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 .
Status of the Claims
Claims 1, 3-6, 8 and 10-15 were previously pending and subject to a final office action mailed 12/08/2025. Claims 1, 3-6, 8, 10, 13-15 were amended; claim 3 was cancelled, and no claim was added in a reply filed 02/23/2026. Therefore claims 1, 4-6, 8 and 10-15 are currently pending and subject to the non-final office action below.
Response to Arguments
Applicant's arguments filed 02/23/2026 in regards to the 101 rejection have been fully considered but they are not persuasive.
Applicant argues “The claims have been amended to recite optimizing the operation of one or more of electrical generation equipment, electrical consumption equipment and electrical storage equipment at a facility according to the synthesized time-series constructed using the claimed approaches. These limitations are not open-ended because they are directed to the control and operation of very specific types of electrical equipment. Using the detailed synthesized time-series to guide the control also takes the claims away from reciting any general or generic control mechanism.” (remarks p. 8).
Applicant’s argument has been considered but is not persuasive. Claim 1 does not affirmatively recite controlling or operating the electrical equipment “according to” the synthesized time series. The actual limitation states: “wherein the synthesized time-series is used to optimize an operation of one or more of electrical generation equipment, electrical consumption equipment and electrical storage equipment at the facility. “ This limitation recites the desired use or intended result of the synthesized time series. It does not specify what equipment parameter is optimized, what optimization calculation is performed; what control determination is made; what control signal is generated; what operating state of the equipment is changed”
The claim therefore does not require a particular technological implementation of the identified abstract idea. It merely directs that the result of the data analysis process, the synthesized time series, be used for the broadly stated purpose of optimizing equipment operation.
Under step 2A, prong Two, merely stating that the result of an abstract calculation is to be applied in a particular technological environment does not, without more, integrate the exception into a practical application. The additional limitation must impose a meaningful technological limit on the exception (please see MPEP 2106.04(d), 2106.05(e)-(h)).
Furthermore, the claims do not recite “control” of the equipment or using the synthesized time series to “guide the control”. Those actions appear only in Applicant’s argument. Claim 1 identifies broad categories of equipment: electrical generation equipment; electrical consumption equipment; and electrical storage. The claim does not identify a particular machine within these categories or require any particular interaction between the time series generator and the equipment. For example, the claim does not require that the time series generator be operably connected to the equipment, that a controller receives the synthesized time series, or that an equipment operating state be physically changed.
Merely limiting the intended use of calculated information to a particular field or technological environment does not integrate an abstract idea into a practical application. The broad categories of electrical equipment amount to the environment in which the calculated time series is intended to be used, rather than a particular machine that meaningfully implements the claimed calculations (please see MPEP 2106.05(g)).
Applicant argues “The claimed approach also provides meaningful technical advantages in the technical field of facility power control and management because the claimed approach is implemented in real time, is automated, and overcomes the art's inability to obtain or supply effective time-series in a quick and efficient way to control electrical equipment. The claimed approach produces time- series that can effectively operate specific pieces of electrical equipment and efficiently operate a specific type of facility with these specific types of equipment. This, in turn, increases the profitability of these facilities in meaningful ways. The electronic control and usage of these specific types of equipment is also not a mental process and cannot be accomplished using a pad of paper and/or a pencil.” (remarks p. 10).
Applicant’s argument is not persuasive. First, claim 1 does not recite that the claimed process is implemented “in real time”. The claim recites a “non-training, execution phase” but a non-training phase is not necessarily a real time phase. It merely distinguishes execution or inference from model training. The execution could occur well before any contemplated equipment operation.
Second, claim 1 does not recite controlling electrical equipment. It only states that the synthesized time-series “is used to optimize an operation.” The claim does not recite the mechanism or steps by which the synthesized time series produces a control action.
Third, the claimed automation consists of generic computer components, a database, time series generator, and reference data acquisition module, used to perform database searching, similarity evaluation, weighted mathematical combination, or machine learning prediction. The claims do not identify an improvement to the operation of these computer components themselves.
