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 Arguments
Applicant’s arguments, see Remarks, filed January 5, 2026, with respect to the rejection of claims 12 – 19 and 21 under 35 U.S.C. § 103 have been fully considered and are persuasive. The claims have been amended with limitations not taught by the previously relied upon art. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground(s) of rejection is made in view of new art found in a search of the prior art prompted by amendments.
Applicant’s remarks indicate that claim 15 was intended to be cancelled in the amendments. However, in the amendments submitted January 5, 2026, claim 15 is not cancelled. In the interest of compact prosecution, the amendments have entered as submitted.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(d):
(d) REFERENCE IN DEPENDENT FORMS.—Subject to subsection (e), a claim in dependent form shall contain a reference to a claim previously set forth and then specify a further limitation of the subject matter claimed. A claim in dependent form shall be construed to incorporate by reference all the limitations of the claim to which it refers.
The following is a quotation of pre-AIA 35 U.S.C. 112, fourth paragraph:
Subject to the following paragraph [i.e., the fifth paragraph of pre-AIA 35 U.S.C. 112], a claim in dependent form shall contain a reference to a claim previously set forth and then specify a further limitation of the subject matter claimed. A claim in dependent form shall be construed to incorporate by reference all the limitations of the claim to which it refers.
Claim 15 is rejected under 35 U.S.C. 112(d) or pre-AIA 35 U.S.C. 112, 4th paragraph, as being of improper dependent form for failing to further limit the subject matter of the claim upon which it depends, or for failing to include all the limitations of the claim upon which it depends. Claim 15 recites “wherein the sensor data and training data are formed as image data”. Claim 12 recites “wherein the sensor data and training data are formed as image data of manufactured components”. The independent claim 12 is narrower than the dependent claim 15, thus claim 15 fails to further limit the subject matter of claim 12. Applicant may cancel the claim, amend the claim to place the claim in proper dependent form, rewrite the claim in independent form, or present a sufficient showing that the dependent claim complies with the statutory requirements.
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.
Claims 12, 13, 14, 18 and 21 are rejected under 35 U.S.C. 103 as being unpatentable over Zhang (NPL, Federated Learning for Internet of Things: A Federated Learning Framework for On-device Anomaly Data Detection, cited in previous action), in view of Williams (NPL, Federated Learning: Private Distributed Machine Learning with Data Locality, cited in previous action), further in view of Mehr et al. (US Patent Application Publication 2018/0341248), hereinafter Mehr.
Regarding claim 12, Zhang teaches a computer-implemented method for anomaly detection in sensor data (Abstract), the method comprising:
a) generating and training a first and at least one second local model (Fig. 1) based on an autoencoder (Section 2.3. Anomaly Data Detection with Deep AutoEncoder), each model comprising local model weightings (Fig. 2) and a local model output variable (Section 2.3 subsection Model Definition, the mean square error/reconstruction error), and determining a local threshold for a respective model output variable aided by at least one of (i) a mean value and (ii) a standard deviation of local data values via a first or at least one second client (Algorithm 1, a personalized threshold is computed based on a mean of mean square errors and a standard deviation (note the lowercase sigma); Section 2.4 subsection Global and Personalized Threshold.);
b) transmitting the local model weightings from the first and the at least one second client to a server (Algorithm 1, each client uploads their weights to the server);
c) generating and training a global model based on the autoencoder (Section 2.3. Anomaly Data Detection with Deep AutoEncoder), utilizing the local model weightings, the global model comprising global model weightings and a local model output variable (Algorithm 1, the global model is trained utilizing an average of local model weightings in update line 8), and determining a global threshold value for a global model output variable aided by at least one of (i) the mean value and (ii) the standard deviation of the local threshold values by the server (Algorithm 1, while indirectly computed, the global threshold value is computed using a mean and standard deviation of local MSEs, which, under the broadest reasonable interpretation, is aided by mean values and standard deviations of the local thresholds);
d) transmitting the global model weightings and global threshold value for the global model output variable to the first client, and adopting the global model weightings for the first local model of the first client (Algorithm 1, the clients receive the new model from the server);
e) capturing first sensor data by a first sensor possessed by the first client (Abstract);
f) applying the first sensor data to the first local model and determining a local model output variable of the first client (Algorithm 1, the local model is trained locally; Section 3 Experiments, the local model is actually run and tested); and
g) detecting an anomaly for the sensor data by the first client, if the local model output variable of the first client is outside a range which is fixed by the global threshold value for the global model output variable (Section 2.3 subsection Anomaly Detection, any sample with a higher reconstruction error than the threshold is detected as an anomaly; Section 3.4 subsection Evaluation using the global threshold, this can be done using the global threshold).
