Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
DETAILED ACTION
This Office Action corresponds to application 18/198,975 which was filed on 5/18/2023.
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 1/28/2026 has been entered.
Response to Amendment
In the reply filed 1/28/2026, claims 1, 8, and 15 have been amended. No additional claims have been added or canceled. Accordingly claims 1-5, 8-12, and 15-19 stand pending.
Response to Arguments
Applicant's arguments filed 1/28/2026 have been fully considered but are moot in view of new grounds of rejection.
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.
Claim(s) 1-2, 4, 8-9, 11, 15-16, and 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Flunkert et al. (US10936947, previously present in ‘892), hereinafter Flunkert, in view of Rath (US2021/0328888, previously present in ‘892), Arik et al. (US2023/0110117, previously present in ‘892), hereinafter Arik, and Amiri Moghadam et al. (US2024/0193488), hereinafter Amiri.
Regarding Claim 1:
Flunkert teaches:
A method, comprising: receiving, by a computing system, a request to forecast a time series absent historical time series values (Flunkert, figures 1 and 3, column 8 lines 13-34, column 9 line 55 – column 10 line 18, note using time series data and metadata to predict a forecasted value for items which were not represent in the training data set or for which very few observations were included in the training data; note forecast request);
receiving, by the computing system, a set of time series and an associated set of metadata text comprising a first subset of metadata text and a second subset of metadata text, the first subset of metadata text associated with the set of time series, the second subset of metadata text associated with the time series (Flunkert, figure 1 and 3, column 2 line 62 – column 3 line 38, column 4 line 63 – column 5 line 9, column 5 line 65 – column 6 line 12, note receiving time series data and associated metadata; note the metadata is associated with the items in the training set and/or with the items for which forecasts are to be generated and since the time series comprises multiple items it is interpreted to have multiple subsets of associated metadata);
generating, by the computing system, a plurality of embeddings, each embedding of the plurality of embeddings comprising a numerical representation of a metadata text of the set of metadata text (Flunkert, figure 1, column 4 line 63 – column 5 line 9, column 6 lines 25-47, column 15 lines 25-58, column 9 line 55 – column 10 line 18, note metadata is converted into vectors, e.g., numerical representation embeddings; note combining inputs, such as time series data and metadata into an encode vector representation, e.g., numerical representation; note this is done for a plurality of metadata/embeddings);
generating, by the computing system, a plurality of vectors, each vector of the plurality of vectors comprising a time series of the set of time series each time series associated with a metadata text of the first subset of metadata text (Flunkert, figures 1 and 3, column 2 line 62 – column 3 line 38, column 9 line 55 – column 10 line 18, note vectors are generated by using the time series data which is associated with metadata; note this is done for a plurality of time series data/vectors);
generating, by the computing system, a plurality of composite embeddings (Flunkert, figures 1 and 3, column 2 line 62 – column 3 line 38, column 9 line 55 – column 10 line 18, note using input elements such as the time series data and feature metadata to generate an encoded vector, e.g., composite embedding);
determining, by the computing system, a composite embedding of the plurality of compositing embeddings based at least in part on the second subset of metadata text (Flunkert, figures 1 and 3, column 2 line 62 – column 3 line 38, column 9 line 55 – column 10 line 18, note using input elements such as the time series data and feature metadata to generate an encoded vector, e.g., composite embedding; note that since each time series has associated metadata it is interpreted to mean it is associated with the second subset of metadata text);
determining, by the computing system, a forecasted value associated with the time series (Flunkert, figures 1 and 3, column 8 lines 13-34, column 9 line 55 – column 10 line 18, note using time series data and metadata to predict a forecasted value; note forecast request).
While Flunkert teaches combining time series data and metadata to forecast values, Flunkert doesn’t specifically teach that the input time series and metadata is converted to embeddings and vector before being combined again. However, Rath is in the same field of endeavor, data analysis and information retrieval, and Rath teaches:
generating, by the computing system, a plurality of embeddings, each embedding of the plurality of embeddings comprising a numerical representation of a metadata text of the set of metadata text (Rath, figure 3C, [0041], note using the metadata to generate a magnitude to apply to a vector, the magnitude is interepted as a numerical representation embedding)
generating, by the computing system, a plurality of vectors, each vector of the plurality of vectors comprising a time series of the set of time series each time series associated with a metadata text of the first subset of metadata text (Rath, figure 3C, [0041], note converting content data to a vector. When combined with the previously cited references, the content data would be the time series data as taught by Flunkert);
generating, by the computing system, a plurality of composite embeddings based at least in part on combining each embedding of the plurality of embeddings with a respective vector of the plurality of vectors (Rath, figure 3C, [0041], note combining the magnitude, e.g. metadata embedding, to the content data vector).
