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 .
Information Disclosure Statement
The information disclosure statement (IDS) submitted on September 28, 2023 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Drawings
The drawings are objected to as failing to comply with 37 CFR 1.84(p)(4) because reference character “104” has been used to designate both "Computing Device" and "Nearest Neighbor Graph". Corrected drawing sheets in compliance with 37 CFR 1.121(d) are required in reply to the Office action to avoid abandonment of the application. Any amended replacement drawing sheet should include all of the figures appearing on the immediate prior version of the sheet, even if only one figure is being amended. Each drawing sheet submitted after the filing date of an application must be labeled in the top margin as either “Replacement Sheet” or “New Sheet” pursuant to 37 CFR 1.121(d). If the changes are not accepted by the examiner, the applicant will be notified and informed of any required corrective action in the next Office action. The objection to the drawings will not be held in abeyance.
New corrected drawings in compliance with 37 CFR 1.121(d) are required in this application because text in figures 5B, and 11A-E are illegible due to text on a grayscale drawing, and figures 12-17 are also illegible. Applicant is advised to employ the services of a competent patent draftsperson outside the Office, as the U.S. Patent and Trademark Office no longer prepares new drawings. The corrected drawings are required in reply to the Office action to avoid abandonment of the application. The requirement for corrected drawings will not be held in abeyance.
Claim Objections
Claim 10 is objected to because of the following informalities: In claim 10 line 2, "a selected node at a center of hypercube", should read "a selected node at a center of a hypercube". Appropriate correction is required.
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 1-15 rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea (mental process) without significantly more.
Claim 1:
Regarding claim 1, in step 1 of the 101-analysis set forth in MPEP 2106, the claim recites “A method for construction of nearest neighbor structures”, and a method or process is one of the four statutory categories of invention.
In step 2A prong 1 of the 101-analysis set forth in the MPEP 2106, the examiner has determined that the following limitations recite a process that, under the broadest reasonable interpretation, covers a mental process but for recitation of generic computer components:
“determining a set of cross-class neighborhood similarities based on a set of distributions of data obtained;” (this is a mental process, a person could mentally determine a set of cross-class neighborhood similarities based on distribution data, see MPEP § 2106.04(a)(2)(III)),
“selecting a first cross-class neighborhood similarity from the set of cross-class neighborhood similarities based on one or more inter-class cross-class neighborhood similarities and one or more intra-class cross-class neighborhood similarities;” (this is a mental process, a person could mentally select a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities, see MPEP § 2106.04(a)(2)(III)),
“building a nearest neighbor graph based on the first cross-class neighborhood similarity” (this is a mental process, a person could mentally build a nearest neighbor graph, see MPEP § 2106.04(a)(2)(III)),
If claim limitations, under the broadest reasonable interpretation, covers performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
In step 2A prong 2 of the 101-analysis set forth in MPEP 2106, the examiner has determined that the following additional elements do not integrate this judicial exception into a practical application:
“by applying a model to data present in a dataset;” (Applying a model to data is considered mere instructions to apply an exception using generic computer – see MPEP § 2106.05(f)),
Since the claim as a whole, looking at the additional elements individually and in combination, does not contain any other additional elements that are indicative of integration into a practical application, the claim is “directed” to an abstract idea.
In step 2B of the 101-analysis set forth in the 2019 PEG, the examiner has determined that the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
As discussed above, additional element iv recites mere instructions to apply an exception using generic computer, which is not indicative of significantly more.
Considering the additional elements individually and in combination, and the claim as a whole, the additional elements do not provide significantly more than the abstract idea. Therefore, the claim is not patent eligible.
Claim 2:
Regarding claim 2, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis to claim 1. Further, claim 2 recites the following additional elements:
“modeling each data point of the data that belongs to a class (C) as a mixture of Gaussian distributions (M).” (this is a mental process, a person could mentally evaluate modeling each data point of data that belongs to a class, see MPEP § 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, covers performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
“The method of claim 1, applying the model to data present in the dataset comprises: receiving data for the dataset in a tabular from via user input;” (In step 2A, prong 2, this is considered insignificant extra-solution activity of mere data gathering – see MPEP § 2106.05(g)). (In step 2B, this is also considered insignificant extra-solution activity of mere data gathering, which is a well understood routine and conventional activity, see receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362).
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 3:
Regarding claim 3, it is dependent upon claim 2, and thereby incorporates the limitations of, and corresponding analysis to claim 2. Further, claim 3 recites the following additional elements:
“determining learned parameters for the set of distributions of data present in the dataset.” (this is a mental process, a person could mentally evaluate determining learned parameters for a set of distributions of data, see MPEP § 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, covers performance of the limitations as a mental process but for the recitation of generic compute components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
“The method of claim 2, wherein applying the model to data present in the dataset further comprises” (In step 2A, prong 2, this is considered mere instructions to apply an exception using generic computer, see MPEP § 2106.05(f)). (In step 2B, this is also considered mere instructions to apply an exception using generic computer - see MPEP § 2106.05(f)).