Zhang does not explicitly state, as claimed in b), that the local threshold values are transmitted from the first and the at least one second client to a server (The personalized thresholds are described as being local to the client, and are not explicitly transmitted to the server).
Williams teaches that, in federated averaging, local values are transmitted from clients to a server (Page 14), and that federated averaging can be applied to any model in which the average is meaningful (Page 24).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention that local thresholds could have been transmitted to the server along with the model weightings as taught by Williams. It would have been obvious because the threshold for detecting an anomaly is, in effect, a part of the machine learning model. It would be clear that the average of thresholds in Zhang is meaningful as shown in Zhang’s equation for computing the global threshold, which involves an average of data from the clients. It would be obvious to one of ordinary skill in the art to try different methods of computing the global anomaly detection threshold to best suit their task, and in their experimentation, treating the local threshold as any other parameter would be a clear direction to try. There would be a clear expectation of success, as federated averaging is applicable to any model in which the average is meaningful (Williams page 24).
Zhang and Williams do not explicitly teach that the sensor data is obtained during an industrial manufacturing process, nor that the sensor data and training data are formed as image data of manufactured components.
Mehr teaches using federated learning (Paragraphs 0003, 0034, 0126, 0153, 0153 and 0164, there may be a plurality of distributed learning systems sharing data) in an industrial manufacturing process (Paragraph 0035), and that the sensor data and training data are formed as image data of manufactured components (Paragraphs 0039, 0121 – 0123, 0164 and Figs. 7A – 7C, the systems described in Mehr may use image data). Mehr is also directed to detecting defects in objects (Paragraph 0006).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention that the federated learning system of Zhang could be applied to an industrial manufacturing process, and that the sensor data and training data would be formed as image data of manufactured components, as taught by Mehr. It would be obvious because machine learning methods are well-known in the art for image analysis tasks including automated object defect classification in manufacturing (Mehr paragraphs 0033, the variety of machine learning algorithms discussed are known to those of skill in the art). These known methods provide advantages in a number of industries (Mehr paragraph 0035). Mehr demonstrates that known machine learning methods, including federated learning, are applicable to detecting defects in image data of manufactured components. It would be clear to one of ordinary skill in the art that the federated learning system of Zhang would, similarly to the other learning systems discussed in Mehr, be applicable to detecting defects in image data of manufactured components.
Regarding claim 13, Zhang in view of Williams and Mehr teaches the method as claimed in claim 12, wherein said training of a respective local model is performed with training data which is assignable to an anomaly-free state in the sensor data (Zhang section 2.3 subsection Anomaly Detection, the AutoEncoder is trained on benign data).
Regarding claim 14, Zhang in view of Williams and Mehr teaches the method as claimed in claim 12, wherein a respective model output is formed by at least one parameter value (Zhang Section 2.3, the reconstruction error is formed from outputs of the autoencoder and by variables), and a respective threshold value is defined by at least one corresponding assignable value of a range limit of a range (Zhang Section 2.3, the threshold tr is defined by a variable, and because a threshold is a minimum/maximum needed to make a consideration, it is also a range limit of a range).
Regarding claim 15, Zhang in view of Williams and Mehr teaches the method as claimed in claim 12, wherein the sensor data and training data are formed as image data (Mehr paragraphs 0039, 0121 – 0123, 0164 and Figs. 7A – 7C, the systems described in Mehr may use image data).
Regarding claim 18, Zhang teaches a system for anomaly detection in sensor data, comprising:
a first and at least one second client each having a client processor and a client memory (Section 3.1 Setup, the 9 IoT devices);
a sensor (Section 2.2 Dataset and Preprocessing); and
a connected server having a server processor and a server memory (Section 3.1 Setup, the server);
wherein the system is configured to perform a method.
The method recites similar language to the method of claim 1, and is similarly rejected under Zhang in view of Williams and Mehr.
Claim 19 recites similar language to claim 15, and is similarly rejected.
Regarding claim 21, Zhang in view of Williams and Mehr teaches a non-transitory electronically readable data carrier encoded with readable control information comprising at least a computer program which, when using the data carrier in a computing facility (Zhang Fig. 1, Section 2.5 FedIoT System Design, the program is used in the IoT devices/FL server), implements the method as claimed in claim 12.