determining, by the computing system, a composite embedding of the plurality of compositing embeddings based at least in part on the second subset of metadata text (Rath, figure 3C, [0041], note combining the magnitude, e.g. metadata embedding, to the content data vector. When combined with the previously cited references, the content data would be the time series data and associated metadata as taught by Flunkert)
It would have been obvious to one of ordinary skill in the art before the effective date of filing to modify the cited references to incorporate the teachings of Rath because all references are directed towards data analysis and information retrieval and because Rath would expand upon the teachings of the previously cited references in data analysis which would improve the efficiency of the system to forecast accurate data by using additional information available such as metadata.
While Flunkert as modified further teaches forecasting, Flunkert as modified doesn’t specifically teach generating, by the computing system, backcasted values for the time series based at least in part on the composite embedding of the plurality of composite embeddings; determining, by the computing system, a forecasted value associated with the time series based at least in a part on the backcasted values of the time series. However, Arik is in the same field of endeavor, data analysis and information retrieval, and Arik teaches:
determining, by the computing system, a composite embedding of the plurality of compositing embeddings based at least in part on the second subset of metadata text (Arik, [0052], note generating an encoded representation of the observed data, such as a context vector or other data structure. When combined with the previously cited references this would be for the composite embeddings and time series data as taught by Flunkert and Rath);
generating, by the computing system, backcasted values for the time series based at least in part on the composite embedding of the plurality of composite embeddings (Arik [0016, 0028], note generating backcasted values for the encoded representation of the selected time series values. When combined with the previously cited references this would be for the composite embeddings and time series data as taught by Flunkert and Rath); and
determining, by the computing system, a forecasted value associated with the time series based at least in a part on the backcasted values of the time series (Arik [0016, 0028], note generating a forecast of one or more time series future points in time using at least in part the backcast decoder machine learning model and the backcasted values. When combined with the previously cited references this would be for the composite embeddings and time series data as taught by Flunkert and Rath).
It would have been obvious to one of ordinary skill in the art before the effective date of filing to modify the cited references to incorporate the teachings of Arik because all references are directed towards data analysis and information retrieval and because Arik would expand upon the teachings of the previously cited references in data analysis which would improve the efficiency of the system to forecast accurate data by using additional information available such as the backcasted values.
While Flunkert as modified further teaches forecasting, Flunkert as modified doesn’t specifically teach generating, by the computing system, backcasted values for the time series based at least in part on the determined composite embedding of the plurality of composite embeddings that are based at least in part on combining each embedding of the plurality of embeddings with the respective vector of the plurality of vectors. However, Amiri is in the same field of endeavor, data analysis and information retrieval, and Amiri teaches:
generating, by the computing system, backcasted values for the time series based at least in part on the determined composite embedding of the plurality of composite embeddings that are based at least in part on combining each embedding of the plurality of embeddings with the respective vector of the plurality of vectors (Amiri, [0044], note combining embeddings with their respective vectors to use as input into a machine learning model to output predictions. When combined with the previously cited references this would be the time series embeddings and vectors and the predictions would be the backcasted and forecasted values as taught by Flunkert, Rath, and Arik).
It would have been obvious to one of ordinary skill in the art before the effective date of filing to modify the cited references to incorporate the teachings of Amiri because all references are directed towards data analysis and information retrieval and because Amiri would expand upon the teachings of the previously cited references in data analysis which would improve the efficiency accuracy of the system to predict data by using modeling and machine learning processes (Amiri, [0003]).
Regarding Claim 2:
Flunkert as modified shows the method as disclosed above;
Flunkert as modified further teaches:
wherein the set of metadata text comprises a table of metadata text, and wherein generating the plurality of embeddings comprises inputting each table row of metadata text into a language model to generate a respective numerical representation of the row of metadata text (Flunkert, column 6 lines 25-47, note metadata may be from a database, which is interpreted to mean that the metadata may be stored in a table) (Rath, [0048], note using language models such as Word2Vec or Doc2vecor to generate numerical embeddings).
Regarding Claim 4:
Flunkert as modified shows the method as disclosed above;
Flunkert as modified further teaches:
wherein generating the plurality of composite embeddings comprises combining a first embedding of the plurality of embeddings with a vector of the plurality of vectors, the embedding generated using a same metadata text of the set of metadata text as associated with the vector (Flunkert, figures 1 and 3, column 2 line 62 – column 3 line 38, column 9 line 55 – column 10 line 18, note using input elements such as the time series data and feature metadata to generate an encoded vector, e.g., composite embedding; note that this would be using the same metadata text associated with the time series data) (Rath, figure 3C, [0041], note combining the magnitude, e.g. metadata embedding, to the content data vector; note this would use the same metadata that is associated with the items).