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 4:
Regarding claim 4, it is dependent upon claim 3, and thereby incorporates the limitations of, and corresponding analysis to claim 3. Further, claim 4 recites the following additional elements:
“The method of claim 3, wherein determining the set of cross-class neighborhood similarities comprises using the learned parameters to compute a value of the cross-class neighborhood similarities, wherein the nearest neighbor graph is built based on the value of the cross-class neighborhood similarities.” (this is a mathematical concept, computing a value of cross-class neighborhood similarities and building a nearest neighbor graph based on the values of the similarities is a mathematical calculation (see Specification paragraphs [0063] and [0110]), see MPEP § 2106.04(a)(2)(I)),
If claim limitations, under their broadest reasonable interpretation, covers performance of the limitations as a mathematical concept but for the recitation of generic computer components, then it falls within the mathematical concept grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 5:
Regarding claim 5, it is dependent upon claim 4, and thereby incorporates the limitations of, and corresponding analysis to claim 4. Further, claim 5 recites the following additional elements:
“The method of claim 4, wherein the value of the cross-class neighborhood similarities is computed using Monte Carlo simulations.” (this is a mathematical concept, computing using Monte Carlo simulation is a mathematical formula (see Specification paragraph [0066]), see MPEP § 2106.04(a)(2)(I)),
If claim limitations, under their broadest reasonable interpretation, covers performance of the limitations as a mathematical concept but for the recitation of generic computer components, then it falls within the mathematical concept grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 6:
Regarding claim 6, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis to claim 1. Further, claim 6 recites the following additional elements:
“The method of claim 1, further comprising: training a graph machine learning model based on the nearest neighbor graph;” (In step 2A, prong 2, this is considered mere instructions to apply an exception using generic computer, see MPEP § 2106.05(f)). (In step 2B, this is also considered mere instructions to apply an exception using generic computer - see MPEP § 2106.05(f)).
“performing predictive tasks using the trained graph machine learning model.” (In step 2A, prong 2, this is considered mere instructions to apply an exception using generic computer, see MPEP § 2106.05(f)). (In step 2B, this is also considered mere instructions to apply an exception using generic computer - see MPEP § 2106.05(f)).
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 7:
Regarding claim 7, it is dependent upon claim 2, and thereby incorporates the limitations of, and corresponding analysis to claim 2. Further, claim 7 recites the following additional elements:
“The method of claim 1, wherein the model applied to the data present in the dataset is a Hierarchical Naïve Bayes model, wherein the Hierarchical Naïve Bayes model models the set of distributions of the data as a mixture of Gaussian distributions, and wherein mixing weights of the mixture are obtained by another mixture of categorical distributions.” (this is a mathematical concept, Hierarchical Naïve Bayes model, mixture of Gaussian distributions, and mixing weights are considered mathematical calculations (see Specification paragraphs [0049], [0050], [0104] and [0105]), see MPEP § 2106.04(a)(2)(I)),
If claim limitations, under their broadest reasonable interpretation, covers performance of the limitations as a mathematical concept but for the recitation of generic computer components, then it falls within the mathematical concept grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 8:
Regarding claim 8, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis to claim 1. Further, claim 8 recites the following additional elements:
“The method of claim 1, wherein determining the set of distributions comprises computing a probability that a first node, belonging to a first class, has a nearest neighbor node, belonging to a second class.” (this is a mathematical concept, computing a probability is a mathematical calculation, see MPEP § 2106.04(a)(2)(I)),
If claim limitations, under their broadest reasonable interpretation, covers performance of the limitations as a mathematical concept but for the recitation of generic computer components, then it falls within the mathematical concept grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 9:
Regarding claim 9, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis to claim 1. Further, claim 9 recites the following additional elements:
“The method of claim 1, wherein selecting the first cross-class neighborhood similarity comprises determining a trade-off between the one or more inter-class cross-class neighborhood similarities and the one or more intra-class cross-class neighborhood similarities.” (this is a mental process, a person could mentally evaluate determining a trade-off between similarities, see MPEP § 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, covers performance of the limitations as a mental process but for the recitation of generic compute components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 10:
Regarding claim 10, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis to claim 1. Further, claim 10 recites the following additional elements:
“The method of claim 1, wherein the nearest neighbor graph comprises: a selected node at a center of hypercube, wherein the hypercube comprises an edge that is optimized; and a set of neighbors of the selected node within the hypercube based on the edge of the hypercube, wherein the hypercube is formed based on the first parameter.” (this is a mental process, this just further elaborates what the nearest neighbor graph comprises of and a person can mentally evaluate that, see MPEP § 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, covers performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 11:
Regarding claim 11, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis to claim 1. Further, claim 11 recites the following additional elements:
“the nearest neighbor graph is built based on the electronic health records present in the dataset;” (this is a mental process, a person could mentally evaluate building a nearest neighbor graph built on electronic health records in a dataset, see MPEP § 2106.04(a)(2)(III)),
“a clinical risk is predicted for a patient” (this is a mental process, a person could mentally evaluate predicting a clinical risk for a patient, see MPEP § 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, covers performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
“The method of claim 1, wherein: the data present in the dataset comprises electronic health records corresponding to a plurality of patients, wherein the electronic health records comprise heart rate, oxygen saturation, weight, height, glucose, temperature associated with each patient in the plurality of patients;” (In step 2A, prong 2, this is considered as mere field of use or technological environment in which to apply a judicial exception, see MPEP § 2106.05(h)). (In step 2B, this is also considered mere field of use or technological environment in which to apply a judicial exception - see MPEP § 2106.05(h)).
