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 .
Response to Arguments
Applicant’s argument filed 02/05/2026 have been fully considered but they are not persuasive.
Applicant’s Argument: On page 8-10 of Applicant’s response to rejections under 35 U.S.C. 101, applicant states “Thus, the first visual representation, i.e., the local level matrix provides an explainability framework for AI by which interpretation of the model's outputs is supported, and system improvement or reinforcement more readily made by reviewers as they can understand the reasoning behind the AI's decisions. Therefore, the claimed method recite a specific way of explaining the reasoning behind the AI's decision by presenting a justification (explanation) of an AI model's prediction by extracting word embeddings of the sentences by the intent classifier in order to solve a problem of conventional black-box AI systems that they may be inefficient in providing an analysis of the solution that they offer.”
Examiner’s Response: Applicant’s argument is not persuasive. During examination, the examiner should analyze the "improvements" consideration by evaluating the specification and the claims to ensure that a technical explanation of the asserted improvement is present in the specification, and that the claim reflects the asserted improvement (see MPEP §2106.05(a)). The MPEP (§2106.05(a)(II)) also warns, “it is important to keep in mind that an improvement in the abstract idea itself (e.g. a recited fundamental economic concept) is not an improvement in technology.” Here, the alleged improvement in the form of “generating similarity scores between the first word embeddings and the second word embeddings” is an improvement to the abstract idea of a mental process that can be performed in the human mind.
An important consideration in determining whether a claim improves technology is the extent to which the claim covers a particular solution to a problem or a particular way to achieve a desired outcome, as opposed to merely claiming the idea of a solution or outcome (see MPEP 2106.05(a)). The amended claims do not provide sufficient details to describe any technological improvement. If the specifications explicitly set forth an improvement but in a conclusory manner (see MPEP 2106.04(d)(1): a bare assertion of an improvement without the detail necessary to be apparent to a person of ordinary skill in the art), the examiner should not determine the claim improves technology.
Applicant argues that “the claimed method recite a specific way of explaining the reasoning behind the AI's decision by presenting a justification (explanation) of an AI model's prediction”. The claim as a whole does not provide any details on how the prediction explanation model generates the rationale behind the intent classifier’s prediction. Claim 1 recites “generating, ..., similarity scores for all possible pairings of word embeddings between the first word embeddings and the second word embeddings” and there are not details recited on how the generation of the similarity scores are performed by the prediction explanation model. The first visual representation does not integrate the abstract idea into a practical application nor cannot provide significantly more than the abstract idea itself because it is only presenting the results of the model onto a display. The presentation of the first visual representation is considered insignificant extra solution activity and well understood, routine and conventional activity of gathering and analyzing information using conventional techniques and displaying the result as identified by the court (see MPEP 2106.05(d)).
Applicant’s Argument: On page 10-13 of Applicant’s response to rejections under 35 U.S.C. 103, applicant states “Malkiel merely teaches about an ITBS - interpreting text-based similarity - model that explains the similarity between a seed item and a recommended item by producing an interpretable explanation for the similarity of textual paragraphs. However, amended claim 1 of the present invention recites about a prediction explanation model, that provides rationale for an intent prediction made by an artificial intelligence based intent classifier. Therefore, Malkiel merely teaches similarity between a seed item and a recommended item but does not teach or suggest the rationale behind recommending the similarity.”
Examiner’s Response: Applicant’s argument is not persuasive. Malkiel (par. 41-42) teaches the use of similarity scores to explain that two textual passages are similar. Malkiel discloses that the similarity scores is the rationale that explains the recommended item output from the model is similar to the query input. Likewise, claim 1 recites “the similarity score for each of the possible pairings are presented as individual cells in the matrix, and indicates a rationale for the selection of the first intent whether or not the intent classifier was correct”. The claim limitation recites that the generated similarity score is a rationale for the selection of the first intent. Under the broadest reasonable interpretation, the claimed invention recites determining similarity scores between the output of the model and the query input and the similarity scores are presented to indicate a rationale for the prediction. Thus, the similarity score disclosed by Malkiel teaches the rationale behind the prediction of the recommended item.
