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 .
Notice to Applicant
The following is a Final Office action. In response to Examiner’s Non-Final Rejection of 2/24/26, Applicant, on 5/26/26, amended claims. Claims 1, 3-5-10, 12-17, 19-20 are pending in this application and have been rejected below.
Response to Amendment
Applicant’s amendments are acknowledged.
The 112b rejections are withdrawn in light of Applicant’s amendments.
The 101 rejections are withdrawn in light of Applicant’s amendments.
The double patenting rejection over 11,403,573 is withdrawn in light of the Terminal Disclaimer approved 5/26/26. The double patenting rejection over 12,165,090 still remains, as it was not included in the Terminal Disclaimer. Examiner suggests filing a terminal disclaimer for the remaining rejection as well.
With regards to 103, Examiner suggests considering incorporating limitations from parent cases, as discussed below, from Allowance in 17/849,185 and 14/729,444 pertaining to – Claim 1 in ‘185 – “model observed values of the at least two demand time series according to a hierarchical negative-binomial state space model” and [0024, 0048, 0054-0055, 0068-0070, 0079, 0096-0098] as published in this Application, and other places in the disclosure referring to these aspects.
Reasons for Overcoming the 101 Rejection
The amendment overcomes the 101 rejection because the claim is now : 1) define a plate notation model for modelling a supply chain comprising one or more supply chain entities, wherein the plate notation model comprises an inner plate that indicates a repetition of variables inside the inner plate; 2) perform inference using a Gaussian Markov Random Field and a sparse tridiagonal precision matrix over the single time series of unobserved supply chain model variables to generate local and global process parameters, a latent variable, and an effective latent variable, wherein the use of the sparse tridiagonal precision matrix increases a computing speed of computing a prior term in the single time series, wherein the local and global process parameters are represented in the outer area of the plate notation model; and 3) in response to a current time period elapsing, repeating the receiving historical sales data, performing inference, generating the forecasted latent variables, and generating the distributional demand forecast based on actual sales data corresponding to the elapsed current time period being added to the historical sales data. When viewing the claim as a whole, when combined with the earlier limitations, and based on Applicant’s specification [0077, 0080-0083] (see Remarks, 5/26/26, pages 13-14), is viewed as a practical application under step 2a, prong 2, as the claim is improving another technology when viewing all the limitations listed above (See MPEP 2106.05a) and/or is viewed as a using a judicial exception in a meaningful way under MPEP 2106.05(e). This is then consistent with Applicant’s arguments in 14/729,444, on 2/16/22 pages 10-12, that use of the Gaussian Markov Random field, and the sparse tridiagonal precision matrix, are used as tools in the domain of “artificial intelligence and machine learning,” where a feedback loop is created to improve the “inference” and final forecast with increased computing speed by formation and use of the particular data structure of the sparse tridiagonal precision matrix. This is viewed as similar to the example in MPEP 2106.05(a)(I) - Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, 1336, 118 USPQ2d 1684, 1689 (Fed. Cir. 2016) and how the invention improved the way the computer stores and retrieves data in memory in combination with the specific data structure recited in the claims the demonstrated eligibility; MPEP 2106.05(a)(II) - Examples that the courts have indicated may be sufficient to show an improvement in existing technology include: iv. Improved, particular method of digital data compression, DDR Holdings, LLC. v. Hotels.com, L.P., 773 F.3d 1245, 1259, 113 USPQ2d 1097, 1107 (Fed. Cir. 2014); Intellectual Ventures I v. Symantec Corp., 838 F.3d 1307, 1315, 120 USPQ2d 1353, 1358 (Fed. Cir. 2016); See also July 2024 Guidance Update on Patent Subject Matter Eligibility, Federal Register, Vol. 89, No. 137, pages 58137-58138 giving example of “Claims to performing error correction and detection encoding where the information bits appear in a variable number of subsets were directed to an improvement of encoding data that relies in part on irregular repetition and not an abstract idea,” Cal. Inst. of Tech v. Broadcom Ltd, 25 F.4th 976, 988 (Fed Cir. 2022) – since here there also is repetition with the in the particular data structure of the limitations to achieve the sparse tridiagonal precision matrix of unobserved variables; See also Abstract Idea Example 48, claim 2, where separating speech from different speech sources without prior knowledge or interaction from a target user in the manner claimed reflected the improvement in the disclosure of how the DNN (deep neural network) aids in the assignments in the mixed speech signal while excluding audio from an unidentified source. Here, a similar set of operations are performed to form a sparse tridiagonal precisions matrix from modelling unobserved variables without interaction from a user.
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.
This application currently names joint inventors. In considering patentability of the claims under pre-AIA 35 U.S.C. 103(a), the examiner presumes that the subject matter of the various claims was commonly owned at the time any inventions covered therein were made absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and invention dates of each claim that was not commonly owned at the time a later invention was made in order for the examiner to consider the applicability of pre-AIA 35 U.S.C. 103(c) and potential pre-AIA 35 U.S.C. 102(e), (f) or (g) prior art under pre-AIA 35 U.S.C. 103(a).