“a graph machine learning model is trained using the nearest neighbor graph;” (In step 2A, prong 2, this is considered mere instructions to apply an exception using generic computer, see MPEP § 2106.05(f)). (In step 2B, this is also considered mere instructions to apply an exception using generic computer - see MPEP § 2106.05(f)).
“using the trained graph machine learning model.” (In step 2A, prong 2, this is considered mere instructions to apply an exception using generic computer, see MPEP § 2106.05(f)). (In step 2B, this is also considered mere instructions to apply an exception using generic computer - see MPEP § 2106.05(f)).
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 12:
Regarding claim 12, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis to claim 1. Further, claim 12 recites the following additional elements:
“the nearest neighbor graph is built based on the genomic activity information present in the dataset;” (this is a mental process, a person could mentally evaluate building a nearest neighbor graph built on genomic activity information in a dataset, see MPEP § 2106.04(a)(2)(III)),
“a suitability of a patient for a drug trial is predicted” (this is a mental process, a person could mentally evaluate predicting a suitability of a patient for a drug trial, see MPEP § 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, covers performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
“The method of claim 1, wherein: the data present in the dataset comprises genomic activity information corresponding to a plurality of patients, wherein the genomic activity of each patient identifies a response of the respective patient to a drug;” (In step 2A, prong 2, this is considered as mere field of use or technological environment in which to apply a judicial exception, see MPEP § 2106.05(h)). (In step 2B, this is also considered mere field of use or technological environment in which to apply a judicial exception - see MPEP § 2106.05(h)).
“a graph machine learning model is trained using the nearest neighbor graph;” (In step 2A, prong 2, this is considered mere instructions to apply an exception using generic computer, see MPEP § 2106.05(f)). (In step 2B, this is also considered mere instructions to apply an exception using generic computer - see MPEP § 2106.05(f)).
“using the graph machine learning model.” (In step 2A, prong 2, this is considered mere instructions to apply an exception using generic computer, see MPEP § 2106.05(f)). (In step 2B, this is also considered mere instructions to apply an exception using generic computer - see MPEP § 2106.05(f)).
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 13:
Regarding claim 13, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis to claim 1. Further, claim 13 recites the following additional elements:
“the nearest neighbor graph is built based on the soil data present in the dataset;” (this is a mental process, a person could mentally evaluate building a nearest neighbor graph built on genomic activity information in a dataset, see MPEP § 2106.04(a)(2)(III)),
“a quality of an input soil type is predicted based on the nearest neighbor graph.” (this is a mental process, a person could mentally predict a soil type based on a nearest neighbor graph, see MPEP § 2106.04(a)(2)(III)),
If claim limitations, under their broadest reasonable interpretation, covers performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
“The method of claim 1, wherein: the data present in the dataset comprises soil data corresponding to a plurality of areas, wherein the soil data comprises humidity, temperature, and performance metrics related to different areas;” (In step 2A, prong 2, this is considered as mere field of use or technological environment in which to apply a judicial exception, see MPEP § 2106.05(h)). (In step 2B, this is also considered mere field of use or technological environment in which to apply a judicial exception - see MPEP § 2106.05(h)).
“a graph machine learning model is trained using the nearest neighbor graph;” (In step 2A, prong 2, this is considered mere instructions to apply an exception using generic computer, see MPEP § 2106.05(f)). (In step 2B, this is also considered mere instructions to apply an exception using generic computer - see MPEP § 2106.05(f)).
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Claim 14:
Regarding claim 14, in step 1 of the 101-analysis set forth in MPEP 2106, the claim recites “A computer system programmed for performing automated sharing of data and analytics across a data space platform”, and a system or machine is one of the four statutory categories of invention.