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 1-4, 7-11, 14-18, and 20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding Claim 1:
Subject Matter Eligibility Analysis Step 1:
Claim 1 recites “A method of providing reasoning for an intent prediction made by an artificial intelligence (AI)-based intent classifier, the method comprising” and is thus a process, one of the four statutory categories of patentable subject matter.
Subject Matter Eligibility Analysis Step 2A Prong 1:
“selecting, from a training corpus ..., a first sentence that is labeled with a first intent” (a mental process that can be performed in the human mind, i.e. judgement)
“extracting, ..., first word embeddings for the first sentence and second word embeddings for the first query” (a mental process that can be performed in the human mind with the aid of pen and paper, i.e. judgement)
“determining, ..., the first query is sufficiently similar to the first sentence to associate the first query with the first intent” (a mental process that can be performed in the human mind with the aid of pen and paper, i.e. judgement)
“generating, ..., similarity scores for all possible pairings of word embeddings between the first word embeddings and the second word embeddings” (a mental process that can be performed in the human mind with the aid of pen and paper, i.e. judgement)
Claim 1 therefore recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
"receiving, ..., a first query” (This step is directed to data gathering, which is understood to be insignificant extra solution activity - see MPEP 2106.05(g))
“..., at/by an/the intent classifier, ...” (mere instructions to apply the exception using a generic computer component - see MPEP 2106.05(f))
"passing, from the intent classifier, the extracted first word embeddings and second word embeddings to a prediction explanation model” (This step is directed to data gathering, which is understood to be insignificant extra solution activity - see MPEP 2106.05(g))
“..., at the prediction explanation model, ...” (mere instructions to apply the exception using a generic computer component - see MPEP 2106.05(f))
"presenting, via the prediction explanation model, a first visual representation of the similarity scores” (This step is directed to data gathering, which is understood to be insignificant extra solution activity - see MPEP 2106.05(g))
"wherein the first visual representation is a local-level matrix in which each of the word embeddings of the first word embeddings are shown on a first axis and each of the word embeddings of the second word embeddings are shown on a second axis, and the similarity score for each of the possible pairings are presented as individual cells in the matrix, and indicates a rationale for the selection of the first intent whether or not the intent classifier was correct” (merely specifies a particular technological environment in which the abstract idea is to take place, ie. a field of use, and thus does not integrate the abstract idea into a practical application nor cannot provide significantly more than the abstract idea itself - see MPEP 2106.05(h))
The additional elements as disclosed above alone or in combination do not integrate the judicial exception into practical application as they are mere insignificant extra solution activity in combination of generic computer functions being implemented with generic computer elements in a high level of generality to perform the disclosed abstract idea above. Therefore, Claim 1 is directed to the abstract idea.
Subject Matter Eligibility Analysis Step 2B:
"receiving, ..., a first query” (This step is directed to transmitting or receiving information, which is understood to be insignificant extra solution activity and well understood, routine and conventional activity of transmitting and receiving data as identified by the court - see MPEP 2106.05(d))
“..., at/by an/the intent classifier, ...” (mere instructions to apply the exception using a generic computer component - see MPEP 2106.05(f))
"passing, from the intent classifier, the extracted first word embeddings and second word embeddings to a prediction explanation model” (This step is directed to transmitting or receiving information, which is understood to be insignificant extra solution activity and well understood, routine and conventional activity of transmitting and receiving data as identified by the court - see MPEP 2106.05(d))
“..., at the prediction explanation model, ...” (mere instructions to apply the exception using a generic computer component - see MPEP 2106.05(f))
"presenting, via the prediction explanation model, a first visual representation of the similarity scores” (This step is directed to transmitting or receiving information, which is understood to be insignificant extra solution activity and well understood, routine and conventional activity of gathering and analyzing information using conventional techniques and displaying the result as identified by the court - see MPEP 2106.05(d))
"wherein the first visual representation is a local-level matrix in which each of the word embeddings of the first word embeddings are shown on a first axis and each of the word embeddings of the second word embeddings are shown on a second axis, and the similarity score for each of the possible pairings are presented as individual cells in the matrix, and indicates a rationale for the selection of the first intent whether or not the intent classifier was correct” (merely specifies a particular technological environment in which the abstract idea is to take place, ie. a field of use, and thus does not integrate the abstract idea into a practical application nor cannot provide significantly more than the abstract idea itself - see MPEP 2106.05(h))
The additional elements as disclosed above alone or in combination do not recite significantly more than the abstract idea itself as they are mere insignificant extra solution activity in combination of generic computer functions being implemented with generic computer elements in a high level of generality to perform the disclosed abstract idea above. Therefore, Claim 1 is subject-matter ineligible.
