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 .
DETAILED ACTION
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-5, 7-20, 22-25 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. Claims 1, 24, 25 recite the term “the intent classification space,” in a manner lacking clear antecedent which renders the claims indefinite. Further, the recited intent classification space creates ambiguity in the claim as it is structurally unrelated to the remainder of the claim and thus unclear as to what data the classification space draws from or embodies. Construing the recitation broadly in light of the specification the amendment appears to resolve a recited “multi-dimensional space … indicative of an intent classification,” as recited in ¶ 65 of the instant specification and similarly recited in ¶ 48 and ¶ 66. The specified multi-dimensional space requires a user input, however the recited first dataset of the independent claims does not mandate input data but only “at least one of user input data, vector data representative of the user input data, and model output data,” and in such a case as the presence of only model output data in the dataset the recited “the intent classification space,” only grows more ambiguous and indefinite. Further Examiner cannot clearly resolve an appropriate antecedent in the specification for the amended recited “determining that a characteristic of a distribution of user input data within the intent classification space of the trained model has changed.” Certainly there is discussion in the specification of determinations of distributions of user intent and determinations of differences or changes to distributions of data but determining the metes and bounds by combining the specified disclosure to arrive at the full recitation as quoted again results in the conclusion of at least indefinite subject matter. Claims 2-5, 7-20, 22, 23 do not remedy and are similarly rejected. Claim 8 additionally recites “the system of claim 1,” where claim 1 recites a method and mentions no system. Appropriate correction is required
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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 of this title, 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.
The factual inquiries set forth in Graham v. John Deere Co., 383 U.S. 1, 148 USPQ 459 (1966), that are applied for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1, 2, 7-11, 13-20, 22-25 rejected under 35 U.S.C. 103 as being unpatentable over Gopalan: 20180285772 hereinafter Gop further in view of Huang: 20210141862 hereinafter Hua and further in view of Wohlwend: 20200151253 hereinafter Woh.
Regarding claim 1
Gop teaches:
A method for monitoring concept drift in a trained model (Gop: ¶ 10-12, 27, 32: system for dynamic update of models wherein the system computes and processes distributions of a training data set comprising labelled data and processes new data to determine a likelihood of new data with respect thereto and implementing training when the likelihood of new data exceeds a characteristic of the distribution such as by the accumulation of low likelihood data points), the method comprising:
receiving a first dataset representing model operations executed by the trained model at a first time, wherein the first dataset comprises at least one of user input data, vector data representative of the user input data, and model output data (Gop: ¶ 14, 28, 33-38: such as by segmenting, time windowing, etc. a stream of data into blocks thereof processed by a deployed model; each block comprising a dataset which is vectorized into a feature space and includes user chat, call, etc. data);
applying a data processing operation to the first dataset to determine a first result data based on the first dataset (Gop: ¶ 32, 33: system determines first result data with respect to a first block, dataset, etc. by computation of a distance between a data distribution over the received block and training data);
receiving a second dataset representing model operations executed by the trained model at a second time (Gop: ¶ 32, 33: system determines second , etc. data by repeating, iterating, etc. over successive blocks, windows, etc.; alternatively the training data may comprise a second dataset);
applying a data processing operation to the second dataset to determine a second result data based on the second dataset (Gop: ¶ 32-34: the same distribution operation applied to the second dataset);
determining, based on the first result data, a difference between the first result data and the second result data (Gop: ¶ 32-34: system determines a distance between the data distribution of the new data, blocked, windowed, etc. segments thereof and that of the training data);
determining, based on the difference, whether concept drift has occurred, wherein determining whether concept drift has occurred comprises determining that a characteristic of a distribution of user input data within a feature space of the trained model has changed (Gop: ¶ 12, 33, 38, 47: system iteratively determines which the distribution of new data is no longer a good fit to the model by analysis of successive changes to the input data with respect to distances determined in the feature space of the trained model); and
in accordance with a determination that concept drift has occurred, transmitting an instruction to update training of the trained model (Gop: ¶ 12, 42, 47: system sets a flag, issues an instruction to a network element, etc. to trigger retaining of the model); and
retraining the trained model (Gop: ¶ 11, 38: model automatically retrained when a counter maintaining drift parameters is incremented beyond a threshold), wherein retraining the trained model comprises:
determining one or more labels associated with the training data and updating the trained model based at least in part on a labeled training dataset (Gop: ¶ 14, 27, 31, 38, 39: system retrains using a transiting set comprising labelled data such as by supervised machine learning).