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 1-4, 6-11, 13-18, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Uhrig (US 2005/0075949) in view of Breuker, “Towards Model-Driven Engineering for Big Data Analytics – An Exploratory Analysis of Domain-Specific Languages for Machine Learning,” 2014, 47th Hawaii International Conference on System Sciences. IEEE, 2014, pages 758-767, and Salemi et al., (“Forecasting the intermittent demand for slow-moving inventories: A modelling approach,” 2014, Proceedings of the Winter Simulation Conference, pages 3809-3820.
Concerning claim 1, Uhrig discloses:
A system, comprising:
a computer comprising a memory and a processor and (Uhrig – par 9 – rules stored in computer memory; software analyzing and planning inventory; See also pa 122 – server is “Dell” processor, with RAM and cache; other software and hardware; See also par 121-122 utilizing servers and programming languages; par 120 - having executable program files; See par 7-14 - perform the method using computer software and memory ) configured to:
define a … model for modelling a supply chain comprising one or more supply chain entities (Applicant’s examples of entities include supplier, retailer, manufacturer, distribution center, online retailers, and/or customers (See FIG. 1, par 33 as filed) Uhrig discloses limitations –See par 31 - Data identified in the import map includes inventory item identifying information (e.g., Stock Keeping Unit or SKU) and the physical location of each item in the inventory [see par 45 – location can be a geographic location /warehouse]. Demand and strategy data are also identified in the import map. The demand data includes usage history for inventory items and future demand predictions for those items, as calculated using probability and statistics algorithms known in the art. Strategy data includes ordering and stocking data for inventory items and typically includes UDF fields in order to accurately represent the stocking logic for each item.; See par 42- Supplier data stored in the inventory fact table 300 includes supplier average lead time (the average lead time a supplier needs to fill an order), and supplier cost. Supplier data is generally used to assist inventory planners with determining how to replenish items that are being stocked; See par 97 - the collaborate process of block 92 can be used to obtain information or feedback on a stocking plan from suppliers of inventory items in a supply chain) wherein the plate notation model comprises an inner plate that indicates a repetition of variables inside the inner plate and an outer area outside the inner plate that indicates a non-repetition of variables in the outer area. (Breuker page 760-761 -
PNG
media_image1.png
592
436
media_image1.png
Greyscale
PNG
media_image2.png
478
410
media_image2.png
Greyscale
see page 760, Table 1 – construction of directed graphical models; includes “plate” that can be repeated N times; page 765, Col 1 – plates; FIG. 4 – showing plats relative to observed variables, random variables, factors, nodes, and more).
Uhrig discloses:
receive historical sales data of the one or more supply chain entities comprising a single time series (Uhrig – See par 31 - Demand and strategy data are also identified in the import map. The demand data includes usage history for inventory items and future demand predictions for those items, as calculated using probability and statistics algorithms known in the art; see par 91 - The generate actions process of block 90 identifies new items that need to be added to the inventory and existing items that need to be deleted from the inventory in order to achieve the new stocking plan; updating the database 14 to reallocate items to another location in the supply chain (i.e. each location constitutes a different demand series); See par 93 - The inventory data, including current demand forecasts, for the Atlanta location is imported into the database 14 for each of 12 months. and explanatory variables, wherein the explanatory variables are represented in the inner plate of the plate notation model (1) There is no special definition for “explanatory variables” in the specification Applicant’s specification in par 24 states that explanatory variables can include promotions, seasonality, special events (such as sporting events), weather). Paragraph 24 as filed gives examples of “explanatory variables” as it states that “input variables, or explanatory variables, which may include, but are not limited to, indicators and data of promotions, seasonality, special events (such as sporting events), weather, and the like; Uhrig – discloses the limitations based on broadest reasonable interpretation in light of the specification – See par 73 - Demand forecast information is used to adjust stocking plans for seasonal changes in demand and other demand spikes;
see also Breuker – see page 760- Table 1 – Constructs of directed graphical models – Parameter – “defines a parameter of the model”; see page 764, col. 2, last paragraph - Both variables and factors are subsumed in the abstract entity type Node which can be given a textual description (the name attribute).).
Uhrig discloses accounting for a slow moving item (See par 71, 120); adjusting stocking plans for seasonal changes (see par 73). However, Uhrig does not disclose:
“perform inference using a Gaussian Markov Random Field and a sparse tridiagonal precision matrix over the single time series of unobserved supply chain model variables to generate local and global process parameters, a latent variable, and an effective latent variable, wherein the use of the sparse tridiagonal precision matrix increases a computing speed of computing a prior term in the single time series.”