In step 2A prong 1 of the 101-analysis set forth in the MPEP 2106, the examiner has determined that the following limitations recite a process that, under the broadest reasonable interpretation, covers a mental process but for recitation of generic computer components:
“determining a set of cross-class neighborhood similarities based on a set of distributions of data obtained” (this is a mental process, a person could mentally determine a set of cross-class neighborhood similarities based on distribution data, see MPEP § 2106.04(a)(2)(III)),
“selecting a first cross-class neighborhood similarity from the set of cross-class neighborhood similarities based on one or more inter-class cross-class neighborhood similarities and one or more intra-class cross-class neighborhood similarities;” (this is a mental process, a person could mentally select a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities, see MPEP § 2106.04(a)(2)(III)),
“building a nearest neighbor graph based on the first cross-class neighborhood similarity” (this is a mental process, a person could mentally build a nearest neighbor graph, see MPEP § 2106.04(a)(2)(III)),
If claim limitations, under the broadest reasonable interpretation, covers performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
In step 2A prong 2 of the 101-analysis set forth in MPEP 2106, the examiner has determined that the following additional elements do not integrate this judicial exception into a practical application:
“A computer system programmed for performing automated sharing of data and analytics across a data space platform, the computer system comprising one or more hardware processors which, alone or in combination, are configured to provide for execution” (Using hardware processors is considered generic computer component being used as tool to perform functions of the judicial exception – see MPEP § 2106.05(f)),
“by applying a model to data present in a dataset;” (Applying a model to data is considered mere instructions to apply an exception using generic computer – see MPEP § 2106.05(f)),
Since the claim as a whole, looking at the additional elements individually and in combination, does not contain any other additional elements that are indicative of integration into a practical application, the claim is “directed” to an abstract idea.
In step 2B of the 101-analysis set forth in the 2019 PEG, the examiner has determined that the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
As discussed above, additional element iv recites generic computer component being used as tool to perform functions of the judicial exception, and additional element v recites mere instructions to apply an exception using generic computer, which is not indicative of significantly more.
Considering the additional elements individually and in combination, and the claim as a whole, the additional elements do not provide significantly more than the abstract idea. Therefore, the claim is not patent eligible.
Claim 15:
Regarding claim 15, in step 1 of the 101-analysis set forth in MPEP 2106, the claim recites “A tangible, non-transitory computer-readable medium for performing automated sharing of data and analytics across a data space platform having instructions thereon”, and a non-transitory computer-readable medium or machine is one of the four statutory categories of invention.
In step 2A prong 1 of the 101-analysis set forth in the MPEP 2106, the examiner has determined that the following limitations recite a process that, under the broadest reasonable interpretation, covers a mental process but for recitation of generic computer components:
“determining a set of cross-class neighborhood similarities based on a set of distributions of data obtained” (this is a mental process, a person could mentally determine a set of cross-class neighborhood similarities based on distribution data, see MPEP § 2106.04(a)(2)(III)),
“selecting a first cross-class neighborhood similarity from the set of cross-class neighborhood similarities based on one or more inter-class cross-class neighborhood similarities and one or more intra-class cross-class neighborhood similarities;” (this is a mental process, a person could mentally select a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities, see MPEP § 2106.04(a)(2)(III)),
“building a nearest neighbor graph based on the first cross-class neighborhood similarity” (this is a mental process, a person could mentally build a nearest neighbor graph, see MPEP § 2106.04(a)(2)(III)),
If claim limitations, under the broadest reasonable interpretation, covers performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
In step 2A prong 2 of the 101-analysis set forth in MPEP 2106, the examiner has determined that the following additional elements do not integrate this judicial exception into a practical application:
“A tangible, non-transitory computer-readable medium for performing automated sharing of data and analytics across a data space platform having instructions thereon, which, upon being executed by one or more processors, provides for execution of the following steps” (Using processors is considered generic computer component being used as tool to perform functions of the judicial exception – see MPEP § 2106.05(f)),
“by applying a model to data present in a dataset;” (Applying a model to data is considered mere instructions to apply an exception using generic computer – see MPEP § 2106.05(f)),
Since the claim as a whole, looking at the additional elements individually and in combination, does not contain any other additional elements that are indicative of integration into a practical application, the claim is “directed” to an abstract idea.
In step 2B of the 101-analysis set forth in the 2019 PEG, the examiner has determined that the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
As discussed above, additional element iv recites generic computer component being used as tool to perform functions of the judicial exception, and additional element v recites mere instructions to apply an exception using generic computer, which is not indicative of significantly more.
Considering the additional elements individually and in combination, and the claim as a whole, the additional elements do not provide significantly more than the abstract idea. Therefore, the claim is not patent eligible.
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, 9 and 14-15 are rejected under 35 U.S.C. 103 as being unpatentable over Ding C. et al., "Similarity-balanced Discriminant Neighborhood Embedding", available at https://ieeexplore.ieee.org/abstract/document/6889611, published on July 6-11, 2014, (hereafter Ding), in view of Niu W. et al., (US. Patent 12393617 B1) effectively filed on September 30, 2022, (hereafter Niu).
