co-occurrence dictionary 104. The weights calculation module 107 calculates weights for each co-occurrence words 108 from co-occurrence matrix 106-1.”) (“[0069] … Then, the flow determines create co-occurrence the same way through using co-occurrence matrix creation module 105-1, 105-2 by using keyword candidates 102-4, dependency information 102-5, labeling information 102-6, and co-occurrence dictionary 104.”) by SUGIMOTO et al. US 20220027558 A1
Sapugay further teaches:
a word filtering module, Sapugay teaches (“[0097] For the embodiment illustrated in FIG. 11, the set of plug-ins 314 of the structure subsystem 172 include: a part of speech (POS) plug-in 318, one or more correction plug-in(s) 320, a variation filter (VF) plug-in 322, a parser plug-in 354, and a final scoring and filtering (FSF) plug-in 326. …”) by Sapugay et al. US 20210004537 A1
Sapugay is considered to be analogous to the claimed invention because it relates to generally to the fields of natural language understanding (NLU) and artificial intelligence (AI), and more specifically, to a hybrid learning system for NLU.
Therefore, it would have been obvious for someone of ordinary skill in the art before the effective filing date of the claimed invention to modify SUGIMOTO, Ng tari and Pai to incorporate the teachings of Sapugay in order to scoring all the part-of-speech tagging results through a preset scoring metric to obtain score values.
One could have been motivated to do because system can achieve accurate similarity scores. (“[0123] … After each comparison, the meaning search subsystem 152 may refine set 430 of similarity scores generated via a previous iteration, or alternatively, replace each previously-generated similarity score of the set 430 with its more accurate counterpart. Indeed, because more processing resources are utilized during application of subsequent comparison functions 410, the set 430 of similarity scores is generally improved in accuracy and/or precision as additional comparison functions 410 are applied. ...”) by Sapugay et al. US 20210004537 A1
Regarding Claim 9, the combination teaches the method of claim 1 as identified above.
SUGIMOTO further teaches:
9. A computer equipment, comprising a memory, a processor and computer-readable instructions stored in the memory and executable by the processor, wherein when the processor executes the computer-readable instructions, the keyword extraction method of claim 1 is realized. SUGIMOTO teaches [0082] FIG. 20 illustrates an example computing environment with an example computer device suitable for use in some example implementations. Computer device 2005 in computing environment 2000 can include one or more processing units, cores, or processors 2010, memory 2015 (e.g., RAM, ROM, and/or the like), internal storage 2020 (e.g., magnetic, optical, solid state storage, and/or organic), and/or IO interface 2025, any of which can be coupled on a communication mechanism or bus 2030 for communicating information or embedded in the computer device 2005…”) (“[0087] Computer device 2005 can use and/or communicate using computer-usable or computer-readable media, including transitory media and non-transitory media. …”) (“[0088] Computer device 2005 can be used to implement techniques, methods, applications, processes, or computer-executable instructions in some example computing environments. Computer-executable instructions can be retrieved from transitory media, and stored on and retrieved from non-transitory media. The executable instructions can originate from one or more of any programming, scripting, and machine languages (e.g., C, C++, C#, Java, Visual Basic, Python, Perl, JavaScript, and others).”) By SUGIMOTO et al. US 20220027558 A1
Regarding Claim 10, The combination teaches the method of claim 1 as identified above.
SUGIMOTO further teaches:
10. One or more readable storage media, storing computer-readable instructions, wherein when the computer-readable instructions are executed by one or more processors, the keyword extraction method of claim 1 is implemented. SUGIMOTO teaches (“[0087] Computer device 2005 can use and/or communicate using computer-usable or computer-readable media, including transitory media and non-transitory media. …”) (“[0088] Computer device 2005 can be used to implement techniques, methods, applications, processes, or computer-executable instructions in some example computing environments. Computer-executable instructions can be retrieved from transitory media, and stored on and retrieved from non-transitory media. The executable instructions can originate from one or more of any programming, scripting, and machine languages (e.g., C, C++, C#, Java, Visual Basic, Python, Perl, JavaScript, and others).”) By SUGIMOTO et al. US 20220027558 A1
Regarding Claim 3, The combination teaches the method of claim 1 as identified above.
SUGIMOTO further teaches:
3. The keyword extraction method of claim 1, acquiring a preset part-of-speech list, wherein the preset part-of-speech list comprises at least one target part of speech; and . SUGIMOTO teaches NN (Noun), ADJ (adjective), ADV (Adverb). (i.e. preset part-f-speech list) (“[0046] FIG. 6 illustrates dependency rules 103, in accordance with an example implementation. Dependency rules 103 are rules to extract keywords and surrounding words from text. These rules are developed by human resources, or otherwise in accordance with the desired implementation. Depending on the desired implementation, the dependency rules 103 can be regular expression of labels with identification of keywords, such as NN (Noun), ADJ (adjective), ADV (Adverb) as shown in FIG. 5, and the “K” after label identifies keywords. The system searches dependency information of the text which fits the dependency rules. Details of the searching of dependency information are described with respect to FIG. 11.”)
