Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendments
Applicant’s amendments filed on 03/02/2026 have been entered. Claims 1-20 are pending in this application of which claims 1, 12, and 20 are independent.
Response to Arguments
Applicant’s arguments in view of amendments, filed on 03/02/2026 have been fully considered and the examiner response is as follows:
Applicant’s arguments, Page 1, regarding the specification objection of the Abstracts length are considered and found convincing. Therefore, the objection is withdrawn.
Applicant’s arguments, Page 1 and 2, regarding 35 USC 112(b) are moot due to Applicant’s amendment. The rejections are withdrawn in light of the amendment.
Applicant’s arguments, Page 2, 3, and 4, regarding 35 USC 103 rejections are considered and not found persuasive as shown in the relevant claim limitation section and for the following reasons.
Applicant argues Attia’s feedback represents only “results of completed experiments” and not “at least one prediction” (Remarks, Pg. 2-3). Examiner respectfully disagrees. Attia expressly states that “our algorithm performs these updates using estimates from the early outcome predictor instead of using the actual cycle lives.” (Pg. 398-399), and Attia’s Fig. 3 caption (Pg. 399) describes the resulting “Evolution of the parameter space per round.” wherein “as more predictions are generated, the BO algorithm updates its cycle life estimates.”. This contradicts the Applicant’s argument that Attia is silent regarding the term “evolving, or a variation thereof” (Remarks, Pg. 3) and reads on the claim limitation in question. Applicant’s own operational definition of evolving at Specification [0014] – [0015] (The evolving may include maintaining diversity within the design parameter space. The maintaining may include employing a clustering method, pareto method, or a combination thereof.).
With respect to claims 10 and 18, Applicant’s argument is that “Okuyama does not remedy the above noted deficiencies of Gao and Attia” is unpersuasive because Gao and Attia are not deficient as Applicant contends as explained above. Therefore, the 35 USC 103 rejections are maintained.
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 non-obviousness.
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) 1-9 and 11 is/are rejected under 35 U.S.C. 103 as being unpatentable over Gao et al. “Machine learning toward advanced energy storage devices and systems” (2020) [herein “Gao”], and in view of Attia et al. “Closed-loop optimization of fast-charging protocols for batteries with machine learning” (2022) [herein “Attia”].
Regarding claim 1, Gao teaches
“A computer-implemented method for generative design of an energy storage device, the computer-implemented method comprising: automatically building at least one model of an energy storage device, the building based on a design parameter space and employing a machine learning process;” “This paper provides a comprehensive review of the application of machine learning technologies in the development and management of energy storage devices and energy storage systems. Machine learning has demonstrated success for solving a range of problems, including state estimation, life prediction, fault and defect diagnosis, property and behavior analysis, as well as modeling, design and optimization. Even more progress is achieved for ESDs with broad commercial applications (including batteries, capacitor/supercapacitors, fuel cells) because of the availability of a large amount of dataset to train the ML models.” (Summary and future directions). Some ML technologies combine multiple algorithms and models (such as combining CNN and RNN, combining RNN and DNN) to achieve the highest learning performance.” (Summary and future directions).
“automatically performing a simulation of the energy storage device, the simulation employing the design parameter space, a design evaluation space, and the at least one model built, the performing including producing at least one prediction of the energy storage device achieving at least one product design objective or the at least one model built achieving at least one model design objective;” “The machine learning technologies can be coupled with other approaches (such as experiments and numerical simulations) more tightly during the development of energy storage. For instance, machine learning can be used as an intermediate step for processing the experimental or numerical data. This will assist the design of the next-step experiments or simulation strategies, which can reduce redundant experiments or simulations and accelerate the development. A typical example…” “… optimizes the parameter space and promotes the efficiency of development (e.g., battery design, material selection, cell manufacture and control, elongation of the battery lifetime, etc.).” (Summary and future directions - Point 3).
“in an event the at least one prediction indicates the at least one product design objective has been achieved or the at least one model design objective has been achieved, automatically converging on the design parameter space evolved, thereby completing a generative design of the energy storage device and, otherwise, repeating the building, performing, and evolving.” “Appropriate ESD design, including choice of structural parameters, material selection, as well as designing operational strategies, is critical in ensuring the target cost, performance, efficiency, durability, and safety. A challenge is to systematically optimize the design for various conditions.” (Introduction and overviews).