Regarding Claim 8:
The claim recites an article of manufacture that performs the method as described in claim 1. Therefore, claim 8 is rejected for the same reasons as disclosed for claim 1. The limitations for additional elements of claim 8 are analyzed below.
Subject Matter Eligibility Analysis Step 2A Prong 1:
Please see Step 2A Prong 1 analysis of claim 1
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to” (mere instructions to apply the exception using a generic computer component - see MPEP 2106.05(f))
Regarding Claim 15:
The claim recites a system that performs the method as described in claim 1. Therefore, claim 15 is rejected for the same reasons as disclosed for claim 1. The limitations for additional elements of claim 15 are analyzed below.
Subject Matter Eligibility Analysis Step 2A Prong 1:
Please see Step 2A Prong 1 analysis of claim 1
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“A system for providing reasoning for an intent prediction made by an artificial intelligence (AI)-based intent classifier, the system comprising one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to” (mere instructions to apply the exception using a generic computer component - see MPEP 2106.05(f))
Regarding Claims 2, 9, and 16:
Subject Matter Eligibility Analysis Step 2A Prong 1:
“concatenating, ” (a mental process that can be performed in the human mind with the aid of pen and paper, i.e. judgement)
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“concatenating, at the prediction explanation model, ” (mere instructions to apply the exception using a generic computer component - see MPEP 2106.05(f))
Regarding Claims 3, 10, and 17:
Subject Matter Eligibility Analysis Step 2A Prong 1: None
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“wherein the similarity scores are based on similarity values calculated between each pairing of word embeddings with respect to one or more of their shared semantic, syntactic, and contextual morphologies” (merely specifies a particular technological environment in which the abstract idea is to take place, ie. a field of use, and thus does not integrate the abstract idea into a practical application nor cannot provide significantly more than the abstract idea itself - see MPEP 2106.05(h))
Regarding Claims 4, 11, and 18:
Subject Matter Eligibility Analysis Step 2A Prong 1: None
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“wherein the first visual representation is a heatmap” (merely specifies a particular technological environment in which the abstract idea is to take place, ie. a field of use, and thus does not integrate the abstract idea into a practical application nor cannot provide significantly more than the abstract idea itself - see MPEP 2106.05(h))
Regarding Claims 7, 14, and 20:
Subject Matter Eligibility Analysis Step 2A Prong 1:
“selecting, from a training corpus ..., a second sentence that is labeled with a second intent” (a mental process that can be performed in the human mind, i.e. judgement)
“extracting, ..., third word embeddings for the second sentence” (a mental process that can be performed in the human mind with the aid of pen and paper, i.e. judgement)
“determining, ..., the first query is sufficiently similar to the second sentence to associate the first query with the second intent” (a mental process that can be performed in the human mind with the aid of pen and paper, i.e. judgement)
“generating, ..., second similarity scores for all possible pairings of word embeddings between the third word embeddings and the second word embeddings” (a mental process that can be performed in the human mind with the aid of pen and paper, i.e. judgement)
Subject Matter Eligibility Analysis Step 2A Prong 2 & 2B:
“..., at/by the intent classifier, ...” (mere instructions to apply the exception using a generic computer component - see MPEP 2106.05(f))
"passing, from the intent classifier, the extracted third word embeddings to a prediction explanation model” (This step is directed to data gathering, which is understood to be insignificant extra solution activity (2106.05(g) in step 2A prong 2) and well understood, routine and conventional activity of transmitting and receiving data as identified by the court (2106.05(d) in step 2B))
“..., at the prediction explanation model, ...” (mere instructions to apply the exception using a generic computer component - see MPEP 2106.05(f))
"presenting, via the prediction explanation model, a second visual representation of the similarity scores” (This step is directed to data gathering, which is understood to be insignificant extra solution activity (2106.05(g) in step 2A prong 2) and well understood, routine and conventional activity of gathering and analyzing information using conventional techniques and displaying the result as identified by the court (2106.05(d) in step 2B))
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.