Fourth, the assertion that the system obtains an effective time series “in a quick and efficient way” is a result, not a claimed technological mechanism. An asserted improvement must be reflected in the claim, including the components or steps responsible for the technological improvement. The Office’s current guidance explains that the specification may illuminate an improvement, but the claim must still reflect the technological advance rather than merely claim a desired result (please see MPEP 2106.04(d)(1) and USPTO’s 2025 memo).
The claim remains broadly directed to: obtaining facility attributes; identifying similar stored attribute sets; selecting associated reference time series; mathematically combining or otherwise processing those time series; and designating the resulting information for use in equipment optimization. That is an information processing solution, not a claimed improvement to the operation of the electrical equipment, computer, database, or time series generator.
Furthermore, as argued above, the claim still does not require the synthesized time series to “operate” any piece of equipment. Producing information that may subsequently be useful in operating equipment is different from affirmatively controlling or operating that equipment. Claim 1 is satisfied once the synthesized time series has been generated and is “used to optimize” an operation in an unspecified manner. It does not require that: the equipment receives the time series; the equipment receive a command derived from the time series; the equipment respond to the time series or operating setpoint be changed. Applicant’s description of what the time series “can” do therefore does not add a limitation to the claim. Patent eligibility is evaluated based on what the claim requires, not on unclaimed uses or possible downstream benefits of the generated information.
Also, Applicant’s argument in regards profitability is a financial or business benefit, not an improvement to computer functionality or to the operation of electrical equipment. Moreover, the claims do not recite: determining profitability; reducing operating costs; increase revenue or controlling equipment according to a profitability calculation. The alleged profitability benefit therefore does not demonstrate that the claimed abstract idea has been integrated into a practical technological application.
Lastly, the office has not identified the physical operation of equipment as the abstract idea. The abstract idea lies in the claimed information processing operations, including: comparing input attributes with stored candidate attributes; evaluating similarity….optimizing weights or predicting time series through mathematical models. These limitations recite mathematical relationships and evaluations of data. The fact that the claim uses a computer to perform the calculations does not, by itself, transform those calculations into an eligible technological improvement.
Second, the claim does not require the physical “electronic control and usage” on which Applicant relies. Because no control signal, equipment state change, or physical equipment operation is recited, Applicant’s assertion that such physical control cannot be performed mentally does not address the claim as written.
Even where the volume of data or speed of calculation makes mental performance impractical, a claim may remain directed to an abstract mathematical or information analysis process when generic computer components merely execute the analysis. The relevant question is not simply whether the complete claim could practically be performed with pencil and paper, but whether the additional elements meaningfully integrate the recited exception into a technological application.
Applicant’s arguments with respect to 103 rejection have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following is a quotation of pre-AIA 35 U.S.C. 112, sixth paragraph:
An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph:
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific structural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Claim limitations in this application that use the word “means” (or “step”) are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action. Conversely, claim limitations in this application that do not use the word “means” (or “step”) are not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier. Such claim limitation(s) is/are: “a time series generator” and “time series modification module” in claim 1, and “a reference data acquisition module” in claim 3.
Because this/these claim limitation(s) is/are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
If applicant does not intend to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1, 4-6, 8 and 10-15 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Claim 1/13 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The claim recites “storing existing site attributes including a first candidate set of attributes and a second candidate set of attributes, and also storing a first reference time- series and a second reference time-series, the first candidate set of attributes mapped to the first reference time-series and the second candidate set of attributes mapped to the second reference time-series; receiving, during a non-training phase, execution phase, an input set of attributes comprising attributes characterizing the facility, searching for existing site attributes and identifying the first candidate set of attributes and the second candidate set of attributes as most similar to the input set of attributes; and constructing a synthesized time-series representing estimated energy data for the facility; during the non-training, execution phase: construct the synthesized time series by either: i) calculating a weighted combination of the first and second reference time-series to form the synthesized time-series, and outputting the weighted combination as the synthesized time-series, or ii) concatenating selected intervals from the first and second reference time-series to form the synthesized time-series, the intervals being selected according to a similarity between the reference sets of attributes respectively associated with the first and second reference time-series and the input set of attributes; optimizing an operation of one or more of electrical generation equipment, electrical consumption equipment and electrical storage equipment at the facility according to the synthesized time-series”
The limitations above, as drafted, is a process that, under its broadest reasonable interpretation, covers a method of synthesizing energy data. That is, the method allows for mathematical relationships and mental processes because it allows for mathematical concepts and concepts that can be done in the human mind (with the help of pen and paper).