Claim 1:
Regarding claim 1, Ding teaches “A method for construction of nearest neighbor structures, the method comprising: selecting a first cross-class neighborhood similarity from the set of cross-class neighborhood similarities based on one or more inter-class cross-class neighborhood similarities and one or more intra-class cross-class neighborhood similarities;”
See Ding in Construction of adjacent graphs section on page 3 describing “Now, we consider the construction of adjacent graphs according to the new similarity function (9). Our scheme is to select the farthest homogeneous neighbors for a sample to construct an intra-class graph Fw, and it's nearest heterogeneous neighbors to build an inter-class graph Fb. ” Here, Ding establishes a method for nearest neighbor structures with the scheme to construct intra and inter-class graphs based on near and far neighbors. The scheme uses a similarity function for the construction of these graphs, and the set of cross-class neighborhood similarities here is with the selected intra-class and inter-class similarities. Further, see Ding in Construction of adjacent graphs section on page 4 describing “For
PNG
media_image1.png
30
27
media_image1.png
Greyscale
, we select the k homogeneous samples with the smallest similarity for
PNG
media_image2.png
26
19
media_image2.png
Greyscale
and preserve their structural relationships.” Further, see Ding in Construction of adjacent graphs section on page 4 describing “On the contrary, for
PNG
media_image3.png
31
25
media_image3.png
Greyscale
, k heterogeneous nearest neighbors with the highest similarity are selected for
PNG
media_image2.png
26
19
media_image2.png
Greyscale
.” The intra and inter class graphs are what is denoted here with the big F and a selection of a first class-class neighborhood similarity is done here based on either the inter or intra-class graph similarities. Further, see in Construction of adjacent graphs section on page 4 describing “We try to maximize the difference between the nearest inter-class distance and the farthest intra-class distance so as to make the distance between the same classes is nearer and the distance between the different classes is farther in the projection sub-space.” Here Ding establishes a difference between the nearest intra-class distance and farthest intra-class distance which can be seen as a selection of a first cross-class similarity.
Further, Ding teaches “and building a nearest neighbor graph based on the first cross-class neighborhood similarity.”
See Ding in Construction of adjacent graphs section on pages 3-4 and Figure 2 describing “Now, we consider the construction of adjacent graphs according to the new similarity function (9). Our scheme is to select the farthest homogeneous neighbors for a sample to construct an intra-class graph Fw, and it's nearest heterogeneous neighbors to build an inter-class graph Fb. The reason can be illustrated by Fig. 2. In Fig. 2(a), there are three classes denoted by solid square, circle and solid triangle. For the hollow circle point, we select the farthest neighbor in the solid circle points, and the nearest neighbors in the solid square and triangle points as shown in Fig. 2(b). Fig. 2(c) ideally gives their images in the subspace. We expect that the farthest homogeneous could be attracted to around the sample and the nearest heterogeneous neighbors could be pushed way from the sample.
PNG
media_image4.png
167
352
media_image4.png
Greyscale
” Here, Ding establishes a subspace of points based on nearest neighbors which can be seen as a nearest neighbor graph. Further, see in Construction of adjacent graphs section on page 4 describing “We try to maximize the difference between the nearest inter-class distance and the farthest intra-class distance so as to make the distance between the same classes is nearer and the distance between the different classes is farther in the projection sub-space.” Here Ding establishes a difference between the nearest intra-class distance and farthest intra-class distance which can be seen as a selection of a first cross-class similarity. Ding then establishes that this difference is projected in a sub-space which as established earlier can be seen as a nearest neighbor graph, the projection can be seen as building the nearest neighbor graph.
However, Ding did not explicitly teach “determining a set of cross-class neighborhood similarities based on a set of distributions of data obtained by applying a model to data present in a dataset;”
In the same field of art, Niu teaches, “determining a set of cross-class neighborhood similarities based on a set of distributions of data obtained by applying a model to data present in a dataset;”
See Niu in column 12 lines 44-53 where it describes “The model training system 120 clusters the documents using connected components of a k-nearest neighbor graph. In some implementations, clustering the documents may result in a skewed distribution of cluster sizes (e.g., the hubness phenomenon, in which a small number of documents are nearest neighbors to a disproportionality large number of documents). In these implementations, the model training system 120 utilizes cross-domain similarity local scaling (CSLS), or any other variant, to establish a range for the nearest neighbors.” Here Niu describes establishing a range for nearest neighbors which can be seen as determining a set of cross-class neighborhood similarities in an analogous system. The model training system utilizes a cross-domain similarity local scaling, which can be a cross-class neighborhood similarity, and the clustering of documents resulting in a skewed distribution can be seen as the similarities being based on a set of distribution of data. This data is obtained by applying a model here, as the data utilizes a model training system. Further, see Niu in column 12 lines 23-28 describing, “The model training system 120 may also determine anchor-positive pairs according to mutually reinforced pretraining. At each iteration of training, a machine learning model such as base ML model 121A mines anchor-positive pairs from a document and reinforces the learning at that iteration by training using the mined pairs.” Here Niu establishes that the model training system utilizes a model for the data.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base reference of Ding with the teachings of Niu by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and incorporate with Niu’s teachings of determining a set of cross-class neighborhood similarities.
