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 .
This action is made non-final.
Claims 1-20 are pending. Claims 1, 10 and 19 are independent claims.
Response to Arguments
Applicant's arguments filed 11/14/2025, regarding the 35 U.S.C. 101 rejections of the previous office action have been fully considered and are persuasive. The 35 U.S.C. 101 rejections have been withdrawn.
Applicant’s arguments filed 11/14/2025, regarding the 35 U.S.C. 103 rejections of the previous office action have been fully considered. Due to the claim amendments, the scope of the claims has changed and new grounds of rejection are applied – see the updated rejection below.
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, 2, 8, 9-11 and 17-20 are rejected under 35 U.S.C. 103 as being unpatentable over Berrebbi et al. (US 20190122164 A1), herein Berrebbi, in view of Turkelson et al. (US 20210004589 A1), herein Turkelson, Griffin et al. (US 20210390587 A1), herein Griffin, and Ilic et al. (US 20200211220 A1), herein Ilic.
Regarding claim 1, Berrebbi teaches: A method comprising: monitoring user interactions with a supply chain system based on a tracked ontology enrichment process (¶30, The ontological graph 173 can further connect restaurants with semantic aspects of the menu items, and can incorporate and consolidate data sets from multiple sources, such as historical data); constructing an explainable reasoning graph based on the monitored user interactions and domain specific reasoning information (¶31, Furthermore, all descriptive terms of all possible comestible items can be organized graphically, with each descriptive term comprising a point in the ontological graph); learning an explainable insight of the monitored user interactions (¶33, to generate a personal preference word vector for the given user); learning a user interaction embedding for an embedding space based on the constructed explainable reasoning graph and the explainable insight (¶33, For example, the profiling engine 140 can generate the personal preference word vector for the given user 197 on the word corpus of descriptive menu item terms); incorporating external data into the embedding space; learning a joint embedding based on the user interaction embedding (¶29, Furthermore, in certain aspects, each of the descriptive terms may be represented by a vector in latent space)… controlling a supply chain based on the revised ontology (¶13, An on-demand delivery service (e.g., an on-demand food delivery service) can be managed by a network-based computing system by connecting requesting users, delivery drivers, and comestible item sources (e.g., restaurants) via designated applications specific to the on-demand delivery service) the controlling comprising dynamically modifying one or more operational parameters of the supply chain (¶13, In various examples, the network computer system acts as a logistical optimizer that preempts menu item demand by coordinating delivery of pre-prepared food items in delivery vehicles and transporting the items to requesters on-demand in response to received item requests – dynamically modifying food inventory levels, i.e., an operational parameter).
Berrebbi fails to teach: identifying missing entities and relationships for incorporation into an ontology based on the user interactions and joint embedding; revising the ontology to incorporate the missing entities and relationships into the ontology to create a revised ontology.
However, in the same field of endeavor, Turkelson teaches: identifying missing entities and relationships for incorporation into an ontology based on the user interactions and joint embedding; revising the ontology to incorporate the missing entities and relationships into the ontology to create a revised ontology (¶292, As another example, a density of clusters of embedding vectors produced by the network may be analyzed. If the density includes large gaps, then this may indicate that additional data should be added to the training data for the model to fill in the missing gaps).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to add missing entities to the ontology as disclosed by Turkelson in the supply chain system disclosed by Berrebbi in order to increase model accuracy (¶292, to improve the model's accuracy).
Berrebbi in view of Turkelson fails to teach: by training a twin neural network… to align the user interaction embedding and the external data embedding in a common latent space.
However, in the same field of endeavor, Griffin teaches: by training a twin neural network… to align the user interaction embedding and the external data embedding in a common latent space (¶57, At step 310, a conventional Siamese network (also referred to as “twin neural network”) can be used to generate a joint embedding matrix across users and advertisements… FIG. 4 illustrates an exemplary Siamese network for joint embedding, according to embodiments described herein).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use a twin neural network to align embeddings in the supply chain system disclosed by Berrebbi in view of Turkelson in order to represent relationships between previously unconnected data (¶57, so as to determine a relationship between users and advertisements (i.e., an interested level of a user relative to an advertisement… if a user is interested in an advertisement, a joint embedding between the advertisement and the user lies as near-by points; otherwise, if a user is not interested in an advertisement, a joint embedding between the advertisement and the user lies as far-away points).
Berrebbi in view of Turkelson and Griffin fails to teach: using a hinge loss function.