This judicial exception is not integrated into a practical application. In particular, the claim recites “a time series generator” and “database” in claim 1, a database in claim 13. The additional limitation is recited at a high level of generality and is no more than mere instructions to apply the exception using a generic computer component. The claims further recite “one or more of electrical generation equipment, electrical consumption equipment and electrical storage equipment “. However, this equipment is generally linked to the abstract idea which amounts to only field of use limitations. Accordingly, these additional elements, alone or in combination, do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. The claim is directed to an abstract idea.
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements, alone or in combination, are nothing more than mere instructions to apply the exception on a general computer.
Dependent claims 4-6, 8, 10-12 and 14-15 are also directed to an abstract idea without significantly more because they further narrow the abstract idea described in relation to claim 1/13 without successfully integrating the exception into a practical application (“time series generator” and “a reference data acquisition module” are recited at a high level of recitation which amounts to mere instructions to apply the exception in a computer environment) or providing significantly more limitations.
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.
The factual inquiries 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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claim(s) 1 and 10-14 is/are rejected under 35 U.S.C. 103 as being unpatentable over Robert Clain, “Generating load profiles from building characteristics”, published by CS229 Final Project Stanford in Oct 25, 2017, hereinafter “Clain” in view of Fahad H. Al-Qahtani, “Multivriate k-nearest neighbour Regression for Time Series data – a novel Algorithm for Forecasting UK Electricity Demand”, published by IEEE Xplore in January 2014, hereinafter “Fahad”, in further view of Miller (US 7274975).
As per claim 1/13, Clain discloses a system for synthesizing energy time-series data for a facility that consumes or generates energy, the system comprising:
a database storing existing site attributes including a first candidate set of attributes and a second candidate set of attributes, the database also storing a first reference time-series and a second reference time-series, the first candidate set of attributes mapped to the first reference time-series and the second candidate set of attributes mapped to the second reference time-series (page 1, “The CBECS survey is conducted approximately every 4 years, and includes building information such as square footage, building use, occupancy, and total energy used per year. The simulation study [2] includes yearly load profiles with 1-hour intervals, in Joules, for most buildings from the survey.”, “The NREL simulation results (existing stock only) were aggregated and then matched by respondent ID to the CBECS features to create the full dataset, containing survey results and load profiles for 4,820 buildings. Buildings with poor survey responses were discarded (mostly NA values), leading to an eventual 4,220 samples.”, each of Clain’s thousands of buildings records is a candidate attribute set mapped by respondent ID to a corresponding yearly hourly load time series. Any two records selected from that database constitute the claimed first and second candidate sets and their first and second reference time series);
a time-series generator configured to receive, during a non-training, execution phase, an input set of attributes comprising attributes characterizing the facility, and to output a synthesized time-series representing estimated energy data for the facility (page 1, “Several models were trained in this study, using a combi nation of survey data collected by the DOE (Department of Energy) and simulation data from NREL (National Renewable Energy Laboratory)., “Because of these issues, the goal of this project was to create a quick and fairly accurate method for generating a load profile given a subset of building characteristics.”, page 2, “The predictive power of each model was estimated using an average of scores from 100 randomly left out samples (left out one at a time from the training set, then tested). While this was not ideal and a more robust cross-validation is desired, such computation is prohibitively expensive given the nature of the similarity metric and the size of the data. The average of each term in (1) and the average final measure have been reported for comparison”. Clain distinguishes training records from left out test inputs and generates an estimated load profile for a building from its characteristics. The test/use operation is the claimed non-training execution phase.);
wherein time-series generator comprises a reference data acquisition module configured to search the database for existing site attributes and to identify the first candidate set of attributes as most similar to the input set of attributes (page 2, “Two simple models were chosen to compare against others. One is k-nearest neighbors (with k=1),which simply picks the closest profile from the data set. Another is DC regression, which regresses on a single frequency component of the response, even though many are present”, “.For the first, the nearest neighbor was determined using the continuous variables (Euclidean distance) with unlikely ties broken by the number of identical categorical variables (Hamming distance), as many of the important variables found in [3] are continuous.” Clain expressly searches mapped building records using distance between the input building characteristics and candidate building characteristics, but its K=1 implementation retains only the closest record.);