One of ordinary skill in the art would be motivated to do so because by integrating Niu’s frameworks into the methods of Ding, which are both in relation to nearest neighbor graph and computing similarity, one of ordinary skill in the art would bring “domain-specific machine learning (ML) models to learn relevant portions of a conversation that trigger assistance and identify relevant assistance information (e.g., documents, articles, phone numbers, website addresses, physical addresses, contact information, etc.). Accurately triggering the need to provide assistance (e.g., when to provide assistance to an agent) and the identification of relevant assistance information (e.g., what assistance to provide to an agent) reduces computing resources required by a system and beneficially provides superior chat assistance in the form of relevant recommended subject matter at relevant time(s) in the conversation” (Niu, column 2 lines 19-30).
Claim 9:
Regarding claim 9, Ding in view of Niu teaches the limitations of claim 1.
Further, Ding teaches “The method of claim 1, wherein selecting the first cross-class neighborhood similarity comprises determining a trade-off between the one or more inter-class cross-class neighborhood similarities and the one or more intra-class cross-class neighborhood similarities.”
See Ding in Construction of adjacent graphs section on page 4 describing “By respectively building intra-class structure graph and inter-class structure graph, each example is able to get the associations with the samples with the same or different classes. In other words, for an example, we can get at least two associations, namely the association with the same class and the association with the different classes. We try to maximize the difference between the nearest inter-class distance and the farthest intra-class distance so as to make the distance between the same classes is nearer and the distance between the different classes is farther in the projection sub-space. That's to say, we need to maximize
PNG
media_image5.png
36
192
media_image5.png
Greyscale
” Here, Ding establishes a trade-off between the inter-class cross-class neighborhood similarities and the intra-class cross-class neighborhood similarities with the maximization of a difference function which can be seen as determining a trade-off. The difference is computed using the respective graphs for the inter-class and intra-class graphs which consists of the neighborhood similarities.
Claim 14:
Regarding claim 14, Ding teaches “selecting a first cross-class neighborhood similarity from the set of cross-class neighborhood similarities based on one or more inter-class cross-class neighborhood similarities and one or more intra-class cross-class neighborhood similarities;”
See Ding in Construction of adjacent graphs section on page 3 describing “Now, we consider the construction of adjacent graphs according to the new similarity function (9). Our scheme is to select the farthest homogeneous neighbors for a sample to construct an intra-class graph Fw, and it's nearest heterogeneous neighbors to build an inter-class graph Fb. ” Here, Ding establishes a method for nearest neighbor structures with the scheme to construct intra and inter-class graphs based on near and far neighbors. The scheme uses a similarity function for the construction of these graphs, and the set of cross-class neighborhood similarities here is with the selected intra-class and inter-class similarities. Further, see Ding in Construction of adjacent graphs section on page 4 describing “For
PNG
media_image1.png
30
27
media_image1.png
Greyscale
, we select the k homogeneous samples with the smallest similarity for
PNG
media_image2.png
26
19
media_image2.png
Greyscale
and preserve their structural relationships.” Further, see Ding in Construction of adjacent graphs section on page 4 describing “On the contrary, for
PNG
media_image3.png
31
25
media_image3.png
Greyscale
, k heterogeneous nearest neighbors with the highest similarity are selected for
PNG
media_image2.png
26
19
media_image2.png
Greyscale
.” The intra and inter class graphs are what is denoted here with the big F and a selection of a first class-class neighborhood similarity is done here based on either the inter or intra-class graph similarities. Further, see in Construction of adjacent graphs section on page 4 describing “We try to maximize the difference between the nearest inter-class distance and the farthest intra-class distance so as to make the distance between the same classes is nearer and the distance between the different classes is farther in the projection sub-space.” Here Ding establishes a difference between the nearest intra-class distance and farthest intra-class distance which can be seen as a selection of a first cross-class similarity.
Further, Ding teaches “and building a nearest neighbor graph based on the first cross-class neighborhood similarity.”
See Ding in Construction of adjacent graphs section on pages 3-4 and Figure 2 describing “Now, we consider the construction of adjacent graphs according to the new similarity function (9). Our scheme is to select the farthest homogeneous neighbors for a sample to construct an intra-class graph Fw, and it's nearest heterogeneous neighbors to build an inter-class graph Fb. The reason can be illustrated by Fig. 2. In Fig. 2(a), there are three classes denoted by solid square, circle and solid triangle. For the hollow circle point, we select the farthest neighbor in the solid circle points, and the nearest neighbors in the solid square and triangle points as shown in Fig. 2(b). Fig. 2(c) ideally gives their images in the subspace. We expect that the farthest homogeneous could be attracted to around the sample and the nearest heterogeneous neighbors could be pushed way from the sample.