Breuker discloses performing inference for the variables:
“perform inference … over the single time series of unobserved supply chain model variables to generate local and global process parameters, a latent variable, and an effective latent variable… (Breuker - see page 762, Section 2.3 - Given a set of latent random variables x1 and a set of observed random variables x0, probabilistic inference can be described as the problem of computing a distribution P(x1| x0) over the latent variables given the observed ones (same as Applicant’s claim and FIG. 3); page 765, col. 1, 2nd paragraph - the factor can lie in a plate in which the variable is not (more factors than variables). This can be resolved by feeding the variable into all the factors (disclosing a global parameter); In any case, if only a specific variable or factor should be used and not all, a selector variable must be defined. The ternary relationship selects codifies this (disclosing a “local” parameter); See page 765, Col. 2, 1st paragraph – when using the model, the purpose is to infer the posterior distributions conditioned on all observations to make predictions for new data.)
Breuker discloses that Markov can be for approximating distributions (page 762, col. 2) and that one wants to “speed up computations” if thousands of variables (See page 760, 1st-2nd paragraphs).
Salemi discloses:
perform inference “using a Gaussian Markov Random Field and a sparse tridiagonal precision matrix” over the single time series of unobserved supply chain model variables to generate local and global process parameters, a latent variable, and an effective latent variable, “wherein the use of the sparse tridiagonal precision matrix increases a computing speed of computing a prior term in the single time series.” (Applicant’s paragraph 47 is that “forecast data 228 comprises data resulting from the inference process.” Applicant’s specification par 94 – The effective latent variable 720 4e,t is conditioned not only by the latent variable 730, but also by the explanatory variables x ,t750 which are known for each historical and future time period and location. Plate-notation model 700 receives the explanatory variables 224 and uses future explanatory variables factors (such as known future promotions, sales, special events, and/or weather, including weather forecasts) to calculate unknown variables. Explanatory variables x ,t750 may be further combined with local coefficients 760 6f, which is the response to the explanatory variables for the -e-th time series. This response may take the form of a lift or a drop to the effective latent variable 720 at each time and location);
Salemi – see page 3810, 2nd paragraph - The Markov structure of GMRFs (Gaussian Markov Random Field) is intuitive for problems in industrial engineering and operations research (Salemi, Staum, and Nelson 2013). For example, if we were interested in predicting the value of the objective function at a feasible prediction point, then the values of the objective function at the feasible points in its neighborhood would typically be sufficient; others would provide very little information. GMRFs are typically defined on lattices, so the use of GMRFs in DOvS problems is more natural than using a GRF with a continuous domain. Most importantly, the Markov structure lends itself to more efficient and numerically stable calculations. A GMRF is defined by its precision matrix, which is the inverse of the covariance matrix. Using the Markov structure of GMRFs, the precision matrix of a GMRF can be constructed to be sparse. Thus, we can use several sparse matrix techniques to calculate expressions which involve the precision matrix; see page 3813 – Qe is a diagonal matrix; see page 3814, Section 3.1, last paragraph – recommend updating the parameters).
Uhrig, Breuker, and Salemi disclose:
wherein the plate notation model comprises an inner plate that indicates a repetition of variables inside the inner plate and an outer area outside the inner plate that indicates a non-repetition of variables in the outer area;
page 760-761 -
PNG
media_image1.png
592
436
media_image1.png
Greyscale
PNG
media_image2.png
478
410
media_image2.png
Greyscale
see page 760, Table 1 – construction of directed graphical models; includes “plate” that can be repeated N times; page 765, Col 1 – plates; FIG. 4 – showing plats relative to observed variables, random variables, factors, nodes, and more).
