DETAILED ACTION
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 .
Claim Interpretation 112(f)
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following is a quotation of pre-AIA 35 U.S.C. 112, sixth paragraph:
An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier. Such claim limitation(s) is/are:
As to claims 1-11, 13-16, and 18, the “determination unit” is considered to read on a computer for operating the determination process (Specification as filed: [0011]; PGPUB: [0019]).
As to claims 1-11 and 13, the “generation unit” is considered to read on a computer for operating the generation process (Specification as filed: [0011]; PGPUB: [0019]).
As to claims 5 and 11, the “setting unit” is considered to read on a computer for operating the setting process (Specification as filed: [0011]; PGPUB: [0019]).
As to claims 7, the “training unit” is considered to read on a computer with a processor for operating the training process (Specification as filed: [0011]; PGPUB: [0019]).
As to claims 14-16 and 18, the “prediction unit” is considered to read on a computer with a processor for operating the prediction process (Specification as filed: [0011]; PGPUB: [0019]).
As to claims 14 and 15, the “notification unit” is considered to read on a computer with a processor for operating the notification process (Specification as filed: [0011]; PGPUB: [0019]).
Because this/these claim limitation(s) is/are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
If applicant does not intend to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 13 and 18 are rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter. The claim(s) does/do not fall within at least one of the four categories of patent eligible subject matter because the claim a “computer program” which is not directed towards statutory subject matter. One way to overcome this rejection would be to claim the computer with a computer program stored thereon, or something in a comparable fashion. Appropriate correction is required.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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 nonobviousness.
Claim(s) 1, 2, 6-8, 12, and 13 are rejected under 35 U.S.C. 103 as being unpatentable over el Kaliouby et al., US 2022/0101146 A1 (Kaliouby).
Regarding claim 1, Kaliouby teaches an information processing device (electronic device or a computing device) ([0025]) comprising:
a determination unit which classifies training data (classifying facial images) ([0023-0024]) for training a model (training a neural network using datasets called training datasets) ([0023]) for each attribute (demographic information) ([0023]) and
determines presence or absence of bias of the training data (detecting bias within a training dataset) ([0024] and [0035]) due to a difference in attributes (based on the demographic groups being disproportionate) ([0027], [0030], and [0035]); and
a generation unit which automatically generates training data of a minor attribute and mitigates the bias when the determination unit determines that there is the bias (when bias is detected within the training dataset, augmentation techniques can be used to mitigate the bias; the augmenting can include supplementing the training dataset with additional images such as facial images that represent underrepresented demographic groups) ([0035]).
Although Kaliouby does not explicitly state determining the “absence” of bias, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention that since Kaliouby teaches detecting bias (Abstract and [0027]), that when it doesn’t detect bias there is obviously an absence of bias.
Regarding claim 2, Kaliouby teaches wherein the determination unit determines at least one of race, gender, or another sensitive attribute (wherein the demographic groups can be based on age bands, ethnicity, gender, etc.) ([0006]).
Regarding claim 6, Kaliouby teaches wherein the generation unit automatically generates training data of a minor attribute (supplementing the training dataset with additional images such as facial images that represent underrepresented demographic groups) ([0035]) by using a generative adversarial network (GAN) (wherein the additional images can be generated using a generative adversarial network (GAN)) ([0036]).
Regarding claim 7, Kaliouby teaches further comprising a training unit which trains the model by using training data (training a neural network using datasets called training datasets) ([0023]) to which training data of a minor attribute automatically generated by the generation unit is added (supplementing the training dataset with additional images such as facial images that represent underrepresented demographic groups) ([0035]).
Regarding claim 8, Kaliouby teaches further comprising a model bias determination unit (detect bias block 310) (Fig. 3; [0040]) which determines presence or absence of bias of the model that has been trained (detecting bias of the neural network) ([0024], [0035], and [0040]) due to a difference in attributes of input data (based on the demographic groups being disproportionate) ([0027], [0030], and [0035]).
Regarding claim 12, Kaliouby teaches an information processing method (computer-implemented method) ([0009]) comprising:
a determination step of classifying training data (classifying facial images) ([0023-0024]) for training a model (training a neural network using datasets called training datasets) ([0023]) for each attribute (demographic information) ([0023]) and determining presence or absence of bias of the training data (detecting bias within a training dataset) ([0024] and [0035]) due to a difference in attributes (based on the demographic groups being disproportionate) ([0027], [0030], and [0035]); and
a generation step of automatically generating training data of a minor attribute and mitigating the bias when it is determined in the determination step that there is the bias (when bias is detected within the training dataset, augmentation techniques can be used to mitigate the bias; the augmenting can include supplementing the training dataset with additional images such as facial images that represent underrepresented demographic groups) ([0035]).
