DETAILED ACTION
NOTE: This office action replaces the Non-Final Rejection with receipt date of 08/26/2026. The purpose of the new office action is only to remove the 35 USC 101 rejection which was improper.
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 .
Information Disclosure Statement
The information disclosure statements (IDS) submitted on 07/15/2026 and 01/07/2025 have been considered by the examiner.
Priority
Receipt is acknowledged of certified copies of papers submitted under 35 U.S.C. 119(a)-(d), of which papers have been placed in the file wrapper.
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1, 4-6, and 9-10 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Zhang et al. (US Pub. No. 2021/0365781 A1).
Regarding claim 1, Zhang discloses, an estimation device comprising: a storage medium configured to store computer-readable instructions; (See Zhang ¶158, “As also illustrated in FIG. 7, instructions for the computer, e.g., executable code, may be stored on a computer readable medium 700.”)
and a processor connected to the storage medium, (See Zhang ¶65, “The device or apparatus may comprise one or more microprocessors which execute appropriate software.”)
wherein the processor executes the computer-readable instructions to acquire, with data as an input, a logit that at least a portion of the data corresponds to a class that represents a certain type by inputting target data to a machine learning model learned to output the logit, (See Zhang ¶81, “More specifically, the output logit vector may be the output of a layer of the DNN preceding the last activation layer of the trained classification model TCM 310. The last activation layer may be, for example, a softmax activation layer or a sigmoid activation layer. For example, if the trained classification model TCM 310 is trained to classify a sample into one or more of K classes, the output logit vector λ may be λ = {λ.sub.0, . . ., λ.sub.K}. Typically, a prediction probability is derived from an output logit vector.”)
calculate a correction value for correcting the logit (See Zhang ¶83-84, “The finetuning submodule FT 320 may comprise a finetuning model, such as a matrix scaling model, a Dirichlet calibration model, temperature scaling, vector scaling or a Beta model, although this list is not exclusive. The finetuning model may be a parametric model, which accepts the output logit vector A as input. In some examples, a Gaussian Process may be used, although a parametric model is preferable in order to reduce complexity.
The matrix scaling model and the Dirichlet calibration model are shown below:
λ.sub.Mtx = Wλ+b (Eq. 1)
λ.sub.Dir = W log softmax(λ)+b (Eq. 2)
where W is a square matrix and b is a bias vector. Both W and b are trainable parameters of the finetuning model.”)
using an output of the machine learning model, (See Zhang ¶86, “In order to train parameters of the finetuning model, a finetuning loss function FTF 350 may be used. The finetuning loss function FTF 350 may be based on a negative log-likelihood (NLL) loss, a Brier score loss, a hinge loss or the like. … and {tilde over (q)}.sub.k,s denotes the prediction probability after finetuning the logits. The model parameters W and b can be determined by minimizing the NLL loss averaged over all validation samples.” Where the prediction probability is an output of the machine learning classifier.)
and correct the logit on the basis of the calculated correction value, (See Zhang ¶84, “The matrix scaling model and the Dirichlet calibration model are shown below:
λ.sub.Mtx = Wλ+b (Eq. 1); λ.sub.Dir = W log softmax(λ)+b (Eq. 2).”)
and estimate a class to which at least a portion of the target data corresponds on the basis of the corrected logit. (See Zhang ¶156, “In some embodiments, the classification, including the adjusted prediction probability, may be output by the calibration module. In some embodiments, the adjusted prediction probability may be output by the calibration module whilst the classification decision (e.g., the most probable class or top-k classes, etc.) may be output by the classification model.”)
Regarding claim 4, Zhang discloses, the estimation device according to the estimation device according to wherein the processor corrects the logit by adding the correction value to the logit and estimates a class where a probability value is maximized based on the corrected logit as the class to which at least a portion of the target data corresponds. (See Zhang ¶84, “The matrix scaling model and the Dirichlet calibration model are shown below:
λ.sub.Mtx = Wλ+b (Eq. 1)
λ.sub.Dir = W log softmax(λ)+b (Eq. 2)
where W is a square matrix and b is a bias vector.” Where the bias vector is considered to be the correction value.)
Regarding claim 5, Zhang discloses, the estimation device according to the estimation device according to wherein the data is an image including a plurality of pixels, and at least a portion of the data is one or more groups of pixels of the image. (See Zhang ¶152, “The input sample may be obtained from a sensor coupled to the system carrying out the method. For example, the input sample may be an image obtained from a camera. However, this is not a limitation. The input sample may be an audio sample, a video, lidar, radar, text, or the like, or a combination thereof.”)