PNG
media_image4.png
167
352
media_image4.png
Greyscale
” Here, Ding establishes a subspace of points based on nearest neighbors which can be seen as a nearest neighbor graph. Further, see in Construction of adjacent graphs section on page 4 describing “We try to maximize the difference between the nearest inter-class distance and the farthest intra-class distance so as to make the distance between the same classes is nearer and the distance between the different classes is farther in the projection sub-space.” Here Ding establishes a difference between the nearest intra-class distance and farthest intra-class distance which can be seen as a selection of a first cross-class similarity. Ding then establishes that this difference is projected in a sub-space which as established earlier can be seen as a nearest neighbor graph, the projection can be seen as building the nearest neighbor graph.
However, Ding did not explicitly teach “A computer system programmed for performing automated sharing of data and analytics across a data space platform, the computer system comprising one or more hardware processors which, alone or in combination, are configured to provide for execution of the following steps: determining a set of cross-class neighborhood similarities based on a set of distributions of data obtained by applying a model to data present in a dataset;”
In the same field of art, Niu teaches “A computer system programmed for performing automated sharing of data and analytics across a data space platform, the computer system comprising one or more hardware processors which, alone or in combination, are configured to provide for execution of the following steps”
See Niu in column 35 lines 44-58 describing “In some examples, a system that implements a portion or all of the techniques described herein can include a general-purpose computer system, such as the computer system 1400 (referred to herein as compute device) illustrated in FIG. 14, that includes, or is configured to access, one or more computer-accessible media. In the illustrated example, the computer system 1400 includes one or more processors 1410 coupled to a system memory 1420 via an input/output (I/O) interface 1430. The computer system 1400 further includes a network interface 1440 coupled to the I/O interface 1430. While FIG. 14 shows the computer system 1400 as a single computing device, in various examples the computer system 1400 can include one computing device or any number of computing devices configured to work together as a single computer system 1400.” Here, Niu establishes a computer system that includes one or more processors to perform all techniques of the embodiment, which can be seen as providing for execution of steps. Further, see Niu in column 7 lines 13-15 describing, “The knowledge base 106 stored in the storage service 116 can be a data structure or collection storing enterprise-specific data.” Niu establishes the knowledge base as a data space here. Further, see Niu in column 10 lines 25-34 describing “At circle (4), the user 118A (such as an admin) using the compute device 122 issues one or more requests (e.g., API calls) to the machine learning manager 138 of the contact center service 180 that indicate the user's 118A desire to train base ML models 121 into domain adapted ML models 162. For example, the model training system 120 uses the enterprise specific data uploaded by an admin such as user 118A (e.g., knowledge base 106 data, conversation data 158, and/or labeled training data 160) to tune the base ML models 121 to domain adapted ML models 162.” Here, Niu establishes use of compute devices, which are established to be referred to as a computer system, and a machine learning manger to train ML models. An example is then giving off using a model training system to use enterprise data, which can be seen as data and analytics, that was shared by a user from knowledge base data, in which the knowledge base is seen here as a data space platform.
Further, Niu teaches “determining a set of cross-class neighborhood similarities based on a set of distributions of data obtained by applying a model to data present in a dataset;”
See Niu in column 12 lines 44-53 where it describes “The model training system 120 clusters the documents using connected components of a k-nearest neighbor graph. In some implementations, clustering the documents may result in a skewed distribution of cluster sizes (e.g., the hubness phenomenon, in which a small number of documents are nearest neighbors to a disproportionality large number of documents). In these implementations, the model training system 120 utilizes cross-domain similarity local scaling (CSLS), or any other variant, to establish a range for the nearest neighbors.” Here Niu describes establishing a range for nearest neighbors which can be seen as determining a set of cross-class neighborhood similarities in an analogous system. The model training system utilizes a cross-domain similarity local scaling, which can be a cross-class neighborhood similarity, and the clustering of documents resulting in a skewed distribution can be seen as the similarities being based on a set of distribution of data. This data is obtained by applying a model here, as the data utilizes a model training system. Further, see Niu in column 12 lines 23-28 describing, “The model training system 120 may also determine anchor-positive pairs according to mutually reinforced pretraining. At each iteration of training, a machine learning model such as base ML model 121A mines anchor-positive pairs from a document and reinforces the learning at that iteration by training using the mined pairs.” Here Niu establishes that the model training system utilizes a model for the data.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base reference of Ding with the teachings of Niu by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and incorporate with Niu’s teachings of determining a set of cross-class neighborhood similarities.
One of ordinary skill in the art would be motivated to do so because by integrating Niu’s frameworks into the methods of Ding, which are both in relation to nearest neighbor graph and computing similarity, one of ordinary skill in the art would bring “domain-specific machine learning (ML) models to learn relevant portions of a conversation that trigger assistance and identify relevant assistance information (e.g., documents, articles, phone numbers, website addresses, physical addresses, contact information, etc.). Accurately triggering the need to provide assistance (e.g., when to provide assistance to an agent) and the identification of relevant assistance information (e.g., what assistance to provide to an agent) reduces computing resources required by a system and beneficially provides superior chat assistance in the form of relevant recommended subject matter at relevant time(s) in the conversation” (Niu, column 2 lines 19-30).