generate forecasted latent variables based, on the generated inferred local and global process parameters, latent variables, and effective latent variable (Breuker – see page 759, section 2.1 - Consider the simple example illustrated in Figure 1. The graph represents a probability distribution over a set of three random variables. The purpose of graphical models is to support a user in asking questions about random variables, very much in the way a database answers questions about the data it stores; see page 762, Section 2.3 - Given a set of latent random variables x1 and a set of observed random variables x0, probabilistic inference can be described as the problem of computing a distribution P(x1| x0) over the latent variables given the observed ones (same as Applicant’s claim and FIG. 3); see page 765, col. 2, 1st paragraph - When using the model, the purpose is to infer the posterior distributions conditioned on all observations to make predictions for new data. it must be indicated which observed variables (priors) are to be replaced with inferred posteriors. This is done using the Boolean attribute infer? of the entity type Observed Variable. After setting the posteriors, training data will be replaced with new data);
generate a distributional demand forecast (Uhrig – 2005/0075949 – See par 73 - Demand forecast information is used to adjust stocking plans for seasonal changes in demand and other demand spikes; See par 82 - The forecasting algorithms are executed on the inventory data to predict the demand for an inventory item at some time in the future; See par 96 - From analyzing the view of the inventory data, the user determines that certain items sell more slowly than others in the inventory. The user sets the forecast method for the new calculate cycle stock rule so that it includes periods in which no sales were made; See par 82 - Generate forecasts process 64 executes forecasting algorithms known in the art, including probability distributions for random variables) comprising probabilities of integer values of demand for one or more time steps into the future based, at least in part, on the and forecasted latent variables.” (Breuker – see page 759, Col. 2, Section 2.1 - Graphical models are a tool to describe probabilistic models visually; The graph represents a probability distribution P (A, B, C) over a set of three random variables {A, B, C}. page 761, col. 2 - It is called a gate and can be used to switch on and off different parts of a model depending on a (discrete) random variable. Each value of this random variable is associated with an area in the model. The entire model is a mixture of all areas weighted with the respective mixing probabilities. see page 762, Section 2.3 - Given a set of latent random variables x1 and a set of observed random variables x0, probabilistic inference can be described as the problem of computing a distribution P(x1| x0) over the latent variables given the observed ones (same as Applicant’s claim and FIG. 3));
in response to the generated distributional demand forecast (See Uhrig par 73 – demand forecast to adjust stocking plans; See par 82 - The forecasting algorithms are executed on the inventory data to predict the demand for an inventory item at some time in the future;), transform the supply chain inventory to a level based, at least in part, on the distributional demand forecast (Uhrig – 2005/0075949 – See par 73 - Demand forecast information is used to adjust stocking plans for seasonal changes in demand and other demand spikes.) .
Uhrig, Breuker, and Salemi are analogous art as they are directed to assessing forecasts/predictions (Uhrig par 39, 73; Breuker Abstract; page 759, Section 2.1 – probability distribution for variables; page 761, col. 2 – predicting; Salemi Abstract, page 3810 – optimization in operations research). 1) Uhrig discloses having modeling of multiple locations and probabilities for demand predictions (See par 31) and supplier data to assist inventory planners (See par 42). Uhrig discloses that accounting for a slow moving item (See par 71, 120); adjusting stocking plans for seasonal changes (see par 73). Breuker improves upon Uhrig by using a plate notation that can be repeated N times for performing the mathematical calculations (See page 760, 765, FIG. 4) and inferring distributions based on observations to make predictions for new data while considering latent variables and random variables (See FIG. 4; page 762, 765). One of ordinary skill in the art would be motivated to further include the known plate rotation model and inferring for unobserved variables. 2) Uhrig discloses that accounting for a slow moving item (See par 71, 120); adjusting stocking plans for seasonal changes (see par 73). Breuker discloses that Markov can be for approximating distributions (page 762, col. 2) and that one wants to “speed up computations” if thousands of variables (See page 760, 1st-2nd paragraphs). Salemi improves upon Uhrig and Breuker by using a Gaussian Markov Random Field and a sparse precision matrix for a more efficient calculation (See page 3810, 3814). One of ordinary skill in the art would be motivated to further include Gaussian Markov Random Field and a sparse precision matrix for a more efficient calculation to efficiently adjust the stocking plans in Uhrig and further approximate many distributions and apply the known Markovian math in Breuker.
Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system and method of adjusting stocking plans based on seasonal changes and “other” demand spikes in Uhrig to further use plate notation for repeating calculations, inferring, considering latent variables and random variables as disclosed in Breuker to further use a Gaussian Markov Random Field and a spare precision matrix for a more efficient calculation as disclosed in Salemi, since the claimed invention is merely a combination of old elements, and in combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art would have recognized that the results of the combination were predictable.
Concerning independent claim 8, Uhrig discloses:
A computer implement method, comprising:
defining, by a computer comprising a memory and a processor (Uhrig par 9 – rules stored in computer memory; software analyzing and planning inventory; See also pa 122 – server is “Dell” processor, with RAM and cache; other software and hardware; See also par 121-122 utilizing servers and programming languages; par 120 - having executable program files; See par 7-14 - perform the method using computer software and memory) …
The remaining limitations are similar to claim 1 above. Accordingly, claim 8 is rejected for the same reasons as stated above for claim 1.
It would have been obvious to combine Uhrig with Breuker and Salemi for the same reasons as discussed above with regards to claim 1.
Concerning independent claim 15, Uhrig discloses:
A non-transitory computer-readable medium embodied with software, the software when executed configured to (See Uhrig – par 9 – rules stored in computer memory; software analyzing and planning inventory; See also pa 122 – server is “Dell” processor, with RAM and cache; other software and hardware; See also par 121-122 utilizing servers and programming languages; par 120 – having executable program files; See par 7-14 - perform the method using computer software and memory).
The remaining limitations are similar to claim 1 above. Accordingly, claim 8 is
rejected for the same reasons as stated above for claim 1.
It would have been obvious to combine Uhrig with Breuker and Salemi for the same reasons as discussed above with regards to claim 1.