It would have been obvious to one of ordinary skill in the art before the effective date of filing to modify the cited references to incorporate the teachings of Rath because all references are directed towards data analysis and information retrieval and because Rath would expand upon the teachings of the previously cited references in data analysis which would improve the efficiency of the system to forecast accurate data by using additional information available such as metadata.
Claim 8 discloses substantially the same limitations as claim 1 respectively, except claim 8 is directed to a system comprising a processor and computer readable medium (Flunkert, figure 12, note processor and memory) while claim 1 is directed to a method. Therefore claim 8 is rejected under the same rationale set forth for claim 1.
Claim 9 discloses substantially the same limitations as claim 2 respectively, except claim 9 is directed to a system comprising a processor and computer readable medium (Flunkert, figure 12, note processor and memory) while claim 2 is directed to a method. Therefore claim 9 is rejected under the same rationale set forth for claim 2.
Claim 11 discloses substantially the same limitations as claim 4 respectively, except claim 11 is directed to a system comprising a processor and computer readable medium (Flunkert, figure 12, note processor and memory) while claim 4 is directed to a method. Therefore claim 11 is rejected under the same rationale set forth for claim 4.
Claim 15 discloses substantially the same limitations as claim 1 respectively, except claim 15 is directed to a non-transitory computer-readable medium comprising a processor (Flunkert, figure 12, note processor and memory) while claim 1 is directed to a method. Therefore claim 15 is rejected under the same rationale set forth for claim 1.
Claim 16 discloses substantially the same limitations as claim 2 respectively, except claim 16 is directed to a non-transitory computer-readable medium comprising a processor (Flunkert, figure 12, note processor and memory) while claim 2 is directed to a method. Therefore claim 16 is rejected under the same rationale set forth for claim 2.
Claim 18 discloses substantially the same limitations as claim 4 respectively, except claim 18 is directed to a non-transitory computer-readable medium comprising a processor (Flunkert, figure 12, note processor and memory) while claim 4 is directed to a method. Therefore claim 18 is rejected under the same rationale set forth for claim 4.
Claim Rejections - 35 USC § 103
Claim(s) 3, 5, 10, 12, 17, and 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Flunkert in view of Rath, Arik, Amiri, and “Using Item Metadata Datasets” by Amazon (published 12/4/2022, , previously present in ‘892).
Regarding Claim 3:
Flunkert as modified shows the method as disclosed above;
Flunkert as modified further teaches:
wherein the set of metadata text comprises a table of metadata text (Flunkert, column 6 lines 25-47, note metadata may be from a database, which is interpreted to mean that the metadata may be stored in a table);
generating a vector of the plurality of vectors by combining the determined each time series of the set of time series comprising data described by the metadata text (Flunkert, figures 1 and 3, column 2 line 62 – column 3 line 38, column 9 line 55 – column 10 line 18, note vectors are generated by using the time series data which is associated with metadata; note this is done for a plurality of time series data/vectors; note when combined with the other cited references this would include the identified items with similar metadata).
While Flunkert as modified teaches metadata text from a table and generating vectors, Flunkert as modified doesn’t specifically teach matching metadata items. However, Amazon is in the same field of endeavor, data analysis and information retrieval, and Amazon teaches:
wherein the set of metadata text comprises a table of metadata text (Amazon, page 3 section “Example: Item Metadata File and Schema”, note metadata table), and wherein generating the plurality of vectors comprises: determining each column cell of a first table column comprising a same metadata text of the set of metadata text (Amazon, page 1 paragraph 3 – page 2 paragraph 3, note identifying time series items with similar metadata, which is interpreted as matching metadata text)
determining each time series of the set of time series comprising data described by the same metadata text (Amazon, page 1 paragraph 3 – page 2 paragraph 3, note identifying time series items with similar metadata); and
It would have been obvious to one of ordinary skill in the art before the effective date of filing to modify the cited references to incorporate the teachings of Amazon because all references are directed towards data analysis and information retrieval and because Amazon would expand upon the teachings of the previously cited references in data analysis which would improve the efficiency of the system to forecast accurate data by using additional information available such as metadata (Amazon, page 1 paragraph 3 – page 2 paragraph 2).