However, Clain does not disclose but Fahad discloses identify the first candidate set of attributes and the second candidate set of attributes as most similar to the input set of attributes (page 2, “In general, the k-NN is an instance based (i.e. local) learning algorithm, which given an unlabelled object finds a group of k most similar objects, i.e. neighbours, in the training dataset given its features, and uses them to assign a class label to match the class of the majority of the k similar cases in the training neighbourhood.”) and wherein the time-series generator s further configured to during the non-training, execution phase: construct the synthesized time-series by either: calculating a weighted combination of the first and second reference time-series to form the synthesized time-series, and outputting the weighted combination as the synthesized time-series, or ii) concatenating selected intervals from the first and second reference time-series to form the synthesized time-series, the intervals being selected according to a similarity between the reference sets of attributes respectively associated with the first and second reference time-series and the input set of attributes (page 2, “Finally, the distances , . (1) for all m-histories are ranked, and the k vectors with the lowest distance to the target feature vector aggregated to form a prediction, often using a simple arithmetic mean with equal weight:” the rejection relies on alternative I. Fahad ranks the candidate feature vectors, selects k=2 nearest neighbors, and averages the corresponding hourly future load values to output a forecast trace. The target vector ranking and averaging occur during generation of the forecast for a new input, the non-training execution phase).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of the invention to include the limitation above as taught by Fahad in the teaching of Clain, in order to finds a group of k most similar objects, i.e. neighbours, in the training dataset given its features, and uses them to assign a class label to match the class of the majority of the k similar cases in the training neighbourhood. (please see Fahad, page 2).
However, Clain still does not disclose but Miller discloses wherein the synthesized time-series is used to optimize an operation of one or more of electrical generation equipment, electrical consumption equipment and electrical storage equipment at the facility (abstract, “Methods and systems are provided for optimizing the control of energy supply and demand. An energy control unit includes one or more algorithms for scheduling the control of energy consumption devices on the basis of variables relating to forecast energy supply and demand. Devices for which energy consumption can be scheduled or deferred are activated during periods of cheapest energy usage”. 1:51-61, “(8) The invention includes various systems and methods for increasing the efficiency with which energy can be managed. In one variation, an integrated control device manages the supply of energy from various sources (e.g., electric grid, battery, photovoltaic cells, fuel cells) and the demand for energy from consumption devices (e.g., hot-water heaters, HVAC systems, and appliances). An optimization algorithm determines based on various factors when to activate the energy sources and when to activate the consumption devices.”).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of the invention to include the limitation above as taught by Miller in the teaching of Clain, in order to scheduling the control of energy consumption devices on the basis of variables relating to forecast energy supply and demand (please see Miller, abstract).
As per claim 10, Clain discloses wherein the time-series generator is configured to generate the synthesized time-series by inputting the input set of attributes to a machine learning model trained to predict the first and second reference time-series on the basis of the respective reference sets of attributes input to the model (page 1, “Several models were trained in this study, using a combi nation of survey data collected by the DOE (Department of Energy) and simulation data from NREL (National Renewable Energy Laboratory). “, “The NREL simulation results (existing stock only) were aggregated and then matched by respondent ID to the CBECS features to create the full dataset, containing survey results and load profiles for 4,820 buildings. Buildings with poor survey responses were discarded (mostly NA values), leading to an eventual 4,220 samples”, page 2, “The first attempt at a more complicated and more accurate model was to create a multivariate multiple linear regression on the amplitude and phase of different frequency components of the responses using a truncated Fast Fourier Transform (FFT) using all of the predictors.”, “A regression approach was also used as a baseline as it is more biased than nearest neighbor and has the ability to generate a profile outside those seen in the training data. As the true responses are high in dimension, regression was performed on the simplest component, the frequency amplitude. Two methods were used here as well. The first method is a pure linear regression, while the second is a random forest regression.”, pages 2-3, “The first attempt at a more complicated and more accurate model was to create a multivariate multiple linear regression on the amplitude and phase of different frequency components of the responses using a truncated Fast Fourier Transform”, page 3 “A second method, using a multivariate regression tree, was also shown as this was done for DC regression.” And “Classification for the group labels used a support vector machine with a linear kernel, with error being calculated between the true signal and the reference signal for the predicted label.” This discloses that the system uses machine learning models (regression, decision trees, support vector machines) that are trained on building characteristics to predict or select appropriate time series profiles.” Clain’s training examples consist of building attribute sets as predictors and respondent matched hourly load profiles as target response. Its regression models learn the relationship between those attributes and profile response components, and then receive a new building’s attributes to generate a synthesized profile. The claimed first and second reference attribute/profile pairs are any two mapped members of that supervised training corpus. Their respective attribute sets are feature inputs and their mapped time series are training targets. Claim 10 does not require the model to output the two stored reference profiles separately at runtime. It requires those reference profiles to serve as training targets and the new facility attributes to produce the synthesized output.)