Gop does not explicitly teach determining, based on the difference, whether concept drift has occurred, wherein determining whether concept drift has occurred comprises determining that a characteristic of a distribution of user input data within the intent classification space of the trained model has changed; and wherein retraining the model comprises: applying one or more labels to data in the first dataset, the one or more labels associated with an intent classification of the data and comprising at least one new intent classification; and updating the trained model based at least in part on the labeled first dataset.
In a related field of endeavor Hua teaches a system and method for adapting and retraining a deployed dialog system model operative for intent classification based on determination of emergent intents which mandate a model update (Hua: Abstract; ¶ 3, 21) comprising tracking differences between input features against a stored prior population thereof (Hua: ¶ 3, 21, 35, 38, 53) the system operative to receive user input data (Hua: ¶ 38, 53; Fig 7: system receives an input question extracts intent and entity data therefrom) and determines an intent/entity vector thereof mapped within an intent classification space (Hua: ¶ 27, 35, 44, 46: Fig 2A, 3: extracted intent/entity pairs concatenated into a feature vector which constitutes an input space of a trained prediction model wherein the dimensionality expands when augmented with new intents) and updating a model based on the input which does not conform to the feature space (Hua: ¶ 46, 53, 56; Fig 4, 7: new features vectorized, stored in a feature database and the emergence of new intent/entity pairs triggers retraining based on normalization of the new pairs) based on applying one or more labels associated with an intent classification of the data (Hua: ¶ 38, 42, 43: each user input received and parsed to generate intent and entity labels, classes, etc.) and additionally comprising at least one new intent classification (Hua: ¶ 21, 44, 57: determinization that a user input does not conform to previously observed, classified data; intents thereof the intent determined and the newly determined intent added to the input feature space) and the dialog model, paths thereof, etc. updated based thereon to the feature vector dataset (Hua: ¶ 21, 44, 57: system discovers new intents, updates a dialog path based thereon). It would have been obvious to one of ordinary skill in the art before the effective filing date of the instant application to perform the drift determination and retraining as taught or suggested by Gop based on the determining of new user input data within the intent classification space of Hua thereby labelling user input with intent classification such that the determination of new intents changes a distribution of the input value space and thereby detect when a deployed models’ intent classification(s) no longer appropriately categorize incoming user queries and for at least the purpose of dynamically updating retraining, etc. a model based comparisons within a feature space thereof; one of ordinary skill in the art would have expected only predictable results therefrom.
Gop in view of Hua strongly suggests determining that a characteristic of a distribution of user input data based on intent classification has changed as the distribution is Gop is computed over data features, such as of input data and Hua evaluates input features against stored feature vectors; Gop in view of Hua does not explicitly discuss an intent classification space within/upon which to reify the distribution with respect to the characteristics of a distribution associated with user input data within the intent classification space.