Claim 16 is rejected under 35 U.S.C. 103 as being unpatentable over Zhang in view of Williams and Mehr as applied to claim 12 above, and further in view of Ouyang (NPL, ClusterFL: A Similarity-Aware Federated Learning System for Human Activity Recognition, cited in previous action).
Regarding claim 16, Zhang in view of Williams and Mehr teaches the method as claimed in claim 12, wherein the global model weightings and the global threshold are transmitted from the server to the at least one second client, (Zhang Algorithm 1, the weightings are transmitted to the clients) which has an autoencoder (Zhang section 2.3 Anomaly Data Detection with Deep AutoEncoder) with a further local model (Zhang section 2.4 FedDetect: A Generic Federated Learning Framework for IoT Anomaly Data Detection, the autoencoder is trained locally and has a similar structure).
Zhang in view of Williams and Mehr does not explicitly teach that similarity is determined via predefined ranges for the local model weightings between the first and at least one second clients (No determination of similarity is explicitly made).
Ouyang teaches a method for federated learning wherein similarity is determined via predefined ranges for the local model weightings between the first and at least one second clients (Page 59 column 1, the similarity is measured by a predefined distance equation for the local model weightings, measuring the distance between the weightings of each client).
It would be obvious to one of ordinary skill in the art before the effective filing date of the invention to use a distance-based similarity measure to determine whether clients are similar. It would be obvious because clustering clients based on similarity advantageously mitigates issues of data heterogeneity (Ouyang page 56), which is an issue Zhang similarly aims to overcome (Zhang Section 2.2 Dataset and Preprocessing column 2, real world data is inconsistent). It would be clear to one of ordinary skill in the art that overcoming this heterogeneity is a core problem in the field of federated learning, and as such, clustering solutions such as Ouyang are a well-explored avenue of modifying federation schemes.
Claim 17 is rejected under 35 U.S.C. 103 as being unpatentable over Zhang in view of Williams and Mehr as applied to claim 12 above, and further in view of Li (NPL, base.py, cited in previous action).
Regarding claim 17, Zhang in view of Williams and Mehr teaches the method as claimed in claim 12.
Zhang in view of Williams and Mehr as applied to claim 12 above teaches the method as claimed in claim 12, wherein for a respective local threshold value, metadata with respect to the local threshold value and the first and at least one second local models is also acquired, and when generating and training the global model, the acquired metadata is applied when weighting individual model weightings (Zhang section 3.4 subsection Understanding the result, the aggregation is weighted with devices that have a lower number of samples, the number of samples being the metadata).
Zhang in view of Williams and Mehr as applied to claim 12 above does not, however, explicitly state that the metadata is acquired by the first or at least one second client, and transmitted to the server (The combination as applied above does not explicitly mention which entity acquires the metadata and transmits it).
Li teaches an implementation of federated learning in which the clients transmit the metadata applied when weighting individual model weightings to the server (Function local_train, the variables in solns are each returned from the client local_train functions, solns including the number of samples of the client; function aggregate, the weighted sum is taken using the number of samples in each item of solns).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention that the clients provide the metadata for weighting the individual model weightings. It would be obvious because there are a finite number of possible means to acquire the metadata, either the clients or the server. Furthermore, errors may occur within individual clients that cause them to lag behind or drop out entirely (Williams page 21). An implementation where individual models transmit metadata would advantageously allow the server to continue training normally if the number of samples actually trained on by the client were to somehow change from the expected number. It would be clear to one of ordinary skill in the art that an implementation utilizing the client to acquire metadata would be an effective option to try.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Von Rundstedt et al. (WIPO Document WO 2021/122616 A1) teaches an image fault detection system for manufacturing, but is not relied upon as it focuses on plant images. Kim et al. (US Patent Application Publication 2020/0005071) similarly teaches a federated image analysis system that is similar to the claims, but is related to identifying business cards. Zwick et al. (US Patent Application Publication 2002/0131633) demonstrates that machine vision for detection in manufacturing is well-understood in the art, but is not relied upon as it is not a federated learning system.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to BRYAN PAI SONG HUANG whose telephone number is (571)272-0510. The examiner can normally be reached Monday - Friday 11:30 AM - 8:30 PM.
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, ASHISH THOMAS can be reached at (571) 272-0631. 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.
/B.P.H./Examiner, Art Unit 2114
/ASHISH THOMAS/Supervisory Patent Examiner, Art Unit 2114