Claims 1-4, 7-11, 14-18, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Malkiel (US20220318504A1) in view of Xu (US20240095460A1) and De “Explainable NLP: A Novel Methodology to Generate Human-Interpretable Explanation for Semantic Text Similarity”.
Regarding claim 1, Malkiel teaches:
“A method of providing reasoning for an intent prediction made by an artificial intelligence (AI)-based intent classifier, the method comprising” (abstract [0023], The proposed ITBS model generates explanations for text-based item recommendations from a pre-trained language model.)
“receiving, at an intent classifier, a first query” ([0027, 0029-0030, 0120], A user selection of a seed item is provided to the recommendation model. The query item may be associated with textual information that includes sentences and words.)
“selecting, from a training corpus and by the intent classifier, a first sentence ” ([0027-0031, 0034, 0105, 0131-0132], The recommendation model may be trained using a domain-specific corpus to create a domain-specific trained language model. The catalog contains a plurality of items for a given subject and includes item descriptions. For example, the domain may be for wines and the description of each item may include color, vineyard, and year. The descriptions of recommended items consist contextual information about the purpose of the item. The text-based descriptions are unlabeled. The catalog is a domain-specific information source that contains a list of items and the recommendation model may be a pre-trained on a domain-specific corpus. From par. 105, it is disclosed that the ITBS model interpret the similarity between each seed and candidate items taken from a fashion dataset. The fashion dataset contains about 1,000 items with descriptions and the ITBS model is fine-tuned on the corpus of item descriptions. Thus, it is implied that the model is trained on the catalog consisting of a plurality of items and item descriptions.)
“extracting, at the intent classifier, first word embeddings for the first sentence and second word embeddings for the first query” ([0034, 0097-0098], The recommendation model extracts feature vectors from candidate items and the seed item. The item descriptions are used to form a meaningful match and analyzing the semantics of each sentence of the item description.)
“determining, at the intent classifier, the first query is sufficiently similar to the first sentence to associate the first query with the first intent” ([0034, 0131-0132, Figure 10 & 12], The recommendation model calculates the cosine similarity between the feature vectors of candidate items and the seed items to determine the most similar item to recommend. A user may provide a seed item of a pair of gloves or a jacket to the recommendation model. The recommendation model searches a fashion catalog of clothing and garments to provide the closest match based on the similarity of the item descriptions. The recommended item represents the closest match to the user’s query and intentions. Figure 10 shows the association between a seed item description and the recommended item description because the system determines the intention of the seed item is to find similar cardigan or gloves.)
“passing, from the intent classifier, the extracted first word embeddings and second word embeddings to a prediction explanation model” ([0050-0052, Figure 2], The recommendation model generates a set of feature vectors that represent the seed item and the recommended item. The set of feature vectors is used by the ITBS model to determine the set of gradients.)
“generating, at the prediction explanation model, similarity scores for all possible pairings of word embeddings between the first word embeddings and the second word embeddings” ([0047, 0052-0053, 0060, 0090], The ITBS model calculate gradients associated with text-based paragraph describing a seed item and a recommended item. Word-pair scores are generated to associate the seed item description with the recommended item description. The model selects the word-pair when the score is above a threshold. Gradient maps are calculated between the first and second paragraph and used to produce saliency scores for every token or words in the paragraph. In some embodiments, the ITBS model is a BERT model that interprets paragraph similarity inferred by a pre-trained BERT model.)
“presenting, via the prediction explanation model, a first visual representation of the similarity scores” ([0043-0044, Figure 10], The word-pair and the scores may be output to a display for the user as shown in Figure 10.)
“wherein the first visual representation is a local-level ” ([0041-0042, 0094-0095, Figure 10], Word-scores are generated to associate a score for each word between the first and second paragraph. Words can be matched by comparing their similarity scores and/or word-pair scores indicating the degree of similarity of the meaning of the words within the domain-specific context of each sentence in which the word appeared. The word-pair and the scores are interpretable explanations that may be output to a display for the user as shown in Figure 10.)