In a related field of endeavor Woh teaches a system and method for processing user input to determine intent by classifying intent of a user input and determining embedding vector thereof mapping the represented message intent into a vector space such that each determined intent is represented by a prototype vector (Woh: Abstract) said user input collected over a time period and clustered with that space and characterized by a centroid, density, and/or variance characteristics (Woh: ¶ 105, 109) wherein a change is determined when such characteristics diverge beyond a threshold from models baseline characteristics, existing intent prototypes thereof; upon such a determination a new intent classification is created from the messages of the diverging cluster (Who: ¶ 61, 107-110, 112). It would have been obvious to one of ordinary skill in the art before the effective filing date of the instant application to the drift of the Gop in view of Hua system and method for a changed characteristic of a distribution of user input data whing the intent classification space taught or suggested by Woh for at least the purpose of identifying intents used by users which the deployed model does not recognized; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 2
Gop in view of Hua in view Woh of teaches or suggests:
The method of claim 1, wherein the data processing operation comprises a statistical analysis operation and wherein the first result data comprises a first statistical output (Gop: ¶ 32, 33: such as by computation of distributions based on Gaussian, Laplacian, etc. methods); (Woh: ¶ 108-110: such as using distances, densities, standard deviations, etc.). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 7
Gop in view of Hua in view Woh of teaches or suggests:
The method of claim 1 wherein the user input data comprises natural language data. (Gop: ¶ 3, 10, 21: such as within incoming text data); (Hua: ¶ 20, 37: such as incoming chat, dialog, etc. data); (Woh: ¶ 25: such as user messages with determinable intent). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 8
Gop in view of Hua in view Woh of teaches or suggests:
The system of claim 1 wherein the model output data comprises an intent classification generated based on user input data of the first dataset (Hua: ¶ 20, 37, 44, 48, etc.: such as for classifying incoming textual data); (Woh: Abstract: such as by selecting an intent with a prototype vector closest to the message embedding). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 9
Gop in view of Hua in view Woh of teaches or suggests:
The method of claim 1, wherein receiving the first dataset comprises receiving the first dataset from a shared memory in communication with the trained model (Gop: ¶ 26: exchange of messages on a system in communication with an active model) (Hua: ¶ 31, 46: features stored in memory accessible to an orchestrator and chatbots); (Woh: ¶ 58: messages in a prototype data store). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 10
Gop in view of Hua in view Woh of teaches or suggests:
The method of claim 1, wherein the first dataset comprises a predetermined amount of data (Gop: ¶ 33: such as by utilizing data with respect to a determined time window of regularly sized blocks). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 11
Gop in view of Hua in view Woh of teaches or suggests:
The method of claim 1, wherein the first dataset comprises data received during a predetermined period of time (Gop: ¶ 33: such as by utilizing data with respect to a determined time window); (Woh: ¶ 104: such as a day, week, etc.). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 13
Gop in view of Hua in view Woh of teaches or suggests:
The method of claim 1, wherein the second dataset comprises a training dataset that was used to train the trained model (Gop: Abstract; ¶ 12, 27: such as by retraining the first, second, etc. dataset); (Hua: ¶ 3, 21, 35, 44, 46, 56; Fig 3: such as by iteratively improving the first dataset base thereon). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 14
Gop in view of Hua in view Woh of teaches or suggests:
The method of claim 12, wherein the second dataset comprises a dataset having a similar distribution to a training dataset that was used to train the trained model (Gop: Abstract; ¶ 12, 27, 32-34: model updates based on new data and thus produces an amended training set with a similar distribution); (Hua: ¶ 3, 21, 35, 44, 46, 56; Fig 3: model updates with respect to emergent, new, etc. data thus producing a similarly distributed dataset save for the inclusion of the new intent, entity, etc.). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 15
Gop in view of Hua in view Woh of teaches or suggests:
The method of claim 12, wherein the second dataset comprises an inference dataset processed by the trained model at a different time than the first dataset (Gop: Abstract; ¶ 12, 27, 32, 33, 37: system practices inference based on a sequence of trained, retrained, etc. models at subsequent times). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 16
Gop in view of Hua in view Woh teaches or suggests:
The method of claim 1, wherein the difference comprises a difference between a first characteristic of the first result data and a second characteristic of the second result data (Gop: ¶ 33: system computes the distance, integral of the distance between two distributions). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 17
Gop in view of Hua in view Woh teaches or suggests:
The method of claim 1, wherein the difference comprises a difference between a first determined number of clusters of the first dataset and a second determined number of clusters in a second dataset (Gop: ¶ 11, 13, 29-31: system operates over various clustering algorithms to determine features with respect to a model which is retrained based on emergently determined features); (Woh: ¶ 105, 113-119: system clusters data message embeddings; two clusters mapping to one prototype are split; a cluster which maps to no extant prototypes generates a new intent; clusters are merged with, removed from other clusters; this is considered to teach comparisons of cluster counts, determination of differences therebetween. The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 18
Gop in view of Hua in view Woh teaches or suggests:
The method of claim 1, wherein determining whether concept drift has occurred comprises determining whether the difference is greater than a threshold value (Gop: Abstract; ¶ 12, 35.: difference of detected drift with respect to predetermined difference threshold iteratively updates the model). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 19
Gop in view of Hua in view Woh teaches or suggests:
The method of claim 1, wherein transmitting the instruction to update training of the trained model comprises transmitting executable program code configured to cause the trained model to be retrained when the code is executed by one or more processor (Gop: Abstract; ¶ 11, 12, 38, 51: system sets a flag to generate an instruction to automatically retrain the model). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 20
Gop in view of Hua in view Woh teaches or suggests:
The method of claim 1, wherein the instruction comprises an indication of the character or magnitude of the detected concept drift (Gop: Abstract; ¶ 15, 40, 41: a count of data drift maintained resolved against a threshold; system further maintains an ordered list of features by relevance score, change in relevance score, and features impactful upon divergence). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 22
Gop in view of Hua in view Woh teaches or suggests:
The method of claim 21, wherein the one or more labels applied to the data are determined based on a spatial clustering analysis of the first dataset (Gop: ¶ 11, 13, 29-31: system operates over various clustering algorithms to determine features with respect to a model which is retrained based on emergently determined features); (Wo h: ¶ 105, 112, 127: system determines intent labels based on hierarchical, centroid based, kmeans, etc. clustering, a prototype vector for a new intent may be created from messages of a cluster; a user may select messages to generate a new intent, such as based thereon). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claim 23
Gop in view of Hua in view Woh teaches or suggests:
The method of claim 1, wherein the trained model is a chatbot (Hua: ¶ 2, 5, 6, etc. : model operative with respect to a chatbot). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom.