Concerning claims 2, 9, and 16, Breuker discloses:
The system of claim 1, wherein the plate notation model comprises an observed value variable, wherein the observed value variable is conditioned by the effective latent variable (Breuker –see page 760, Table 1 –
PNG
media_image3.png
366
450
media_image3.png
Greyscale
see page 761, FIG. 2 –
PNG
media_image4.png
476
406
media_image4.png
Greyscale
See also page 765, FIG. 4).
It would have been obvious to combine Uhrig with Breuker and Salemi for the same reasons as discussed above with regards to claim 1.
Concerning claims 3, 10, and 17, Breuker discloses:
The system of claim 2, wherein the effective latent variable is conditioned by the latent variable, an observed explanatory variable and a regression coefficient (Breuker – see page 761, FIG. 2 – “graphical model for polynomial regression”; col. 2, 3rd paragraph - set of training data consisting of N pairs of real variables (xn, yn) is given. The task is to predict x′ using a new data point _′. To explain the data, the goal is to fit a polynomial F(x, θ) of order k with coefficient vector θ = (θ1, … , θk)T.
It would have been obvious to combine Uhrig with Breuker and Salemi for the same reasons as discussed above with regards to claim 1.
Concerning claims 4, 11, and 18, Breuker discloses:
The system of claim 1,
It would have been obvious to combine Uhrig with Breuker and Salemi for the same reasons as discussed above with regards to claim 1.
Concerning claims 6, 13, and 20, Richard discloses:
The system of claim 1, wherein the computer is further configured to: linearly combine the explanatory variables (Uhrig discloses “ a particular group of rules configured in a certain way yields the best stocking plan for slow moving inventory. The user can create a "solution" specifically for slow moving inventory or other types of inventory” (See par 105) and “a characteristic of an inventory demand is whether a particular item has a previous demand history. Demand forecast information is used to adjust stocking plans for seasonal changes in demand and other demand spikes.” (See par 73).) through regression coefficients (Breuker – see page 761, col. 2 – regression used in FIG. 2; multiple parameters formalized).
It would have been obvious to combine Uhrig with Breuker and Salemi and Richard for the same reasons as discussed above with regards to claim 1.
Concerning claims 7 and 14 and 20, Breuker discloses:
The system of claim 6, wherein the combination of the explanatory variables additively shifts the latent variable (Uhrig discloses “ a particular group of rules configured in a certain way yields the best stocking plan for slow moving inventory. The user can create a "solution" specifically for slow moving inventory or other types of inventory” (See par 105) and “a characteristic of an inventory demand is whether a particular item has a previous demand history. Demand forecast information is used to adjust stocking plans for seasonal changes in demand and other demand spikes.” (See par 73).).
It would have been obvious to combine Uhrig with Breuker and Salemi for the same reasons as discussed above with regards to claim 1.
Claims 5, 12, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Uhrig (US 2005/0075949) in view of Breuker, “Towards Model-Driven Engineering for Big Data Analytics – An Exploratory Analysis of Domain-Specific Languages for Machine Learning,” 2014, 47th Hawaii International Conference on System Sciences. IEEE, 2014, pages 758-767, and Salemi et al., (“Forecasting the intermittent demand for slow-moving inventories: A modelling approach,” 2014, Proceedings of the Winter Simulation Conference, pages 3809-3820, as applied to claims x above, and further in view of Richard, “Regularization methods for prediction in dynamic graphs and e-marketing applications,” 2013, Doctoral dissertation, General Mathematics, École normale supérieure de Cachan - ENS Cachan, available at https://theses.hal.science/tel-00906066/document, pages 1-124.
Concerning claims 5, 12, and 19, Applicant’s specification Admits that “Gaussian Markov Random Field” is equivalent to an “autoregressive process” in [0027] as published:
“The series-expected value evolves through a latent autoregressive process of the log-expected value. This process constitutes a mean-reverting first-order autoregressive prior on the log-expected demand at each time step. The mean-reverting form of the process is appropriate for long-lifetime slow-movers, for which the long-run expected demand is assumed to not shift significantly from its past levels. The mathematical form of this process constitutes a Gaussian Markov Random Field, for which inference in this case is a nearly linear-time operation in the number of observations.”
Breuker discloses:
The system of claim 1, wherein the plate notation model (Breuker – see page 760 – showing how plate defines area that is repeated N times; page 761, FIG. 2 – showing plate; page 765, FIG. 4 – showing plate repeating) follows an autoregressive process (Salemi – see page 3810, 2nd paragraph - a Gaussian Markov Random Field).
It would have been obvious to combine Uhrig with Breuker and Salemi for the same reasons as discussed above with regards to claim 1.
To any extent Breuker and Salemi do not disclose the limitations, Richard discloses:
The system of claim 1, wherein the plate notation model follows an autoregressive process (Richard page 49, Section 2.2.2 - Predicting the future value of a time series is a challenge of interest for a wide range of applications as meteorology, economics, finance and supply chain management; see page 50 – Auto-Regressive models – time series is autoregressive).