Gao does not teach but Attia teaches,
“automatically evolving at least one of (i) the design parameter space and (ii) the design evaluation space, the evolving based on the at least one prediction produced and employing the machine learning process; and” “Optimal experimental design (OED) approaches are widely used to reduce the cost of experimental optimization. These approaches often involve a closed-loop pipeline where feedback from completed experiments informs subsequent experimental decisions, balancing the competing demands of exploration—that is, testing regions of the experimental parameter space with high uncertainty—and exploitation—that is, testing promising regions based on the results of the completed experiments. Adaptive OED algorithms have been successfully applied to physical science domains, such as materials science1,2,12,13,14, chemistry15,16, biology17 and drug discovery18, as well as to computer science domains, such as hyperparameter optimization for machine learning19,20.” (Main para. 1).
“Here we develop and demonstrate a machine learning methodology to efficiently optimize a parameter space specifying the current and voltage profiles of six-step, ten-minute fast-charging protocols for maximizing battery cycle life, which can alleviate range anxiety for electric-vehicle users8,9. We combine two key elements to reduce the optimization cost: an early-prediction model5…”. (Abstract).
“Figure 3 presents the predictions and selected protocols (Fig. 3a), as well as the evolution of cycle life estimates over the parameter space…”. (Pg. 3).
“Crucially, to reduce the total optimization cost, our algorithm performs these updates using estimates from the early outcome predictor instead of using the actual cycle lives.”. (Pg 2-3).
“Finally, these predicted cycle lives from early-cycle data are fed into the BO algorithm (Methods and Supplementary Discussion 2), which recommends the next round of 48 charging protocols that best balance the exploration–exploitation tradeoff.”. (Pg. 2).
“Evolution of the parameter space per round. The colour scale represents cycle life, as estimated by the BO algorithm. The initial cycle life estimates are equivalent for all protocols; as more predictions are generated, the BO algorithm updates its cycle life estimates.”. (Pg. 4).
Attia describes a closed loop pipeline over a parameter space of fast charging protocols. The feedback driving that loop comprises a machine learning generated life cycle predictions from an early outcome predictor, which Attia expressly teaches the evolution of the parameter space per round as those predictions update the algorithm’s estimates. This reads on the limitation under the Applicant’s own operational definition of evolving at Specification [0014 – 0015].
It would have been obvious to one of the ordinary skills in the art before the effective
filing date of the applicants claimed invention to combine Gao and Attia. Gao discloses building a ESD model based on parameter space using a machine learning process, an ESD simulation model that makes a prediction towards a design objective parameter, and repeating this process to converge on the design objective. Attia teaches a loop of optimization that includes testing to converge on a desired attribute using machine learning in different scientific domains. Gao states “Technology advancement demands energy storage devices (ESD) and systems (ESS) with better performance, longer life, higher reliability, and smarter management strategy. Designing such systems involve a trade-off among a large set of parameters, whereas advanced control strategies need to rely on the instantaneous status of many indicators. Machine learning can dramatically accelerate calculations, capture complex mechanisms to improve the prediction accuracy, and make optimized decisions based on comprehensive status information.” (Summary). Attia states “Simultaneously optimizing many design parameters in time-consuming experiments causes bottlenecks in a broad range of scientific and engineering disciplines.” “Our closed-loop methodology automatically incorporates feedback from past experiments to inform future decisions and can be generalized to other applications in battery design and, more broadly, other scientific domains that involve time-intensive experiments and multi-dimensional design spaces.” (Abstract). One would have been motivated to combine these two references, as this would allow an implementation of machine learning with a feedback-based process to accelerate and improve energy storage devices. This is supported by Gao “The machine learning technologies can be coupled with other approaches (such as experiments and numerical simulations) more tightly during the development of energy storage.” (Summary and Future Directions) and Attia which states “Our closed-loop methodology automatically incorporates feedback from past experiments to inform future decisions…” (Abstract) supporting that this combination would be predictively effective. Therefore, one would be motivated to combine Gao with Attia for the reasons stated above.
Regarding Claim 2, Gao teaches
The computer-implemented method of Claim 1, wherein the design parameter space and design evaluation space are associated with at least one of the energy storage device and the at least one model built, wherein the design parameter space includes variables,
wherein the variables include i) material variables, ii) system variables, or a combination i) and ii), and
“The goal for current ESD development can be grouped into five categories: (1) lowering the cost, (2) improving the performance and efficiency, (3) ensuring the usage safety, (4) promoting the reliability and durability, and (5) reducing the environmental impact. Achieving these goals rely on accurate ESD modeling that guides the performance analysis and design. One associated challenge is the identification of model parameters. Appropriate ESD design, including choice of structural parameters, material selection, as well as designing operational strategies, is critical in ensuring the target cost, performance, efficiency, durability, and safety.” (Introduction and overviews).
