DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
This action is responsive to communications: Application filed on 5/29/2024.
Claims 1-20 are pending. Claims 1, 8, and 15 are independent.
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, 4-9, 11-16, and 18-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Curtis et al. (US12,158,880) in view of Shluzas et al. (US2022/0301666A1).
In regards to claim 1, Curtis et al. discloses a method for training a time series classification model, the method comprising:
transforming time-series ground-truth data into continuous data (Curtis et al. col26 ln19-32, check that there are sufficiently many unique values in the series to justify treating the data as a numeric and continuous-valued metric (e.g., if there are only the values 0 and 1, the variable is binary and not suited to time-series analysis), which involves binning the values and looking at the distribution of the bin counts; check that the range of the time-series' values does not span too many orders of magnitude; check that the series is relatively smooth, has some degree of structure and predictability, and is distinguishable from sheer random noise);
training a time-series forecasting model to predict future values based on the transformed continuous data, yielding a forecast output (Curtis et al. fig. 3 306 col15 ln6-20, data used in forecasting machine learning model to predict points in dataset based on prior points);
projecting the forecast output into a two-dimensional representation of estimated continuous data (Curtis et al. fig. 2B, col14 ln39-59, generates envelope 252 representing values of the data stream that have been determined to be benign);
training a residual model to determine residuals between the continuous data and the estimated continuous data (Curtis et al. col18 ln50-55, determines difference between predicted value and actual value in dataset);
integrating the determined residuals into embeddings of the time-series forecasting model (Curtis et al. col23 ln21-37, integrates residuals to adjust the sensitivity threshold).
Curtis et al. does not explicitly disclose retraining the time-series forecasting model with the embedded residuals to minimize cross-entropy loss associated with the time-series forecasting model; and
upon convergence of the cross-entropy loss, outputting a trained time-series forecasting model configured to predict classifications for time-series data.
However Shluzas et al. substantially discloses retraining the time-series forecasting model with the embedded residuals to minimize cross-entropy loss associated with the time-series forecasting model (Shluzas et al. para[0222]-[0223], retrains model based on cross-entropy loss to reduce loss from class imbalance) ; and
upon convergence of the cross-entropy loss, outputting a trained time-series forecasting model configured to predict classifications for time-series data (Shluzas et al. fig. 22 para[0204], rescaling can be applied to bound the domain of variable to smaller numbers to help models converge sooner. Model evaluation can be performed by an automated system or with human intervention, or both).
It would have been obvious to one of ordinary skill in the art before he filing date of the invention to have combined the forecasting method of Curtis et al. with the monitoring method of Shluzas et al. in order to identify predicted risks from time-series data (Shluzas et al. para[0082]).
In regards to claim 2, Curtis et al. as modified by Shluzas et al. discloses the method of claim 1, wherein the classifications predicted by the trained model are binary and comprise normal data and abnormal data (Curtis et al.col19 ln20-35, flags anomalous data based on heuristic assessment).
In regards to claim 4, Curtis et al. as modified by Shluzas et al. discloses the method of claim 1, further comprising projecting residuals from the model to a two-dimensional space (Curtis et al. fig. 5 col23 ln15-20, displays plot of dataset and envelope determined based on the sensitivity threshold).
In regards to claim 5, Curtis et al. as modified by Shluzas et al. discloses the method of claim 4, wherein the projecting of residuals includes utilizing a numerical optimization to align the projected residuals with the two-dimensional representation of estimated continuous data values (Curtis et al. col 32 ln46-64, list of values is analyzed according to an assessment according to the gap in error values of a group of three neighboring points).
In regards to claim 6, Curtis et al. as modified by Shluzas et al. discloses the method of claim 1, further comprising adjusting weights of a cross-entropy loss function based on class frequency to mitigate effects of class imbalance during the retraining of the time-series forecasting model (Shluzas et al. para[0222]-[0223], factor added to standard cross entropy criterion to account for focal loss due to class imbalance).
In regards to claim 7, Curtis et al. as modified by Shluzas et al. discloses the method of claim 1, wherein the time series classification model utilizes a transformer-based architecture to perform sequence-to-sequence prediction tasks (Curtis et al. col37 ln7-28, Transformation of data into events can include event generation and event indexing).
Claims 8-9, and 11-14 recite substantially similar limitations to claims 1-2, and 4-7. Thus claims 8-9 and 11-14 are rejected along the same rationale as claims 1-2 and 4-7.
Claims 15-16 and 18-20 recite substantially similar limitations to claims 1-2 and 4-6. Thus claims 15-16 and 18-20 are rejected along the same rationale as claims 1-2 and 4-6.
Claim(s) 3, 10, and 17 is/are rejected under 35 U.S.C. 103 as being unpatentable over Curtis et al. in view of Shluzas et al. and Woo et al. (US2023/0244947).
In regards to claim 3, Curtis et al. as modified by Shluzas et al. discloses the method of claim 1. Curtis et al. does not explicitly disclose wherein the transforming includes utilizing exponential smoothing that assigns weights to data points in the time-series ground-truth data for calculating the continuous data.
However Woo et al. discloses wherein the transforming includes utilizing exponential smoothing that assigns weights to data points in the time-series ground-truth data for calculating the continuous data (Woo et al. para[0047], the exponential smoothing attention layer may be a weighted average with weights that decrease exponentially looking back further in the sequence).
It would have been obvious to one of ordinary skill in the art before the filing date of the invention to have combined the of Curtis et al. with the time-series forecasting method of Woo et al. in order to identify a pattern to time-series data without overfitting to the noise (Woo et al. para[0003]).
Claim 10 recites substantially similar limitations to claim 3. Thus claim 10 is rejected along the same rationale as claim 3.
Claim 17 recites substantially similar limitations to claim 3. Thus claim 17 is rejected along the same rationale as claim 3.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Catledge et al. (US2025/0256864) teaches transforming discrete values into continuous, high dimensional vectors.
Singh et al. (US2024/0005177) teaches using machine learning to generate a time-series forecast.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to NICHOLAS HASTY whose telephone number is (571)270-7775. The examiner can normally be reached Monday-Friday 8:30am-5:00pm.
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, Matt Ell can be reached at (571)270-3264. 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.H/Examiner, Art Unit 2141
/MATTHEW ELL/Supervisory Patent Examiner, Art Unit 2141
Prosecution Insights