It would have been obvious to combine Uhrig with Breuker and Salemi for the same reasons as discussed above with regards to claim 1. In addition, Uhrig, Breuker, and Salemi and Richard are analogous art as they are assessing forecasts/predictions (Uhrig par 39, 73; Breuker Abstract; page 759, Section 2.1 – probability distribution for variables; page 761, col. 2 – predicting; Salemi Abstract, page 3810 – optimization in operations research; Richard page 5, Overview). Breuker discloses a plate that repeats calculations (See page 760, 761, 765). Salemi discloses a Gaussian Markov Random Field (See page 3810). Richard improves upon Uhrig, Breuker, and Salemi by disclosing explicitly autoregressive models (See page 49-50) and tridiagonalizing matrices to reduce computational cost (See page 104). One of ordinary skill in the art would be motivated to further include the known autoregressive model to efficiently improve upon the adjust the stocking plans in Uhrig and the known Markovian math in Breuker, and the Gaussian Markov Random Field in Salemi.
**Suggestions for 103
Examiner suggests incorporating similar aspects from Allowance in 17/849,185 and 14/729,444 pertaining to – Claim 1 ‘185 – “model observed values of the at least two demand time series according to a hierarchical negative-binomial state space model” and [0024, 0048, 0054-0055, 0068-0070, 0079, 0096-0098] as published in this Application, and other places in the disclosure referring to these aspects.
Double Patenting
The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the “right to exclude” granted by a patent and to prevent possible harassment by multiple assignees. A nonstatutory double patenting rejection is appropriate where the conflicting claims are not identical, but at least one examined application claim is not patentably distinct from the reference claim(s) because the examined application claim is either anticipated by, or would have been obvious over, the reference claim(s). See, e.g., In re Berg, 140 F.3d 1428, 46 USPQ2d 1226 (Fed. Cir. 1998); In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) or 1.321(d) may be used to overcome an actual or provisional rejection based on nonstatutory double patenting provided the reference application or patent either is shown to be commonly owned with the examined application, or claims an invention made as a result of activities undertaken within the scope of a joint research agreement. See MPEP § 717.02 for applications subject to examination under the first inventor to file provisions of the AIA as explained in MPEP § 2159. See MPEP § 2146 et seq. for applications not subject to examination under the first inventor to file provisions of the AIA . A terminal disclaimer must be signed in compliance with 37 CFR 1.321(b).
The filing of a terminal disclaimer by itself is not a complete reply to a nonstatutory double patenting (NSDP) rejection. A complete reply requires that the terminal disclaimer be accompanied by a reply requesting reconsideration of the prior Office action. Even where the NSDP rejection is provisional the reply must be complete. See MPEP § 804, subsection I.B.1. For a reply to a non-final Office action, see 37 CFR 1.111(a). For a reply to final Office action, see 37 CFR 1.113(c). A request for reconsideration while not provided for in 37 CFR 1.113(c) may be filed after final for consideration. See MPEP §§ 706.07(e) and 714.13.
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/patent/patents-forms. The actual filing date of the application in which the form is filed determines what form (e.g., PTO/SB/25, PTO/SB/26, PTO/AIA /25, or PTO/AIA /26) should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to www.uspto.gov/patents/apply/applying-online/eterminal-disclaimer.
Claims 1, 3-5-10, 12-17, 19-20 are rejected on the ground of nonstatutory double patenting as being unpatentable over claims 1-20 of U.S. Patent No. 12,165,090 (app. no. 17/849,185) in view of Breuker, “Towards Model-Driven Engineering for Big Data Analytics – An Exploratory Analysis of Domain-Specific Languages for Machine Learning,” 2014, 47th Hawaii International Conference on System Sciences. IEEE, 2014, pages 758-767 and Brockett, “Flexible Purchase Frequency Modeling,” 1996, Journal of Marketing Research, Vol. XXXIII, pages 94-107, and Snyder (“Forecasting the intermittent demand for slow-moving inventories: A modelling approach,” 2012, International Journal of Forecasting, Vol. 28, pages 485-496.