Regarding claim 6, Zhang discloses, the estimation device according to the estimation device according to wherein the data is a vocal sound, and at least a portion of the data is a section of the vocal sound. (See Zhang ¶152, “The input sample may be obtained from a sensor coupled to the system carrying out the method. For example, the input sample may be an image obtained from a camera. However, this is not a limitation. The input sample may be an audio sample, a video, lidar, radar, text, or the like, or a combination thereof.”)
Regarding claim 9, Zhang discloses, an estimation method comprising: by a computer, (See Zhang ¶65, “The device or apparatus may comprise one or more microprocessors which execute appropriate software.”)
acquiring, with data as an input, a logit that at least a portion of the data corresponds to a class that represents a certain type by inputting target data to a machine learning model learned to output the logit; (See Zhang ¶81, “More specifically, the output logit vector may be the output of a layer of the DNN preceding the last activation layer of the trained classification model TCM 310. The last activation layer may be, for example, a softmax activation layer or a sigmoid activation layer. For example, if the trained classification model TCM 310 is trained to classify a sample into one or more of K classes, the output logit vector λ may be λ = {λ.sub.0, . . ., λ.sub.K}. Typically, a prediction probability is derived from an output logit vector.”)
calculating a correction value for correcting the logit (See Zhang ¶83-84, “The finetuning submodule FT 320 may comprise a finetuning model, such as a matrix scaling model, a Dirichlet calibration model, temperature scaling, vector scaling or a Beta model, although this list is not exclusive. The finetuning model may be a parametric model, which accepts the output logit vector A as input. In some examples, a Gaussian Process may be used, although a parametric model is preferable in order to reduce complexity.
The matrix scaling model and the Dirichlet calibration model are shown below:
λ.sub.Mtx = Wλ+b (Eq. 1)
λ.sub.Dir = W log softmax(λ)+b (Eq. 2)
where W is a square matrix and b is a bias vector. Both W and b are trainable parameters of the finetuning model.”)
using an output of the machine learning model; (See Zhang ¶86, “In order to train parameters of the finetuning model, a finetuning loss function FTF 350 may be used. The finetuning loss function FTF 350 may be based on a negative log-likelihood (NLL) loss, a Brier score loss, a hinge loss or the like. … and {tilde over (q)}.sub.k,s denotes the prediction probability after finetuning the logits. The model parameters W and b can be determined by minimizing the NLL loss averaged over all validation samples.” Where the prediction probability is an output of the machine learning classifier.)
and correcting the logit on the basis of the calculated correction value; (See Zhang ¶84, “The matrix scaling model and the Dirichlet calibration model are shown below:
λ.sub.Mtx = Wλ+b (Eq. 1); λ.sub.Dir = W log softmax(λ)+b (Eq. 2).”)
and estimating a class to which at least a portion of the target data corresponds on the basis of the corrected logit. (See Zhang ¶156, “In some embodiments, the classification, including the adjusted prediction probability, may be output by the calibration module. In some embodiments, the adjusted prediction probability may be output by the calibration module whilst the classification decision (e.g., the most probable class or top-k classes, etc.) may be output by the classification model.”)
Regarding claim 10, Zhang discloses, a computer-readable non-transitory storage medium that stores a program causing a computer to execute: (See Zhang ¶158, “As also illustrated in FIG. 7, instructions for the computer, e.g., executable code, may be stored on a computer readable medium 700.”)
acquiring, with data as an input, a logit that at least a portion of the data corresponds to a class that represents a certain type by inputting target data to a machine learning model learned to output the logit, (See Zhang ¶81, “More specifically, the output logit vector may be the output of a layer of the DNN preceding the last activation layer of the trained classification model TCM 310. The last activation layer may be, for example, a softmax activation layer or a sigmoid activation layer. For example, if the trained classification model TCM 310 is trained to classify a sample into one or more of K classes, the output logit vector λ may be λ = {λ.sub.0, . . ., λ.sub.K}. Typically, a prediction probability is derived from an output logit vector.”)
calculating a correction value for correcting the logit (See Zhang ¶83-84, “The finetuning submodule FT 320 may comprise a finetuning model, such as a matrix scaling model, a Dirichlet calibration model, temperature scaling, vector scaling or a Beta model, although this list is not exclusive. The finetuning model may be a parametric model, which accepts the output logit vector A as input. In some examples, a Gaussian Process may be used, although a parametric model is preferable in order to reduce complexity.