“…in order to ensure the consistency of electrochemical characteristics of battery cells and to avoid property discrepancies caused by material variation and fluctuations in manufacturing precision.” (Graphical Abstract).
wherein the variables are associated with a a) chemical space of the energy storage device, b) formula space of the energy storage device, c) material space of the energy storage device, d) configuration space of the energy storage device, e) process space of the energy storage device, or a combination of a)-e).
“In this paper, we provide a comprehensive review of recent advances and applications of machine learning in ESDs and ESSs. These include state estimation, lifetime prediction, fault and defect diagnosis, property and behavior analysis, modeling, design and optimization for ESDs…” (Introduction and overviews).
“…in the field of energy storage and renewable energy materials for rechargeable batteries, photovoltaics, catalysis, superconductors, and solar cells, specifically focusing on how machine learning can assist the design, development, and discovery of novel materials. These reviews mainly focus on the application of certain types of machine learning algorithms in a specific subarea.” (Introduction and overviews).
Gao teaches the current state of the art ESD design using machine learning with target parameters, and variables on different spaces as shown from summary excerpts. The Gao - Attia combination discussed for claim 1 render claim 2 obvious for the same
reasons.
Regarding Claim 3, Gao teaches
The computer-implemented method of Claim 1, wherein the design evaluation space includes at least one test of the energy storage device and wherein performing the simulation includes simulating the at least one test using the at least one model built.
“Machine learning is also used for the modeling, simulation, and design optimization of fuel cells, which is mostly embodied in constructing regression models between the input and output variables and further using the regression models in specific optimization algorithms.” (Application of machine learning for fuel cells).
Gao teaches using machine learning to model, simulate, and optimize ESD’s on different evaluation spaces. The Gao - Attia combination discussed for claim 1 render claim 3 obvious for the same reasons.
Regarding Claim 4, Attia teaches
The computer-implemented method of Claim 1, wherein the at least one product design objective includes at least one user-specified criterion associated with the energy storage device, at least one machine-generated criterion associated with the energy storage device, or a combination thereof, wherein the at least one product design objective includes a target product profile (TPP), and wherein the at least one model design objective includes at least one error threshold associated with a difference between simulated and real-world measurement, a target size for the design parameter space, or a combination thereof.
“Optimal experimental design (OED) approaches are widely used to reduce the cost of experimental optimization. These approaches often involve a closed-loop pipeline where feedback from completed experiments informs subsequent experimental decisions, balancing the competing demands of exploration—that is, testing regions of the experimental parameter space with high uncertainty—and exploitation—that is, testing promising regions based on the results of the completed experiments. Adaptive OED algorithms have been successfully applied to physical science domains, such as materials science1,2,12,13,14, chemistry15,16, biology17 and drug discovery18, as well as to computer science domains, such as hyperparameter optimization for machine learning19,20.” (Attia: Main para. 1).
Attia teaches machine learning optimization towards a target criteria using the simulated data, and experiments. Regarding TPP the definition from Gao of target data is noted from claim 1 as TPP which is target attributes wanted in the product. The Gao - Attia combination discussed for claim 1 render claim 4 obvious for the same reasons.
Regarding Claim 5, Attia teaches
The computer-implemented method of Claim 1, wherein the evolving includes (a) pruning the design parameter space, (b) pruning the design evaluation space, (c) expanding the design parameter space, (d) expanding the design evaluation space, or (e) a combination of (a)-(d).
PNG
media_image1.png
345
685
media_image1.png
Greyscale
(Figure 1. Attia: Shows a sample closed loop system exemplifying a-d.)
“First, batteries are tested. The cycling data from the first 100 cycles (specifically, electrochemical measurements such as voltage and capacity) are used as input for an early outcome prediction of cycle life. These cycle life predictions from a machine learning (ML) model are subsequently sent to a BO algorithm, which recommends the next protocols to test by balancing the competing demands of exploration (testing protocols with high uncertainty in estimated cycle life) and exploitation (testing protocols with high estimated cycle life). This process iterates until the testing budget is exhausted. In this approach, early prediction reduces the number of cycles required per tested battery, while optimal experimental design reduces the number of experiments required. A small training dataset of batteries cycled to failure is used both to train the early outcome predictor and to set BO hyperparameters. In future work, design of battery materials and processes could also be integrated into this closed-loop system.”