18/943,270
12,165,090 (app. no. 17/849,185)
A system, comprising: a computer comprising a memory and a processor and configured to:
A system, comprising: a computer comprising a memory and a processor and configured to:
define a plate notation model for modelling a supply chain comprising one or more supply chain entities wherein the plate notation model comprises an inner plate that indicates a repetition of variables inside the inner plate and an outer area outside the inner plate that indicates a non-repetition of variables in the outer area;
model a supply chain comprising one or more supply chain entities and a supply chain inventory comprising one or more supply chain products;
receive historical sales data of the one or more supply chain entities comprising a single time series and explanatory variables, wherein the explanatory variables are represented in the inner plate of the plate notation model;
receive historical sales data comprising at least two demand time series and explanatory variables;
model observed values of the at least two demand time series according to a hierarchical negative-binomial state space model;
model a set of global parameters that are constant at each of the plurality of past time periods and each location, wherein the set of global parameters comprise global process parameters that condition the local process parameters;
perform inference using a Gaussian Markov Random Field and a sparse tridiagonal precision matrix over the single time series of unobserved supply chain model variables to generate local and global process parameters, a latent variable, and an effective latent variable, wherein the use of the sparse tridiagonal precision matrix increases a computing speed of computing a prior term in the single time series, wherein the local and global process parameters are represented in the outer area of the plate notation model, and wherein the latent variable and the effective latent variable are represented in the inner plate of the plate notation model;
perform inference using a Gaussian Markov Random Field and a sparse tridiagonal precision matrix over a time series of unobserved supply chain model variables to generate the local and global process parameters, a latent variable, and an effective latent variable, wherein the use of the sparse tridiagonal precision matrix increases the computing speed of computing a prior term in the time series;
generate forecasted latent variables based on the generated inferred local and global process parameters, latent variable, and effective latent variable;
generate forecasted latent variables based on the generated inferred local and global process parameters, latent variable, and effective latent variable;
generate a distributional demand forecast comprising probabilities of integer values of demand for one or more time steps into the future based, at least in part, on the forecasted latent variables;
generate a distributional demand forecast comprising probabilities of integer values of demand for one or more time steps into the future based, at least in part, on the forecasted latent variables;
in response to a current time period elapsing, repeating the receiving historical sales data of the one or more supply chain entities and the explanatory variables, performing inference, generating the forecasted latent variables and generating the distributional demand forecast based on actual sales data corresponding to the elapsed current time period being added to the historical sales data;
in response to a current time period elapsing to become one of the plurality of past time periods, repeating the receiving historical data, modelling, performing inference and generating forecasting steps based on the actual sales data corresponding to the elapsed current time period being added to the historical sales data;
and in response to the generated distributional demand forecast, generate a supply chain plan.
and in response to the generated distributional demand forecast, generate one or more replenishment orders for the one or more supply chain products.
Although the claims at issue are not identical, they are not patentably distinct from each other because: the claims here are slightly broader, they also have broader instances of many limitations.
Although the claims at issue are not identical, they are not patentably distinct from each other because: the claims here are slightly broader, they also have broader instances of some limitations.
Turning to limitations present in one application or another, they are obvious in view of secondary references:
The ’270 application here has a “plate notation model.” This is made obvious by Breuker as detailed in the 103 rejection above (see page 760, Table 1 – construction of directed graphical models; includes “plate” that can be repeated N times; page 765, Col 1 – plates; FIG. 4 – showing plats relative to observed variables, random variables, factors, nodes, and more).
The ‘090 patent has limitation “model observed values of the at least two demand time series according to a hierarchical negative-binomial state space model; model a set of global parameters that are constant at each location, wherein the set of global parameters comprise global process parameters that condition local process parameters”
Brockett discloses the limitations:
model observed values of the at least two demand time series according to a hierarchical negative-binomial state space model (Brockett – See page 94, Col. 1, 1st paragraph – negative binomial distribution (NBD) for stochastic modeling of purchase frequencies; modeling purchasing frequencies used for estimation of product sales; estimation of sales volume; estimation of stock-out quantities; Note: Snyder 2012 applied below also discloses the limitation - Snyder 2012 – See page 492, col. 2 - They confirm that better predictions may be obtained from distributions which allow for overdispersion, with the negative binomial distribution being the best option. This outcome was to be expected, because the negative binomial distribution has been used widely in inventory control for slow moving items, presumably because it has been found to work well in practice);
model a set of global parameters that are constant at each location, wherein the set of global parameters comprise global process parameters that condition local process parameters.
Snyder discloses the limitations (Applicant’s specification paragraph 55 - the long-run expected demand for a slow- mover, when projected far in the future, should fall back to a constant level in spite of any past transient disturbances; Snyder 2012 discloses the limitations – See page 486, col. 1, 1st paragraph - When volumes are low, the exponential smoothing framework must be based upon a distribution that describes count data, rather than the normal distribution. see page 493, col. 1, 3rd paragraph - In our framework, the undamped hurdle shifted Poisson model is the closest to the modified Croston model. Instead of smoothing the time gaps, we smooth the demand occurrence indicator variable using Eq. (2) [see page 488, col. 2]. See page 494, Co. 1, 4th paragraph - The mean was then revised in the light of this new simulated demand using simple exponential smoothing (with α = 0.1), to give a new mean of 0.675, with this change being a reflection of presumed permanent changes in the market for the inventory.)
Claim 6 of ‘090 patent recites:
6. The system of claim 1, wherein the observed values are conditioned by the effective latent variable, wherein the effective latent variable is conditioned by the latent variable and an explanatory variable of the received explanatory variables, and the latent variable is conditioned by a prior latent variable and local process parameters.