Regarding Claim 5:
Flunkert as modified shows the method as disclosed above;
Flunkert as modified further teaches:
wherein determining the forecasted value comprises: an embedding of the plurality of embeddings generated using the second subset of metadata text with the plurality of composite embeddings (Flunkert, figures 1 and 3, column 2 line 62 – column 3 line 38, column 4 line 63 – column 5 line 9, column 6 lines 25-47, column 15 lines 25-58, column 9 line 55 – column 10 line 18, note metadata is converted into vectors, e.g., numerical representation embeddings; note combining inputs, such as time series data and metadata into an encode vector representation, e.g., numerical representation; note this is done for a plurality of metadata/embeddings; note using input elements such as the time series data and feature metadata to generate an encoded vector, e.g., composite embedding) (Rath, figure 3C, [0041], note using the metadata to generate a magnitude to apply to a vector; note combining the magnitude, e.g. metadata embedding, to the content data vector); and
determining the forecasted value using a composite embedding of the plurality of composite embeddings (Flunkert, figures 1 and 3, column 8 lines 13-34, column 9 line 55 – column 10 line 18, note using time series data and metadata to predict a forecasted value; note forecast request).
It would have been obvious to one of ordinary skill in the art before the effective date of filing to modify the cited references to incorporate the teachings of Rath because all references are directed towards data analysis and information retrieval and because Rath would expand upon the teachings of the previously cited references in data analysis which would improve the efficiency of the system to forecast accurate data by using additional information available such as metadata.
While Flunkert as modified teaches forecasting using embeddings and composite embeddings, Flunkert as modified doesn’t specifically teach comparing to find the nearest neighbor. However, Amazon is in the same field of endeavor, data analysis and information retrieval, and Amazon teaches:
wherein determining the forecasted value comprises: comparing an embedding of the plurality of embeddings generated using the second subset of metadata text with the plurality of composite embeddings to determine a nearest neighbor of the embedding (Amazon, page 1 paragraph 3 – page 2 paragraph 3, note identifying time series items with similar metadata, which is interpreted as matching metadata text, to identify its nearest neighbor to generate forecast. When combined with the previously cited references this would be for the metadata embeddings and composite embeddings); and
determining the forecasted value using a composite embedding of the plurality of composite embeddings that is the nearest neighbor of the embedding of the plurality of embedding (Amazon, page 1 paragraph 3 – page 2 paragraph 3, note identifying time series items with similar metadata, which is interpreted as matching metadata text, to identify its nearest neighbor to generate forecast. When combined with the previously cited references this would be for the metadata embeddings and composite embeddings).
It would have been obvious to one of ordinary skill in the art before the effective date of filing to modify the cited references to incorporate the teachings of Amazon because all references are directed towards data analysis and information retrieval and because Amazon would expand upon the teachings of the previously cited references in data analysis which would improve the efficiency of the system to forecast accurate data by using additional information available such as metadata (Amazon, page 1 paragraph 3 – page 2 paragraph 2).
Claim 10 discloses substantially the same limitations as claim 3 respectively, except claim 10 is directed to a system comprising a processor and computer readable medium (Flunkert, figure 12, note processor and memory) while claim 3 is directed to a method. Therefore claim 10 is rejected under the same rationale set forth for claim 3.
Claim 12 discloses substantially the same limitations as claim 5 respectively, except claim 12 is directed to a system comprising a processor and computer readable medium (Flunkert, figure 12, note processor and memory) while claim 5 is directed to a method. Therefore claim 12 is rejected under the same rationale set forth for claim 5.
Claim 17 discloses substantially the same limitations as claim 3 respectively, except claim 17 is directed to a non-transitory computer-readable medium comprising a processor (Flunkert, figure 12, note processor and memory) while claim 3 is directed to a method. Therefore claim 17 is rejected under the same rationale set forth for claim 3.
Claim 19 discloses substantially the same limitations as claim 5 respectively, except claim 19 is directed to a non-transitory computer-readable medium comprising a processor (Flunkert, figure 12, note processor and memory) while claim 5 is directed to a method. Therefore claim 19 is rejected under the same rationale set forth for claim 5.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Wong (US2022/0138552) and Herzog (US2017/0206452) teach generating backcasted values and using those values for forecasting.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JOHN J MORRIS whose telephone number is (571)272-3314. The examiner can normally be reached M-F 6:00-2:00 PM EST.
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, Neveen Abel-Jalil can be reached at 571-270-0474. 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.
/JOHN J MORRIS/Examiner, Art Unit 2152 2/20/2026
/NEVEEN ABEL JALIL/Supervisory Patent Examiner, Art Unit 2152