Regarding claims 24, 25 – the claims are considered to recite substantially similar subject matter to that of claim 1 and are similarly rejected.
Claims 3, 4 rejected under 35 U.S.C. 103 as being unpatentable over Gopalan: 20180285772 hereinafter Gop further in view of Huang: 20210141862 hereinafter Hua and further in view of Wohlwend: 20200151253 hereinafter Woh as applied to claims 1, 2, 7-11, 13-20, 22-25 supra and further in view of Khatami: 20210406726 hereinafter Kha
Regarding claim 3
Gop in view of Hua in view Woh teaches or suggests:
The method of claim 2, wherein the statistical analysis operation comprises one or more of the following: a Kolmogorov-Smirnov (KS) test (Kha: ¶ 6, 32, 37-39, 44, 56; fig 1, 2, 7: system compares detected first, second MSE values such as based on a KD approach), a maximum mean discrepancy (MMD) test, a least-squares density difference (LSDD) test, a KMeans and chi square test, an equal intensity KMeans (EIKMeans) and chi square test, a Jensen-Shannon (JS) divergence test, and an uncertainty classifier. It would have been obvious to one of ordinary skill in the art before the effective filing date of the instant application to implement the statistical analysis operation of Gop in view of Hua in view of Woh as the KS test taught or suggested by Kha and for at least the purpose of testing whether two distributions differ sufficiently to warrant update; one of ordinary skill in the art would have expected only predictable results therefrom.
Examiner had taken official notice of the well-known nature of such algorithms in performing the types of statistical tests claims which Applicant had failed to timely and specifically traverse in the response to the NF action filed by Applicant 8/8/25 and the well-known nature was accepted as Admitted Prior Art in the final action of 10/8/25. In the arguments accompanying the RCE filed 2/9/26 Applicant traversed, arguing against the well-known nature of the claimed features. Examiner duly provided instant and unquestionable demonstration of the well-known nature of the methods recited to provide similar functionality, however Kha teaches the necessary one of the recited one or more statistical analysis operations.
Regarding claim 4
Gop in view of Hua in view Woh of teaches or suggests:
The method of claim 1, wherein the data processing operation comprises a model error rate determination analysis and wherein the first result data comprises a first error rate (Kha: ¶ 31-39: system maintains error distribution over time to determine changes in time series data and a need for update of a model upon which system invokes the update). The claim is considered obvious over Gop as modified by Hua and Woh as addressed in the base claim as it would have been obvious to apply the further teaching of Gop, Hua, and/or Woh to the modified device of Gop, Hua, and Woh; one of ordinary skill in the art would have expected only predictable results therefrom. It would have been obvious to one of ordinary skill in the art before the effective filing date of the instant application to implement the data processing operation of Gop in view of Hua in view of Woh as the model error rate determination taught or suggested by Kha for at least the purpose determining when the model can no longer accurately predict the data pattern; one of ordinary skill in the art would have expected only predictable results therefrom.