Malkiel does not explicitly disclose an implementation of “selecting, ..., a first sentence that is labeled with a first intent” and “wherein the first visual representation is a local-level matrix in which each of the word embeddings of the first word embeddings are shown on a first axis and each of the word embeddings of the second word embeddings are shown on a second axis, and the similarity score for each of the possible pairings are presented as individual cells in the matrix”. However, Xu discloses in the same field of endeavor:
“selecting, ..., a first sentence that is labeled with a first intent” ([0025, 0034, 0036, 0040], The retrieval component may select question/answer pairs that are relevant to the input text data. The input text data may be a user query in the form of a question. The question/answer pairs may be stores with intents, which is a label that represents the classification type, which as tire pressure intent or open gas compartment intent. Xu explicitly teaches a sentence being associated with a label that indicates the classification type.)
It would be obvious to one of the ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teaching of “selecting, ..., a first sentence that is labeled with a first intent” from Xu into the teaching of Malkiel. Doing so can improve dialogue systems by implementing the use of language models to analyze text data and generate answers for user’s questions (Xu, abstract).
Malkiel in view of Xu does not explicitly disclose an implementation of “wherein the first visual representation is a local-level matrix in which each of the word embeddings of the first word embeddings are shown on a first axis and each of the word embeddings of the second word embeddings are shown on a second axis, and the similarity score for each of the possible pairings are presented as individual cells in the matrix”. However, De discloses in the same field of endeavor:
“wherein the first visual representation is a local-level matrix in which each of the word embeddings of the first word embeddings are shown on a first axis and each of the word embeddings of the second word embeddings are shown on a second axis, and the similarity score for each of the possible pairings are presented as individual cells in the matrix, and indicates a rationale for the selection of the first intent whether or not the intent classifier was correct” ([pg. 10, Section 4.1, par. 1; pg. 13-14, Section 4.2; pg. 17, Table 4], Text is broken down into smaller units called tokens, which may consists of words and the Word2Vec model creates embeddings of the tokens. A similarity matrix is created with the similarity scores and the similarity score defines the similarity between two sentences. Table 4 shows an example of a similarity matrix with one axis showing resume sentences and the other axis showing job description sentences and each individual cell containing the score. It would be obvious to a person having ordinary skills in the arts to modify Table 4 to represent the axes with words instead of sentences.)
It would be obvious to one of the ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teaching of “wherein the first visual representation is a local-level matrix in which each of the word embeddings of the first word embeddings are shown on a first axis and each of the word embeddings of the second word embeddings are shown on a second axis, and the similarity score for each of the possible pairings are presented as individual cells in the matrix” from De into the teaching of Malkiel in view of Xu. Doing so can create human-interpretable explanation for the text similarity match score (De, abstract).
Regarding Claim 8:
Claim 8 recites an article of manufacture that performs the same process as described in Claim 1. Therefore claim 8 is rejected under the same reasons mention for claim 1. The additional elements of claim 8 is addressed below by Malkiel:
“A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to” ([0121], A computer-readable medium storing instructions to perform the process of interpreting similarities between paragraph pairs.)
Regarding claim 15:
Claim 15 recites a system that performs the same process as described in Claim 1. Therefore claim 15 is rejected under the same reasons mention for claim 1. The additional elements of claim 15 is addressed below by Malkiel:
“A system for providing reasoning for an intent prediction made by an artificial intelligence (AI)-based intent classifier, the system comprising one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to” ([0121, 0144, 0147], The system for interpreting similarities between paragraph pairs is performed on various computer systems consisting of memory components.)
Regarding claims 2, 9, and 16, Malkiel teaches:
“concatenating, at the prediction explanation model, a vector representation of the first query and the first sentence” ([0098, 0100-0101, Figure 9], The seed item and the recommended item are associated with paragraph descriptions and paired together as input for the BERT model.)
Regarding claims 3, 10, and 17, Malkiel teaches:
“wherein the similarity scores are based on similarity values calculated between each pairing of word embeddings with respect to one or more of their shared semantic, syntactic, and contextual morphologies” ([0128], The word pairs are generated by matching words having similar semantic meaning.)