Attia teaches a closed-loop system for battery testing incorporates all four steps. The use of early outcome prediction after only 100 cycles effectively prunes the design evaluation space (b) by eliminating the need for lengthy full-cycle tests on unpromising protocols. The Bayesian Optimization (BO) algorithm works to prune the design parameter space (a) by focusing on the most promising designs and expands the design parameter space (c) by intentionally exploring uncertain protocols to discover new options. The system further facilitates expanding the design evaluation space (d) by taking the promising designs identified by the BO and integrating them back into the testing loop. The Gao - Attia combination discussed for claim 1 render claim 5 obvious for the same reasons.
Regarding Claim 6, Gao teaches
The computer-implemented method of Claim 1, wherein the evolving includes maintaining diversity within the design parameter space and wherein the maintaining includes employing a clustering method, pareto method, or a combination thereof.
“Unsupervised learning performs learning on unlabeled dataset and is typically used in the problems of clustering. A commonly used algorithm is the k-mean clustering algorithm (such as the k-nearest neighbor algorithm (k-NN)).” Gao shows different clustering methods that can be used in machine learning for energy storage devices. The Gao - Attia combination discussed for claim 1 render claim 6 obvious for the same reasons.
Regarding Claim 7, Attia teaches
The computer-implemented method of Claim 1, further comprising:
employing at least one monitored parameter in the building, evolving, or combination thereof, the at least one monitored parameter employed in at least one iteration of the building, evolving, or combination thereof, the at least one monitored parameter representing at least one real-world result generated via at least one real-world experiment employing the energy storage device, the real-world result effectuated based on employing, in the at least one real-world experiment, the design parameter space evolved.
Attia establishes a closed feedback loop that integrates testing data into an automated design methodology. Real-world electrochemical measurements (monitored parameters) are collected from batteries undergoing testing (real-world experiments). These results are then used as input for an ML model to make predictions, which informs a Bayesian Optimization algorithm. The BO algorithm generates the next set of testing protocols (evolving the design parameter space), which are implemented in the subsequent round of physical testing, continuously linking tangible outcomes back to the building and evolving phases of loop. The Gao - Attia combination discussed for claim 1 render claim 6 obvious for the same reasons.
Regarding Claim 8, Gao teaches
The computer-implemented method of Claim 1, further comprising employing a generative adversarial network (GAN), deep neural network (DNN), Bayesian optimization (BAO), genetic function approximation method, or a combination thereof, in the machine learning process.
“This paper reviews recent progresses in this emerging area, especially new concepts, approaches, and applications of machine learning technologies for commonly used energy storage devices (including batteries, capacitors/supercapacitors, fuel cells, other ESDs) and systems (including battery ESS, hybrid ESS, grid and microgrid-containing energy storage units, pumped-storage system, thermal ESS).” (Summary).
“A schematic of commonly used machine learning models
(A) Single-layer feed-forward neural network (SLFNN).
(B) Deep neural network (DNN).
(C) Auto encoder (AE).
(D) Convolution neural network (CNN).
(E) Recurrent neural network (RNN).
(F) Reinforcement learning architecture.
(G) Generative adversarial network (GAN).” (Figure 1.: Gao explain in the “Overview of machine learning technologies” section type of methodologies that are used for ESD’s.). Gao provides a background of machine learning methods that can be used for design of ESD’s, and further detailed examples that are referenced in the article. The Gao - Attia combination discussed for claim 1 render claim 8 obvious for the same reasons.
Regarding Claim 9, Gao teaches
The computer-implemented method of Claim 1, wherein the design parameter space includes semantically structured real-world evidence (RWE) data associated with experimentation of the energy storage device.
“The input to the discriminator is the fabricated dataset and the real dataset, whereas the output from the discriminator is the result showing whether the fabricated dataset is fake or real. ““The real dataset for the DC-GAN model are microstructural images. A comparison of the real and the synthetic data (based on morphological parameters, transport properties, and the two-point correlation function) …” (Summary: Battery design and optimization). The Gao article gives examples of using real and synthetic data in the machine learning process to enhance design choices. The Gao - Attia combination discussed for claim 1 render claim 9 obvious for the same reasons.