As per claim 11, Clain discloses wherein the time-series generator is further configured to modify the synthesized time-series in a post-processing step to match one or more predetermined key attributes of the facility (page 3, “Since it appeared there was a common pattern to the mean and spread scores for which medoid clustering poorly predicted, a second stage prediction for mean value was carried out within the predicted group using a random forest. The mean was then divided by the mean of the centroid to get an appropriate scaling factor”. This shows that the system performs post processing modifications (second stage predictions and scaling) to adjust the synthesized time series to better match specific building characteristics like mean energy usage).
As per claim 12, Clain discloses wherein attributes include one or more of: a classification of the facility; annual energy consumption or generation at the facility; operation hours; geographic location; equipment used by the facility (page 1, table 1, shows building features such as government owned and PBA8 as classification, electricity used as annual energy consumption at the building, wkhrs8 weekly operating hours and region as geographic location).
As per claim 14, Clain discloses wherein the post-processing step comprises rescaling the synthesized time-series to match one or more of a total energy consumption of the facility, a total energy generation of the facility, a peak power of the facility, or an average power of the facility (page 3-4 “Since it appeared there was a common pattern to the mean and spread scores for which medoid clustering poorly predicted, a second stage prediction for mean value was carried out within the predicted group using a random forest. The mean was then divided by the mean of the centroid to get an appropriate scaling factor. This method had very good results. The correlation score was very similar to what was achieved earlier and the mean/spread scores improved.” Second stage prediction is equivalent to “post processing step”. “divided by the mean of the centroid to get an appropriate scaling factor” is “rescaling” and the medoid profile being scaled is the synthesized output. The scaling operation is performed to improve mean scores, meaning to match the mean characteristics. “total energy consumption” is equivalent to the mean of a load profile over time directly relates to total energy consumption. Total energy = mean power x time duration. Rescaling the mean rescales the total energy consumption proportionally. The target facility’s characteristics which are the input attributes determine the scaling factor. Therefore, rescaling by mean inherently rescales total energy consumption.)
Claim(s) 4-6 is/are rejected under 35 U.S.C. 103 as being unpatentable over “Clain” in view “Fahad”, in further view of Miller (US 7274975), as disclosed in the rejection of claim 1, in further view Leen De Baets, “detection of unidentified appliances in non-intrusive load monitoring using Siamese neural networks”, published by Electrical Power and Energy systems, on July 2018, hereinafter “Baets”.
As per claim 4, the combination of Clain in view of Fahad discloses the two ranked facility records and their weighted reference profiles. Furthermore, Clain does not disclose but Fahad discloses wherein the reference data acquisition module selects first and second reference time-series based upon a similarity metric (page 2-3, Fahad uses distance-based metric to select the closest nearest neighbors). However, Clain in view of Fahad does not disclose but Baets discloses that the similarity metric satisfying a predetermined similarity threshold (page 647, To determine which cluster a test sample belongs to (if any), its feature vector is first transformed to the calculated lower-dimensional space. Next, the Euclidean distance is calculated to all core samples and the minimal distance is selected. If this distance is smaller than threshold ∊, then the sample belongs to the same cluster as the closest core sample. Otherwise, it will be assigned the label ‘unidentified’.).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of the invention to include the limitation above as taught by Baets in the teaching of Clain in view of Fahad, since the claimed invention is merely a combination of old elements, and in the 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.