Brockett discloses the limitations:
such as known future promotions, sales, special events, and/or weather, including weather forecasts) to calculate unknown variables. Explanatory variables x ,t750 may be further combined with local coefficients 760 6f, which is the response to the explanatory variables for the -e-th time series. This response may take the form of a lift or a drop to the effective latent variable 720 at each time and location; par 116 – latent variables = process log-mean; par 132 - latent mean – demonstrate the effect of seasonality; Latent in FIG. 3 – is “mean (log) demand”; Brockett – See page 94, Col. 1, 1st paragraph – negative binomial distribution (NBD) for stochastic modeling of purchase frequencies; assessment of purchase frequency shifts due to changes or strategic variables (e.g. price, sales promotion) (sales promotion discloses an explanatory variable); See page 99, col. 1, last paragraph – ETNBD (extended truncated negative binomial distribution) parameterized as a truncated NBD; page 99, col. 2 – The other members of the (a.b,l) subclass are the logarithmic and truncated versions of the Poisson, binomial, geometric, and negative binomial distributions. The truncation is carried out by removing the probability mass at zero and normalizing the remaining probabilities so that they sum to one. These results are summarized in Table I, and we subsequently use them when applying Equation 18 to any member of the Class (a,b) distributions for subsequent period sales forecasts
and the latent variable is conditioned by a prior latent variable and local process parameters (Brockett – see page 94, col. 1, 1st paragraph – trend analysis is prediction of next period’s sales based on observed values of this period’s sales; See page 97, Col. 1, last paragraph – page 98, col. 2, 1st paragraph – researcher can increase complexity; can generalize the existing class of aggregate purchase-event frequency models to incorporate differential levels of data availability, which can include scanner data at the store level (i.e. local), and aggregation to the population-wide levels; See page 95, Col. 2 – decomposing purchase frequencies into individual consumer-level (or alternatively store-level) with parameters that vary across the population is called a mixture model [see equation 1]).
Breuker, Brockett, and Snyder are analogous art as they are directed to assessing forecasts/predictions (Breuker Abstract; page 759, Section 2.1 – probability distribution for variables; page 761, col. 2 – predicting; Brockett page 102 – considering seasonal promotion; See Snyder Abstract). 1) Breuker discloses recompiling code each time a constant value changes; observed variables can be changed without recompilation (See page 764, section 4.1) and having plates for repeating different calculations (See Table 1, FIG. 2, 4). Brockett improves upon Breuker by disclosing using a negative binomial distribution as well as Laplace transforms based on observed values of sales (See page 94) that is the basis for assessing mean, variance, and skewness for purchases (See page 102). One of ordinary skill in the art would be motivated to further include determining parameters using negative binomial distribution to efficiently and systematically compute the impact of different situations (e.g. seasonality, sales promotions) on estimated sales. 2) Breuker discloses some variables are for all factors and other specific variables or factors have a selector which is sometimes not used (See page 765, col. 1, 2nd paragraph); and Brockett discloses that its GCPP also includes a negative binomial secondary distribution (See pg. 99, Col. 1, last paragraph – page 100) and that purchase frequency models can have an extra parameterization that allows flexibility to accommodate the “spike-at-zero” problem associated with nonbuyers (See page 98, Col. 2, equation 7, and surrounding paragraphs); where the NBD and GCPP models “could” be fit by utilizing an extra parameter in equation 7 (See page 102, col. 1, 3rd paragraph). Snyder improves upon Breuker and Brockett by explicitly smoothing time gaps, revising the mean, and changing the mean based on any presumed permanent changes in the market (See par 486, 493-494) and calculating based on a “log” to estimate future values (See page 490). One of ordinary skill in the art would be motivated to further include smoothing demand and average demand to handle various types of seasonal, or volume effects on the forecast and would be motivated to further include determining parameters, utilizing a “log,” to efficiently and systematically compute the impact of different situations (e.g. seasonality, sales promotions) on estimated sales.
Accordingly, Claims 1, 3-5-10, 12-17, 19-20 are rejected on the ground of nonstatutory double patenting over claims 1-20 of U.S. Patent No. 12,165,090 since the claims, if allowed, would improperly extend the “right to exclude” already granted in the patent.
Response to Arguments
Applicant’s arguments 5/26/26 with respect to 103 appears to be that Uhrig does not disclose “plate notation model”. Remarks, pages 14-16. In response, Examiner respectfully disagrees with this analysis. Breuker discloses the “plate notation model” as cited in page 760 (Table 1), page 761 and FIG. 2, page 765, FIG. 4. In response to applicant's arguments against the references individually, one cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986).
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 IVAN R GOLDBERG whose telephone number is (571)270-7949. The examiner can normally be reached 830AM - 430PM.
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, Anita Coupe can be reached at 571-270-3614. 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.
/IVAN R GOLDBERG/Primary Examiner, Art Unit 3619
Prosecution Insights