Claims 5 rejected under 35 U.S.C. 103 as being unpatentable over Gopalan: 20180285772 hereinafter Gop further in view of Huang: 20210141862 hereinafter Hua and further in view of Wohlwend: 20200151253 hereinafter Who as applied to claims 1, 2, 7-11, 13-20, 22-25 supra and further in view of Rafael de Lima Cabral, “Concept drift detection based on Fisher’s Exact test,” (copy provided by Examiner, published 2018 and hereinafter Cab).
Regarding claim 5
Gop in view of Hua in view Woh teaches or suggests:
The method of claim 4, wherein the model error rate determination analysis comprises one or more of the following: Fisher’s test, and a statistical test of equal proportions (STEPD). Examiner had taken official notice of the well-known nature of such algorithms in performing the types of statistical tests claims which Applicant had failed to timely and specifically traverse in the response to the NF action filed by Applicant 8/8/25 and the well-known nature was accepted as Admitted Prior Art (APA: please see MPEP 2144.03) in the final action of 10/8/25. In the arguments accompanying the RCE filed 2/9/26 Applicant traversed, arguing against the well-known nature of the claimed features. Examiner duly provided instant and unquestionable demonstration. Cab teaches that Fisher’s text and a statistical test of equal proportions comprise well-known algorithms by which to detect concept drift (Cab: Abstract, etc.). It would have been obvious to one of ordinary skill in the art before the effective filing date of the instant application to utilize well known tests such as the Cab taught or suggested Fisher’s and STEPD for computation of drift such as by quantification of errors, error rate, etc. to thereby improve the Gop in view of Hua in view of Woh drift detection system and method; one of ordinary skill in the art would have expected only predictable results therefrom. Please see additionally “Boschloo’s Test” Wikipedia page provided by Examiner and available at least 1/28/22 which conflates Fisher’s test with a statistical determination of equal proportion.
Response to Arguments
Applicant’s arguments in concert with amendments to the claims, see Remarks and Claims, filed 6/25/26, with respect to the rejection(s) of claim(s) 1-5, 7-11, 13-20, 22-25 under 35 USC 103 over Khatami in view of Huang in view of Gopalan, and/or Khatami in view of Huang in view of Gopalan in view of Ackerman, and/or Khatami in view of Huang in view of Gopalan in view of Rafael de Lima Cabral, and/or Khatami in view of Huang in view of Gopalan in view of Bjorelind have been fully considered and are persuasive. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground(s) of rejection is made in view of Gopalan in view of Huang in view of Wohlwend, and/or Gopalan in view of Huang in view of Wohlwend in view of Khatami, and/or Gopalan in view of Huang in view of Wohlwend in view of Rafael de Lima Cabral as addressed above.
Applicant’s remarks repeatedly state that Examiner “agreed” that the proposed amendment “would obviate the pending rejection.” Examiner indeed stated that the proposed amendment “appeared to obviate,” and agreed to conduct further search and consideration as mandated by the amendment and in keeping with MPEP 713. The results of the further search and consideration appear supra as the newly formed 35 USC 112, second paragraph rejection and the rejection of the claims over Gopalan in view of Huang in view of Wohlwend, and/or Gopalan in view of Huang in view of Wohlwend in view of Khatami, and/or Gopalan in view of Huang in view of Wohlwend in view of Rafael de Lima Cabral.
The amendments mandated the addition of Wohlwend which teaches a vector space in which user message embeddings and intent prototype vectors co-reside and is therefore to susceptible to Applicant’s arguments that a feature space “can be independent of the intent classification space.”
Applicant sole substantive argument relevant to the instant art rejection supra is that “Gopalan merely describes computing a distribution over the features of the training dataset,” and argues that such features “can be independent of the intent classification space.” This is not found persuasive for the reasons shown in the art rejection supra. Further the argument attacks Gopalan individually—Gopalan is not relied upon for the intent classification space. Additionally, Applicant tacitly concedes that the feature space of Gopalan may be the intent classifier space as Gopalan in ¶ 11 applies expressly to various types of machine learning algorithms, models, etc. as well as in ¶ 27 to classifier type models.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). 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 PAUL C MCCORD whose telephone number is (571)270-3701. The examiner can normally be reached 730-630 M-F.
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, CAROLYN EDWARDS can be reached at (571) 270-7136. 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.
/PAUL C MCCORD/Primary Examiner, Art Unit 2692
Prosecution Insights