As per claim 5, Clain in view of Fahad, Miller and Baets disclose all the limitations of claim 4. Clain does not disclose but Fahad discloses wherein the reference data acquisition module is configured to determine the similarity metric for each of the first and second candidate set of attributes by computing a distance metric representing the distance between the first and second candidate set of attributes and the input set of attributes (page 2 discloses Fahad uses distance based selection to two nearest candidates)(please see claim 1 rejection for combination rationale).
As per claim 6, Clain in view of Fahad, Miller and Baets disclose all the limitations of claim 4. Clain discloses each of the first and second candidate set of attributes and the input set of attributes as a feature vector (page 2, “The distance measure used for knn was non-obvious, as there is a mixture of continuous and categorical variables in the dataset. Two different methods were used, both of which first scaled the data to be centered and have unit variance. For the first, the nearest neighbor was determined using the continuous variables (Euclidean distance) with unlikely ties broken by the number of identical categorical variables (Hamming distance), as many of the important variables found in [3] are continuous. The second method used all of the variables with dummy variables introduced for categorical features. The score results are summarized in Table II and Figure 1 shows the scores of the second method.” Clain represents the input facility and each stored candidate facility through scaled continuous variables and dummy coded categorical variables. These collections of numerical facility attributes constitute feature vectors.).
However, Clain in view of Fahad does not disclose but Baets discloses wherein the reference data acquisition module is configured to determine the similarity metric for each candidate set of attributes by inputting each of the first and second candidate set of attributes and the input set of attributes as a feature vector to a trained classifier trained to predict, as the target, a similarity metric on the basis of an input feature vector comprising two such sets of attributes (page 647, “As input, two binary VI images must be given and as label, a binary value indicating whether or not the images belong to the same class. The output of the siamese network are two vectors, forming a lower-dimensional representation of the two input images. The idea is to learn the representation in such a way, that the distance between these re presentations will be smaller than a given threshold if the two belong to the same class and larger if not.” Baets explain that the binary label identifies whether the two inputs belong to the same class. The paired inputs pass through neural networks having identical architectures and shared weights. The networks generate feature vector representations whose distance is trained to be smaller for similar pairs and larger for dissimilar pairs).
Since each individual element and its function are shown in the prior art, albeit shown in separate references, the difference between the claimed subject matter and the prior art rests not on any individual element or function but in the very combination itself. That is in the substitution of the learned paired inputs metric taught by Baets for the manually constructed Euclidean/Hamming metric of Clain.
Thus, the simple substitution of one known element for another producing a predictable result renders the claim obvious.
Claim(s) 8 is/are rejected under 35 U.S.C. 103 as being unpatentable over Clain in view of Fahad and Miller, as disclosed in the rejection of claim 1, in further view of Shen, “optimizing weights by genetic algorithm for neural network ensemble”, published in Advances in Neural Networks, Springer 2004 (pages 323-331), hereinafter “Shen”.
As per claim 8, Clain does not disclose but Shen discloses wherein the time-series modification module is further configured to use an optimization algorithm to find optimal weights for the weighted combination (abstract “After the training of component neural networks, genetic algorithm is used to optimize the combining weights of component networks”, “In this chapter we present experiments where the dynamic weighted ensemble algorithm is used with neural networks generated by bagging and AdaBoost.”, page 327-328 “After the training of component neural networks, the genetic algorithm employed in our study is realized by the Genetic Algorithms for Optimization Toolbox (GAOT) developed by Houck et al.. The genetic operators, including select, crossover, and mutation, and the system parameters, including the crossover probability, the mutation probability, are all set to the default values of GAOT. The maximum generation and the population are set to 200, and a float coding scheme that represents each weight in 64 bits is used. The evolving process finished as the maximum generation was reached…In order to know how well the compared ensemble methods work, for example, how significant the predicting accuracy and generalization ability are improved by utilizing these ensemble methods, in the experiments we also test the performance of single neural networks. For each dataset, fifty single neural networks are trained…”)
Therefore, it would have been obvious to one of ordinary skill in the art at the time of the invention to include the limitations above as taught by Shen in the teaching of Clain, in order to achieve high predicting accuracy (please see Shen, abstract).