However, in the same field of endeavor, Ilic teaches: using a hinge loss function (¶52, In order to learn the mapping, a so-called Siamese network is used, which takes two inputs instead of one, and a specific cost function. The cost function is defined in such a way that, for similar objects, the square of the Euclidean distance between them is minimized and for dissimilar objects, the hinge loss function is used, which forces the objects apart by means of a difference term).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use a hinge loss function in a twin neural network as disclosed by Ilic in the supply chain system disclosed by Berrebbi in view of Turkelson and Griffin in order to ensure model accuracy upon training with existing testing data (¶240, indicate that a training data set includes enough data for a model to product accurate results).
Regarding claim 2, Berrebbi in view of Griffin and Ilic fails to teach: the method of claim 1, further comprising: triggering data collection based on the identified missing entities; generating a model health score; and retraining a forecasting model based on the model health score.
However, in the same field of endeavor, Turkelson teaches: the method of claim 1, further comprising: triggering data collection based on the identified missing entities (¶292, In particular, based on the location of these gaps in the embedding space, a determination may be made as to what data should be obtained (e.g., added to the training data)); generating a model health score; and retraining a forecasting model based on the model health score (¶240, A threshold for determining whether a model includes enough training data may be if the MCC score may be a value selected from a range of values between 0.1 and 0.9 – and in response to adding new data the model is retrained – ¶201, Therefore, upon retraining the computer-vision object recognition model, parameters of the model may be enriched).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to add previously missing data and use a score to determine whether or not to retrain the model as disclosed by Turkelson in the supply chain system disclosed by Berrebbi in view of Griffin and Ilic in order to ensure model accuracy upon training with existing testing data (¶240, indicate that a training data set includes enough data for a model to product accurate results).
Regarding claim 8, Berrebbi further teaches: the method of claim 1, further comprising: learning a vector representation for the user interactions with explainable insights using an attributed graph embedding (¶33, to generate a personal preference word vector for the given user).
Berrebbi in view of Griffin and Ilic fails to teach: triggering a notification to a user for verification of auto-generated constraints and the missing entities; and triggering retraining of forecasting models based on a supply chain forecasting pipeline health score.
However, in the same field of endeavor, Turkelson teaches: triggering a notification to a user for verification of auto-generated constraints and the missing entities (¶304, In some embodiments, the instruction may automatically cause the video recording process to end, however alternatively the instruction may provide a notification to the user to manually cause the video recording process to end – the video recording is used to train the model in Turkelson); and triggering retraining of forecasting models based on a supply chain forecasting pipeline health score (¶240, A threshold for determining whether a model includes enough training data may be if the MCC score may be a value selected from a range of values between 0.1 and 0.9).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the score-based triggered data collection disclosed by Turkelson with the model disclosed by Berrebbi in view of Griffin and Ilic in order to ensure sufficient model accuracy when training with existing testing data (¶240, indicate that a training data set includes enough data for a model to product accurate results).
Regarding claim 9, Berrebbi further teaches: the method of claim 1, wherein controlling the supply chain comprises taking at least one physical action with respect to the supply chain (¶39, In various examples, the menu item request can include location information indicating the current location of the user 197, or the user 197 can input a rendezvous location to meet with a delivery driver).
Regarding claim 10, it is an apparatus of claim 1 and is rejected on the same grounds as presented above.
Regarding claim 11, it is an apparatus of claim 2 and is rejected on the same grounds as presented above.
Regarding claim 17, it is an apparatus of claim 8 and is rejected on the same grounds as presented above.
Regarding claim 18, it is an apparatus of claim 9 and is rejected on the same grounds as presented above.
Regarding claim 19, it is a computer program product of claim 1 and is rejected on the same grounds as presented above.
Regarding claim 20, it is a computer program product of claim 2 and is rejected on the same grounds as presented above.
Claims 6 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Berrebbi in view of Turkelson, Griffin and Ilic as applied to claim 1 above, and further in view of Saylor et al. (US 20190228372 A1), herein Saylor.
Regarding claim 6, Berrebbi teaches: the method of claim 1, further comprising: optimizing a generic forecasting model based on a spatial-temporal characteristic (¶13, In doing so, the network computer system can implement deep learning techniques to forecast item demand for individual food items (e.g., based on historical data and a variety of other factors, such as weather, time of week, time of day, etc.), provide real-time recommendations based on contextual user information (e.g., search inputs, location, historical preferences)). Since the supply chain system disclosed by Berrebbi is able to provide real-time recommendations, it must be performing some sort of periodic re-evaluation that takes into account the current spatial-temporal dimension.