Regarding Claim 11, Gao teaches
The computer-implemented method of Claim 1, wherein the energy storage device is a battery and wherein converging on the design parameter space evolved includes identifying at least one of a compound, ingredient, additive, formula, recipe, or combination thereof that enables the at least one product design objective of the energy storage device to be achieved.
“The input to the ANN are electrode design, property, and manufacturing parameters including the active material volume fraction, particle radius, binder/additives volume fraction, electrolyte conductivity, and the compaction process pressure. The output from the ANN are resistance parameters, including the reaction resistance, the electrolyte resistance, and the diffusion resistance.”. (Battery design and optimization). The Gao - Attia combination discussed for claim 1 render claim 11 obvious for the same reasons.
Claim(s) 10 is/are rejected under 35 U.S.C. 103 as being unpatentable over Gao et al. “Machine learning toward advanced energy storage devices and systems” (2020) [herein “Gao”], and in view of Attia et al. “Closed-loop optimization of fast-charging protocols for batteries with machine learning” (2022) [herein “Attia”], and OKUYAMA et al. “JP 2019159864 A” (2019) [herein “OKUYAMA”].
Regarding Claim 10, neither Gao nor Attia teaches, but OKUYAMA teaches
The computer-implemented method of Claim 1, further comprising automatically storing, in a database, the at least one model built in association with the at least one prediction produced, at least one input of the at least one model built, and at least one output from the at least one model built,
“Next, the central processing unit 104 reads the target output parameter value (target processing result) and initial learning data from the database 105 and passes them to the target setting unit 107. The target setting unit 107 sets a target processing result (target output parameter value) (step S105). The set target output parameter value is transferred to the central processing unit 104 and stored in the database 105.” (DESCRIPTION-OF-EMBODIMENTS).
wherein performing the simulation includes inputting the at least one input to the at least one model built and, in response to the at least one input, generating the at least one output from the at least one model built.
“In correspondence with FIG. 1, the database 105 is implemented as a ROM 117 and a RAM 118, and each block in the search device is implemented as a program (search program) stored in the ROM 117.” (descriptions of embodiments).
“The RAM 118 stores learning data, a learning model, and the like generated in the search process.
In correspondence with FIG. 1, the database 105 is implemented as a ROM 117 and a RAM 118, and each block in the search device is implemented as a program (search program) stored in the ROM 117.” (BACKGROUND-ART).
OKUYAMA teaches storing in a database a model, associated prediction data, associated output data, associated input data, and simulation data.
It would have been obvious to one of ordinary skill in the art before the effective
filing date of the applicants claimed invention to combine Gao, Attia, and OKUYAMA, in order to enhance processing for ML models. One would have been motivated to combine the combination of Gao-Attia with OKUYAMA. This motivation is supported by OKUYAMA’S statement: “The present invention relates to a search device that searches for an optimal solution for processing…” and from Gao “Data pre-processing is needed in many cases in order to achieve high training accuracy for ML models.”. The relation is further expressed by OKUYAMA stating “…manufacturing and development stages of various products…” such as “…batteries, magnets, pharmaceuticals, and the like.”.
Claims 12-19 recite substantially the same limitations as claims 1, 4, 5, 6, 8, 9, 10, and 11 except these claims are directed to a “A computer-based system for generative design of an energy storage device, the computer-based system comprising: at least one memory; and at least one processor coupled to the at least one memory, the at least one processor configured to:”. Therefore, these claims are rejected under the same rationale as addressed above.
Claim 20 recites substantially the same limitations as claims 1 except these claims are directed to a “A non-transitory computer-readable medium for generative design of an energy storage device, the non-transitory computer-readable medium having encoded thereon a sequence of instructions which, when loaded and executed by at least one processor, causes the at least one processor to:”. Therefore, these claims are rejected under the same rationale as addressed above.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
US20190115778A1 by ERMON et al teaches a multidimensional parameter space of battery cell test protocols.
US11226374B2 by SEVERSON et al teaches predictive modeling to predict and classify battery cells.
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to NARCISO EDUARDO MONTES whose telephone number is (571)272-5773. The examiner can normally be reached Mon-Fri 8-5.
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, REHANA PERVEEN can be reached at (571) 272-3676. 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.
/N.E.M./Examiner, Art Unit 2189
/REHANA PERVEEN/Supervisory Patent Examiner, Art Unit 2189
Prosecution Insights