Claim(s) 15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Clain in view of Fahad and Miller, as disclosed in the rejection of claim 11, in view of Hanno Teiwes, “Energy Load profile analysis on machine level”, published by ScienceDirect in 2018, hereinafter “Teiwes”.
As per claim 15, Clain discloses operation attributes defining facility schedules such as OPEN248 (open 24/hr), OPNMF8 (open Monday to Friday), OPNWE8 (open weekend), WKHRS8 (weekly operating hours) and MONUSE8 (months used). It also discloses post processing modification as shown in claim 14 rejection. a load pattern is the temporal pattern of energy consumption in the synthesized time series and production pattern of the facility is operational schedule indicating when production is running vs. not running). However, Clain fails to disclose but Teiwes discloses wherein the post-processing step comprises adjusting the synthesized time series so that its load pattern follows a production pattern of the facility (abstract “The presented approach focuses on the evaluation of electric load profiles on machine and workgroup level in dependency on available complementing data such as product, scheduling or machine data”, page 271-272 “One widely applied starting point for each analysis is to meter electricity demands of machine equipment in order to receive their individual, dynamic load profiles… In Figure 1 a typical load profile of a machining center as single consumer of a manufacturing system is displayed. Typical machine states such as off, standby, ready for processing and processing as well as the active components in each state have been highlighted… By adding information about ambient conditions such as temperature or humidity data in a next step, correlations of observed deviations with climatic factors such as temperature or humidity become visible. Hence, the production conditions could be adapted to improve the energy efficiency, e.g. by changing heating, ventilation, air conditioning and cooling (HVAC) facilities or settings [13]… As different product types often require different process conditions, relevant process parameters such as temperature, cutting parameters etc. can also be analyzed in order to explain deviations. A different approach is to combine the acquired load profiles with machine scheduling information. This makes it possible to distinguish between the value-adding process intervals and other nonvalue-adding consumer states such as waiting” page 275 “The basic idea is to acquire the energy demand for a specific process basic step…. The energy demand for each basic step is calculated from its average energy value to determine the corresponding load levels. These load levels for each basic step are documented in a process related energy database… each process basic step is further combined with additional data on used machinery, robots and tools. The result is a process basic step and production equipment specific database, allowing to recombine different basic steps to a new manufacturing process… The load profile was metered in a body shop manufacturing cell of a German automotive OEM. The cell contains one robot that performs various handling, tool exchange and welding operations… After the initial metering step, the load profile is combined with the available basic process step chart. This chart consists of standardized process elements (e.g. handling, tool exchange, welding). Accordingly, the average load level for each basic step is calculated, identifying the welding as the most critical basic step… Now, these elements can be used in the planning phase of a new body shop cell when they are re-combined according to the new process.” Teiwes explicitly teaches analyzing load profiles based on machine operating states (processing vs. non processing periods) and production scheduling. It also teaches adjusting energy profiles based on production scheduling to distinguish value adding (production running) from non-value adding (production not running) periods. It discloses reconstructing load profiles by combining basic steps according to production sequences, adjusting profiles to match production patterns. Therefore, it explicitly discloses adjusting synthesized load profiles to match production sequences by combining process steps according to operational patterns.)
Therefore, it would have been obvious to one of ordinary skill in the art at the time of the invention to include the limitations above as taught by Teiwes in the teaching of Clain, in order to distinguish between the value-adding process intervals and other nonvalue-adding consumer states such as waiting (Teiwes, p. 272)
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to OMAR ZEROUAL whose telephone number is (571)272-7255. The examiner can normally be reached Flex schedule.
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, Lynda Jasmin can be reached at (571) 272-6782. 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.
OMAR . ZEROUAL
Examiner
Art Unit 3628
/OMAR ZEROUAL/Primary Examiner, Art Unit 3629
Prosecution Insights