performing part-of-speech tagging on all the entity recognition results based on all the target parts of speech to obtain part-of-speech tagging results. SUGIMOTO teaches (“ [0054] At 204, the flow checks to determine if keyword Ai satisfies the dependency rule (204). This can be done as follows. First, the flow selects the text which includes the keyword Ai. Next, from dependency information 102-5 of the text, dependency paths are created. Dependency paths are all paths which follow the parents to the children in the text. For example, three dependency paths are created from the text “please verify the location of the well”, which includes “Please verify location the”, “please verify location well of”, “please verify location well the”. Then, the flow replaces each word in the dependency path with a label from 102-6. In this step, for the keyword Ai, “K” is added at the end of the label. For example, “please verify location the” are converted into “VB VB NNK DT”, “please verify location well of” is converted into “VB VB NNK NN IN”, and “please verify location well the” are converted into “VB VB NNK NN DT”. Finally, the flow checks if the label satisfies the dependency rule which involve the regex pattern of the labels. For example, there are two dependency rules, those are “NNK (ADJ|ADV|NN)+” and “(ADV)*VB NNK”. “VB VB NNK” that is “please verify location” matches “(ADV)*VB NNK”, and “NNK NN” that is “location well” matches “NNK (ADJ|ADV|NN)+”. If some dependency path matches the dependency rule, a count is incremented for keyword A at 205. If the dependency paths do not match the dependency rule, a count is incremented for non-keyword A at 207.”) By SUGIMOTO et al. US 20220027558 A1
Ng Tari further teaches:
wherein the step of performing part-of-speech tagging on all the entity recognition results to obtain part-of-speech tagging results comprises: FIG. 5, ELEMENT 510. Ng Tari teaches (“[0013] FIG. 5 is a flow chart of a method for extracting ontological information from a body of text shown in FIGS. 2 and 4.”) (“[0049] Further, from subset of noun phrases 412, concept extraction system 400 classifies noun phrases as either an entity 430, such as "inlet plenum" (shown in FIG. 3 as first noun 326 and second noun 328), or a property 432, such as proper noun 316 "water". … … Named-entity-recognition module 402 stores classification information for entity 430 and property 432 in parse tree database 212. …”) (“[0052] … The above PTQL query defines the pattern for entity-property relation extraction. The constructs "Tag=`Entity`" and "Tag=`Property`" correspond to noun phrases 412 that have been identified as entities 430 and properties 432. This PTQL query defines a syntactic constraint that, within a noun phrase 412, an identified property is followed by a prepositional phrase that includes a preposition and an identified entity. The returning entities and properties, i.e., the values of kw1 and kw2, are deemed to have entity-property relations, denoted as <entity, property>. …”) (“[0053] In the exemplary embodiment, named-entity-recognition module 402 and concept-extraction module 420 submit PTQL queries 220 (shown in FIG. 2) to parse tree database 212 when performing execution tasks such as, without limitation, identifying subset of sentences 410 with verb phrase 406, identifying subset of noun phrases 412 related to verb phrase 406 within the subset of sentences 410, classifying noun phrases as either an entity or a property, and identifying conceptual relationship 434 between entity 430 and property 432.”) [0054] FIG. 5 is a flow chart of an exemplary method 500 for extracting ontological information from body of text 202 (shown in FIGS. 2 and 4). More specifically, method 500 extracts conceptual relationship 434 (shown in FIG. 4) from body of text 202 (shown in FIG. 2). Body of text 202 is converted 502 into parse tree format. Verb phrase 406 (shown in FIG. 4) is identified 504. Subset of sentences 410 (shown in FIG. 4) is identified 506 from body of text 202 using verb phrase 406. From subset of sentences 410, subset of noun phrases 412 (shown in FIG. 4) is identified 508. From subset of noun phrases 412, entity 430 (shown in FIG. 4) and property 432 (shown in FIG. 4) are classified 510. Conceptual relationship 434 (shown in FIG. 4) between entity 430 and property 432 is identified 512, and is then output 514. In some embodiments, conceptual relationship 434 may be output 514 to user 156 (shown in FIG. 1) using computing system 120 (shown in FIG. 1). …”) by Ng Tari et al. US 20140163955 A1
Ng Tari is considered to be analogous to the claimed invention because it relates to generally to language processing systems and, more particularly, to techniques for extracting ontological information from a body of text.
Therefore, it would have been obvious for someone of ordinary skill in the art before the effective filing date of the claimed invention to modify SUGIMOTO, Ng tari and Pai and Sapugay, to incorporate the teachings of Ng Tari in order to including part-of-speech tagging on entity recognition.
One could have been motivated to do because system can detect duplicate entities. (“[0050] … Alternatively, normalizing module 450 may use any methods of normalizing like nouns and noun phrases. Methods of normalizing help to reduce what may be duplicate entities. …‘’) by Ng Tari et al. US 20140163955 A1
The combination does not explicitly teach step of performing part-of-speech tagging on all the entity recognition results to obtain part-of-speech tagging results.
Pai teaches:
wherein the step of performing part-of-speech tagging on all the entity recognition results to obtain part-of-speech tagging results comprises: Pai teaches (“(127) According to an embodiment of the present invention, the task of clustering of articles corresponding to similar events may involve the following sub-tasks: (a) Clean the articles in the corpus in a manner similar to de-duplication; (b) Tokenize the articles and apply Named Entity Recognition (NER) to extract and remove all named entities from the corpus; (c) Perform position (POS) tagging of tokens and keep only nouns, adjectives, adverbs and verbs for each article (e.g., union set of leftover tokens forms the vocabulary for clustering); (d) Create various matrices for trying clustering like tf-idf, count, binary, etc.; (e) Use clustering algorithms on each matrix and get cluster labels; …” col. 18, lines 5-30) by Pai et al. US 12235826 B2
Pai is considered to be analogous to the claimed invention because it relates to system performing machine driven analysis and determination of integrity due diligence risk associated with third party entities.
Therefore, it would have been obvious for someone of ordinary skill in the art before the effective filing date of the claimed invention to modify SUGIMOTO, and Ng Tari, to incorporate the teachings of Pai in order to including part-of-speech tagging on entity recog