Although Kaliouby does not explicitly state determining the “absence” of bias, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention that since Kaliouby teaches detecting bias (Abstract and [0027]), that when it doesn’t detect bias there is obviously an absence of bias.
Regarding claim 13, Kaliouby teaches a computer program described in a computer-readable format so as to cause a computer (computer program product embodied in a non-transitory computer readable medium that includes code executable by one or more processors) ([0034]) to function as:
a determination unit which classifies training data (classifying facial images) ([0023-0024]) for training a model (training a neural network using datasets called training datasets) ([0023]) for each attribute (demographic information) ([0023]) and
determines presence or absence of bias of the training data (detecting bias within a training dataset) ([0024] and [0035]) due to a difference in attributes (based on the demographic groups being disproportionate) ([0027], [0030], and [0035]); and
a generation unit which automatically generates training data of a minor attribute and mitigating the bias when the determination unit determines that there is the bias (when bias is detected within the training dataset, augmentation techniques can be used to mitigate the bias; the augmenting can include supplementing the training dataset with additional images such as facial images that represent underrepresented demographic groups) ([0035]).
Although Kaliouby does not explicitly state determining the “absence” of bias, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention that since Kaliouby teaches detecting bias (Abstract and [0027]), that when it doesn’t detect bias there is obviously an absence of bias.
Claim(s) 3 is rejected under 35 U.S.C. 103 as being unpatentable over el Kaliouby et al., US 2022/0101146 A1 (Kaliouby), and further in view of Nushi et al., US 2020/0349395 A1 (Nushi).
Regarding claim 3, Kaliouby teaches wherein the determination unit determines an attribute (demographic groups based on age bands, ethnicity, gender, etc.) ([0006]) of a face image as training data (in a facial image of the training data) ([0006]).
However, Kaliouby does not explicitly teach determining the demographic based on “a skin color” of the face image.
Nushi teaches evaluating the performance of a machine learning system in connection with a test dataset (Abstract); and wherein a test instance refers to an image of a face including demographic identifiers such as skin color ([0019]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Kaliouby to include detecting skin color for determining the demographic of the person, so that a bias in the test data can be detected for a certain group while accomplishing a higher degree of accuracy for the resulting machine learning system (Nushi; [0015-0016]).
Claim(s) 4 and 5 are rejected under 35 U.S.C. 103 as being unpatentable over el Kaliouby et al., US 2022/0101146 A1 (Kaliouby), and further in view of JI et al., US 2021/0158094 A1 (JI).
Regarding claim 4, Kaliouby teaches wherein the determination unit determines that there is bias when a difference in number of pieces of training data between attributes (when one of the demographic groups is underrepresented; i.e. the demographic group has less images thus creating a larger difference) ([0035]).
However, Kaliouby does not explicitly teach determining that the difference “is equal to or greater than a predetermined threshold”.
JI teaches training a machine learning model to classify images (Abstract); wherein pre-process the training data set to generate a balanced training data set such that each user-defined category includes a similar number of images for use in training the initial machine learning model (e.g., such that the difference between the number of images in a first user-defined category and the number of images in a second user-defined category is within a threshold amount) ([0024]); and wherein if it were greater than the threshold a bias would be introduced ([0024]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Kaliouby to include using a threshold since allows for detecting of a bias and the ability to avoid biasing the machine learning model (JI; [0024]).
Regarding claim 5, JI teaches further comprising a setting unit which sets the threshold (wherein the feature map generator sets the threshold) ([0024]).
Claim(s) 9 is rejected under 35 U.S.C. 103 as being unpatentable over el Kaliouby et al., US 2022/0101146 A1 (Kaliouby), and further in view of Lee et al., US 2021/0319333 A1 (Lee).
Regarding claim 9, Kaliouby teaches wherein the model bias determination unit (detect bias block 310) (Fig. 3; [0040]) determines presence or absence of bias of the model that has been trained (detecting bias of the neural network) ([0024], [0035], and [0040]).
However, Kaliouby does not explicitly teach detecting bias based “on a basis of ratios of a prediction label in respective attributes of input data”.
Lee teaches detecting biases in predictive models and the root cause of those biases (Abstract); and wherein detecting bias based on a basis of ratios of a prediction label in respective attributes of input data (wherein a bias is detected when the predictive label ratio is wide) ([0042] and [0060-0062]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Kaliouby to include using a ratio based on the predicted labels since it allows for detection and isolation of bias in predictive models and thus can improve the performance (accuracy) of the predictive model (Lee; [0018]).