The matrix scaling model and the Dirichlet calibration model are shown below:
λ.sub.Mtx = Wλ+b (Eq. 1)
λ.sub.Dir = W log softmax(λ)+b (Eq. 2)
where W is a square matrix and b is a bias vector. Both W and b are trainable parameters of the finetuning model.”)
using an output of the machine learning model, (See Zhang ¶86, “In order to train parameters of the finetuning model, a finetuning loss function FTF 350 may be used. The finetuning loss function FTF 350 may be based on a negative log-likelihood (NLL) loss, a Brier score loss, a hinge loss or the like. … and {tilde over (q)}.sub.k,s denotes the prediction probability after finetuning the logits. The model parameters W and b can be determined by minimizing the NLL loss averaged over all validation samples.” Where the prediction probability is an output of the machine learning classifier.)
and correcting the logit on the basis of the calculated correction value, (See Zhang ¶84, “The matrix scaling model and the Dirichlet calibration model are shown below: λ.sub.Mtx = Wλ+b (Eq. 1); λ.sub.Dir = W log softmax(λ)+b (Eq. 2).”)
and estimating a class to which at least a portion of the target data corresponds on the basis of the corrected logit. (See Zhang ¶156, “In some embodiments, the classification, including the adjusted prediction probability, may be output by the calibration module. In some embodiments, the adjusted prediction probability may be output by the calibration module whilst the classification decision (e.g., the most probable class or top-k classes, etc.) may be output by the classification model.”)
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 set forth in Graham v. John Deere Co., 383 U.S. 1, 148 USPQ 459 (1966), that are applied 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.
Claims 7-8 are rejected under 35 U.S.C. 103 as being unpatentable over Zhang et al. (US Pub. No. 2021/0365781 A1) in view of Hoogeboom et al. (US Pub. No. 2022/0101050 A1).
Regarding claim 7, Zhang discloses, the estimation device according to the estimation device according to claim 1, but he fails to disclose a vehicle control device comprising, wherein the processor controls traveling of a vehicle on the basis of a result of estimation by the estimation device.
However, Hoogeboom discloses vehicle control device comprising, wherein the processor controls traveling of a vehicle on the basis of a result of estimation by the estimation device. (See Hoogeboom ¶73, “At least if the conformance value indicates a sufficient reliability of the image classifier for the input image, the output of the image classifier may be used for controlling the vehicle. For example, the system may use the output of the image classifier for steering the wheels 42 of the vehicle, e.g., for keeping the vehicle in lane.”)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include the vehicle that is controlled by output from an image classifier as suggested by Hoogeboom to Zhang’s image classifier. This can be done using known engineering techniques, with a reasonable expectation of success. The motivation for doing so is because using an image classifier for vehicle driving enables computer vision systems to interpret real-time road conditions, detect obstacles, and make instant navigation choices from visual inputs.
Regarding claim 8, Zhang and Hoogeboom disclose, a vehicle comprising: the vehicle control device according to claim 7. (See Hoogeboom ¶32, “Optionally, the input image is an image of an environment of a vehicle, for example an autonomous or semi-autonomous vehicle. The image may be obtained from a camera of the vehicle, and at least if the conformance value indicates a sufficient reliability of the image classifier for the input image, an output of the image classifier may be used to control the vehicle.”)
Allowable Subject Matter
Claim 2-3 are objected to as being dependent upon a rejected base claim but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
Regarding claim 2, the estimation device according to the estimation device according to wherein the processor calculates the correction value using a preset hyperparameter and an output of the machine learning model. (The disclosed prior art of record fails to disclose the limitations of this claim.)
Regarding claim 3, this claim is objected to since it depends on objected to claim 2.
Conclusion
Listed below are the prior arts made of record and not relied upon but are considered pertinent to applicant’s disclosure.
Menon et al. (US Pub. No. 2023/0017505 A1) Methods, systems, and apparatus, including computer programs encoded on computer storage media, for accounting for long-tail training data.
Qi et al. (US Pub. No. 2024/0330705 A1) Example aspects of the present disclosure provide a novel, resource-efficient approach for learning image representation with federated learning, which can be referred to as federated sampled SoftMax. According to example aspects of the present disclosure, the federated learning clients sample a set of negative classes and optimize only the corresponding model parameters with respect to a sampled SoftMax objective that approximates the global full SoftMax objective.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DAVID PERLMAN whose telephone number is (571) 270-1417.
The examiner can normally be reached on Monday - Friday; 10:00am -6:30pm.
Examiner interviews are available via telephone 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, Chineyere Wills-Burns can be reached at (571) 272-9752. 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.
/DAVID PERLMAN/Primary Examiner, Art Unit 2673
Prosecution Insights