Berrebbi in view of Griffin and Ilic fails to teach: periodically evaluating a performance of the optimized forecasting model in different spatial-temporal dimensions and estimating a model health score; triggering data collection based on the model health score and a corresponding budget; and forecasting a pipeline evaluation.
However, in the same field of endeavor, Saylor teaches: periodically evaluating a performance of the optimized forecasting model in different spatial-temporal dimensions (¶155, Cycle Check 936 sets the amount of simulation time that will pass between requisition manager evaluations of the pipeline) and forecasting a pipeline evaluation (¶57, The supply model accounts for various actions/behaviors that impact the accuracy of a forecasted supply model).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to add the periodic evaluation disclosed by Saylor to the supply chain system disclosed by Berrebbi in view of Turkelson, Griffin and Ilic in order to make it easier to manage (¶201, reduces the time and costs required to develop supply chain network simulations).
Berrebbi in view of Saylor fails to teach: estimating a model health score and triggering data collection based on the model health score and a corresponding budget.
However, in the same field of endeavor, Turkelson teaches: estimating a model health score (¶240, A threshold for determining whether a model includes enough training data may be if the MCC score may be a value selected from a range of values between 0.1 and 0.9) and triggering data collection based on the model health score and a corresponding budget (¶240, A threshold for determining whether a model includes enough training data may be if the MCC score may be a value selected from a range of values between 0.1 and 0.9 – budget considerations are discussed in Berrebbi, ¶37, In various implementations, the profiling engine 140 can determine a price sensitivity metric for each user preference profile). Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the score-based triggered data collection disclosed by Turkelson with the supply chain system disclosed by Berrebbi in view of Griffin, Ilic, and Saylor in order to ensure sufficient model accuracy when training with existing testing data (¶240, indicate that a training data set includes enough data for a model to product accurate results).
Regarding claim 15, it is an apparatus of claim 6 and is rejected on the same grounds as presented above.
Claims 7 and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Berrebbi in view of Turkelson, Griffin and Ilic as applied to claim 1 above, and further in view of Koerner et al. (US 20160179877 A1), herein Koerner.
Regarding claim 7, Berrebbi in view of Turkelson, Griffin and Ilic fails to teach: the method of claim 1, further comprising: receiving a natural language query as an input to analyze an impact of climatic variations using explainable insights; identifying one or more additional entities and relationships with constraints by parsing the natural language query; issuing one or more questions to understand a user's query based on auto-generated explainable insights and curated knowledge in a form of the ontology; generating one or more explainable insights using an explainable model based on the identified constraints; and storing the generated explainable insights and user feedback. Berrebbi does teach analyzing the impact of climactic variation when they account for weather in their model (¶13, e.g., based on historical data and a variety of other factors, such as weather, time of week, time of day, etc.).
However, in the same field of endeavor, Koerner teaches: further comprising: receiving a natural language query as an input to analyze an impact of climatic variations using explainable insights (¶19, The business intelligence system 100 includes a query engine 120 that receives a query from the query box 104 of the user interface layer 102… The inputted text may be in the form of natural human language); identifying one or more additional entities and relationships with constraints by parsing the natural language query (¶21, For example, the keyword parser 122 parses the natural language text of the query to determine one or more keywords); issuing one or more questions to understand a user's query based on auto-generated explainable insights and curated knowledge in a form of the ontology (¶25, In some examples, the query logic unit 126 may ask the user to explain the keyword for future mappings); generating one or more explainable insights using an explainable model based on the identified constraints; and storing the generated explainable insights and user feedback (¶25, Stated another way, the semantic layer 110 continues to learn new keyword mappings and becomes more intelligent as it receives more and more feedback from the end users).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to add the natural language query and insight learning disclosed by Koerner to the supply chain system disclosed by Berrebbi in view of Turkelson, Griffin and Ilic in order to allow the system to continuously improve itself (¶47, Stated another way, the semantic layer 110 continues to learn new keyword mappings and becomes more intelligent).
Regarding claim 16, it is an apparatus of claim 7 and is rejected on the same grounds as presented above.
Allowable Subject Matter
Claims 3, 4, 5, 12, 13 and 14 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Wang et al. (“L2 Mispronunciation Verification Based on Acoustic Phone Embedding and Siamese Networks”, 2018) discloses twin networks combined with a hinge loss.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to HARRISON CHAN YOUNG KIM whose telephone number is (571)272-0713. The examiner can normally be reached Monday - Thursday 9:00 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, Cesar Paula can be reached at (571) 272-4128. 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.
/HARRISON C KIM/ Examiner, Art Unit 2145
/CESAR B PAULA/ Supervisory Patent Examiner, Art Unit 2145