Regarding claims 4, 11, and 18, Malkiel in view of Xu does not explicitly disclose an implementation of “wherein the first visual representation is a heatmap”. However, De discloses in the same field of endeavor:
“wherein the first visual representation is a heatmap” ([pg. 17, Table 4], The table shows similarity scores between resume sentences and job description sentences as a heatmap.)
It would be obvious to one of the ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teaching of “wherein the first visual representation is a heatmap” from De into the teaching of Malkiel in view of Xu. Doing so can create human-interpretable explanation for the text similarity match score (De, abstract).
Regarding claims 7, 14, and 20, Malkiel teaches:
“selecting, from the training corpus and by the intent classifier, a second sentence ” ([0027-0031, 0034, 0131-0132, Figure 10], The recommendation model may select the recommended item from a list of candidate items within a particular domain. The description of candidate items may include a plurality of sentences. An example of a description of the recommended item is shown in par. 131. The first sentence of the description has the intent of gloves that are good for running and it is made out of windproof fabric. The third sentence shows the intent of smartphone-friendly design. Each of these sentences and intent is compared with the user query to determine similarities.)
“extracting, at the intent classifier, third word embeddings for the second sentence” ([0034, 0097-0098], The recommendation model extracts feature vectors from candidate items. The item descriptions for a set of candidate items are processed by the recommendation model. The model processes multiple sentences of item description. The item descriptions are used to form a meaningful match and analyzing the semantics of each sentence of the item description.)
“determining, at the intent classifier, the first query is also sufficiently similar to the second sentence to associate the first query with the second intent” ([0034, 0131-0132, Figure 10], The recommendation model identifies that word “glove” from the seed item description is similar to the same word “glove” from the recommended item description. The model also identifies that the word “touchscreen compatible” from the seed item description is similar to “smartphone-friendly” in the recommended item description. These words may be different but the words carry similar contextual meaning and intent of being functional with smartphone devices.)
“passing, from the intent classifier, the extracted third word embeddings to the prediction explanation model” ([0050-0052, Figure 2], The recommendation model generates a set of feature vectors that represent the seed item and the recommended items. The set of feature vectors is used by the ITBS model to determine the set of gradients. The candidate item descriptions contains multiple sentences which are processed by the ITBS model.)
“generating, at the prediction explanation model, second similarity scores for all possible pairings of word embeddings between the third word embeddings and the second word embeddings” ([0052-0053, 0060, 0090], The ITBS model calculate gradients associated with text-based paragraph describing a seed item and a recommended item. Word-pair scores are generated to associate the seed item description with the recommended item description. The model selects the word-pair when the score is above a threshold. Gradient maps are calculated between the first and second paragraph and used to produce saliency scores for every token or words in the paragraph. The candidate item descriptions contains multiple sentences which are processed by the ITBS model.)
“presenting, via the prediction explanation model, a second visual representation of the second similarity scores” ([0029, 0043-0044, Figure 10], The word-pair and the scores may be output to a display for the user as shown in Figure 10. One or more recommended items may be generated. It is implied that for a single query, the model may produce a visual representation for each recommended item.)
Malkiel does not explicitly disclose an implementation of “selecting, ..., a second sentence that is labeled with a second intent”. However, Xu discloses in the same field of endeavor:
“selecting, ..., a second sentence that is labeled with a second intent” ([0025, 0034, 0036, 0040, Figure 4], The retrieval component may select question/answer pairs that are relevant to the input text data. In Figure 4, question 302(1) and 302(5) may be retrieved based on the input text data. Question 302(1) describes tire pressure level and question 302(5) describes when to change the tires, which consists of a different intent.)
It would be obvious to one of the ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teaching of “selecting, ..., a first sentence that is labeled with a first intent” from Xu into the teaching of Malkiel. Doing so can improve dialogue systems by implementing the use of language models to analyze text data and generate answers for user’s questions (Xu, abstract).
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to GARY MAC whose telephone number is (703)756-1517. The examiner can normally be reached Monday - Friday 8:00 AM - 5:00 PM.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Abdullah Kawsar can be reached at (571) 270-3169. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/GARY MAC/Examiner, Art Unit 2127
/ABDULLAH AL KAWSAR/Supervisory Patent Examiner, Art Unit 2127
Prosecution Insights