Claim(s) 10 and 11 are rejected under 35 U.S.C. 103 as being unpatentable over el Kaliouby et al., US 2022/0101146 A1 (Kaliouby), Lee et al., US 2021/0319333 A1 (Lee), and further in view of JI et al., US 2021/0158094 A1 (JI).
Regarding claim 10, Kaliouby teaches wherein the model bias determination unit (detect bias block 310) (Fig. 3; [0040]) determines that there is bias (detecting bias of the neural network) ([0024], [0035], and [0040]) when a difference in prediction labels between attributes of input data (based on the demographic groups being disproportionate) ([0027], [0030], and [0035]). Lee teaches a determination unit which determines presence or absence of bias of a prediction result (detecting bias within the result) ([0042] and [0060-0061]) due to a difference in attributes of input data (determining that there is an under-representation of training data associated with the feature group) ([0060-0061]).
However, neither explicitly teaches determining that the difference “is equal to or greater than a predetermined threshold”.
JI teaches training a machine learning model to classify images (Abstract); wherein pre-process the training data set to generate a balanced training data set such that each user-defined category includes a similar number of images for use in training the initial machine learning model (e.g., such that the difference between the number of images in a first user-defined category and the number of images in a second user-defined category is within a threshold amount) ([0024]); and wherein if it were greater than the threshold a bias would be introduced ([0024]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Kaliouby to include using a threshold since allows for detecting of a bias and the ability to avoid biasing the machine learning model (JI; [0024]).
Regarding claim 11, JI teaches further comprising a setting unit which sets the threshold (wherein the feature map generator sets the threshold) ([0024]).
Claim(s) 14-18 are rejected under 35 U.S.C. 103 as being unpatentable over Lee et al., US 2021/0319333 A1 (Lee).
Regarding claim 14, Lee teaches an information processing device comprising:
a prediction unit which makes a prediction about input data by using a trained model (wherein a predictive model makes a prediction about input data based on training data; such as female executives as having below-average performance) ([0042]); and
a determination unit which determines presence or absence of bias of a prediction result (detecting bias within the result) ([0042] and [0060-0061]) due to a difference in attributes of input data (determining that there is an under-representation of training data associated with the feature group) ([0060-0061]).
Although Lee does not explicitly state determining the “absence” of bias, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention that since Lee teaches detecting bias (Abstract), that when it doesn’t detect bias there is obviously an absence of bias.
Regarding claim 15, Lee teaches further comprising a notification unit which gives notification of a determination result obtained by the determination unit to an external device (giving a notification in the form of a graphical user interface for display of the determination result, such as the bias detected in the predictive model) ([0073]).
Regarding claim 16, Lee teaches wherein a model parameter in which bias has been mitigated according to the notification is received, and set in the trained model (the model correction receives the identification of the cause of the bias and generates a model correction that, when applied to the predictive model, reduces or eliminates the bias) ([0062-0063]).
Regarding claim 17, Lee teaches an information processing method comprising:
a prediction step of making a prediction about input data by using a trained model (wherein a predictive model makes a prediction about input data based on training data; such as female executives as having below-average performance) ([0042]); and
a determination step of determining presence or absence of bias of a prediction result (detecting bias within the result) ([0042] and [0060-0061]) due to a difference in attributes of input data (determining that there is an under-representation of training data associated with the feature group) ([0060-0061]).
Although Lee does not explicitly state determining the “absence” of bias, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention that since Lee teaches detecting bias (Abstract), that when it doesn’t detect bias there is obviously an absence of bias.
Regarding claim 18, Lee teaches a computer program described in a computer-readable format so as to cause a computer to function as (program code may be resident in the memory device, or any suitable computer-readable medium, and may be executed by the processor) ([0127]):
a prediction unit which makes a prediction about input data by using a trained model (wherein a predictive model makes a prediction about input data based on training data; such as female executives as having below-average performance) ([0042]); and
a determination unit which determines presence or absence of bias of a prediction result (detecting bias within the result) ([0042] and [0060-0061]) due to a difference in attributes of input data (determining that there is an under-representation of training data associated with the feature group) ([0060-0061]).
Although Lee does not explicitly state determining the “absence” of bias, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention that since Lee teaches detecting bias (Abstract), that when it doesn’t detect bias there is obviously an absence of bias.
Contact
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MICHAEL J VANCHY JR whose telephone number is (571)270-1193. The examiner can normally be reached Monday - Friday 9am - 5pm.
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, Emily Terrell can be reached at (571) 270-3717. 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.
/MICHAEL J VANCHY JR/Primary Examiner, Art Unit 2666 Michael.Vanchy@uspto.gov
Prosecution Insights