Claim 15:
Regarding claim 15, Ding teaches “selecting a first cross-class neighborhood similarity from the set of cross-class neighborhood similarities based on one or more inter-class cross-class neighborhood similarities and one or more intra-class cross-class neighborhood similarities;”
See Ding in Construction of adjacent graphs section on page 3 describing “Now, we consider the construction of adjacent graphs according to the new similarity function (9). Our scheme is to select the farthest homogeneous neighbors for a sample to construct an intra-class graph Fw, and it's nearest heterogeneous neighbors to build an inter-class graph Fb. ” Here, Ding establishes a method for nearest neighbor structures with the scheme to construct intra and inter-class graphs based on near and far neighbors. The scheme uses a similarity function for the construction of these graphs, and the set of cross-class neighborhood similarities here is with the selected intra-class and inter-class similarities. Further, see Ding in Construction of adjacent graphs section on page 4 describing “For
PNG
media_image1.png
30
27
media_image1.png
Greyscale
, we select the k homogeneous samples with the smallest similarity for
PNG
media_image2.png
26
19
media_image2.png
Greyscale
and preserve their structural relationships.” Further, see Ding in Construction of adjacent graphs section on page 4 describing “On the contrary, for
PNG
media_image3.png
31
25
media_image3.png
Greyscale
, k heterogeneous nearest neighbors with the highest similarity are selected for
PNG
media_image2.png
26
19
media_image2.png
Greyscale
.” The intra and inter class graphs are what is denoted here with the big F and a selection of a first class-class neighborhood similarity is done here based on either the inter or intra-class graph similarities. Further, see in Construction of adjacent graphs section on page 4 describing “We try to maximize the difference between the nearest inter-class distance and the farthest intra-class distance so as to make the distance between the same classes is nearer and the distance between the different classes is farther in the projection sub-space.” Here Ding establishes a difference between the nearest intra-class distance and farthest intra-class distance which can be seen as a selection of a first cross-class similarity.
Further, Ding teaches “and building a nearest neighbor graph based on the first cross-class neighborhood similarity.”
See Ding in Construction of adjacent graphs section on pages 3-4 and Figure 2 describing “Now, we consider the construction of adjacent graphs according to the new similarity function (9). Our scheme is to select the farthest homogeneous neighbors for a sample to construct an intra-class graph Fw, and it's nearest heterogeneous neighbors to build an inter-class graph Fb. The reason can be illustrated by Fig. 2. In Fig. 2(a), there are three classes denoted by solid square, circle and solid triangle. For the hollow circle point, we select the farthest neighbor in the solid circle points, and the nearest neighbors in the solid square and triangle points as shown in Fig. 2(b). Fig. 2(c) ideally gives their images in the subspace. We expect that the farthest homogeneous could be attracted to around the sample and the nearest heterogeneous neighbors could be pushed way from the sample.
PNG
media_image4.png
167
352
media_image4.png
Greyscale
” Here, Ding establishes a subspace of points based on nearest neighbors which can be seen as a nearest neighbor graph. Further, see in Construction of adjacent graphs section on page 4 describing “We try to maximize the difference between the nearest inter-class distance and the farthest intra-class distance so as to make the distance between the same classes is nearer and the distance between the different classes is farther in the projection sub-space.” Here Ding establishes a difference between the nearest intra-class distance and farthest intra-class distance which can be seen as a selection of a first cross-class similarity. Ding then establishes that this difference is projected in a sub-space which as established earlier can be seen as a nearest neighbor graph, the projection can be seen as building the nearest neighbor graph.
However, Ding did not explicitly teach “A tangible, non-transitory computer-readable medium for performing automated sharing of data and analytics across a data space platform having instructions thereon, which, upon being executed by one or more processors, provides for execution of the following steps: determining a set of cross-class neighborhood similarities based on a set of distributions of data obtained by applying a model to data present in a dataset;”
In the same field of art, Niu teaches “A tangible, non-transitory computer-readable medium for performing automated sharing of data and analytics across a data space platform having instructions thereon, which, upon being executed by one or more processors, provides for execution of the following steps:”
See Niu in column 37 lines 12-17 describing “In some examples, the system memory 1420 can be one example of a computer-accessible medium configured to store program instructions and data as described above. However, in other examples, program instructions and/or data can be received, sent, or stored upon different types of computer-accessible media. Generally speaking, a computer-accessible medium can include any non-transitory storage media or memory media such as magnetic or optical media, e.g., disk or DVD/CD coupled to the computer system 1400 via the I/O interface 1430.” Here, Niu establishes a non-transitory computer readable medium as part of a computer system. Further, see Niu in column 35 lines 44-58 describing “In some examples, a system that implements a portion or all of the techniques described herein can include a general-purpose computer system, such as the computer system 1400 (referred to herein as compute device) illustrated in FIG. 14, that includes, or is configured to access, one or more computer-accessible media. In the illustrated example, the computer system 1400 includes one or more processors 1410 coupled to a system memory 1420 via an input/output (I/O) interface 1430. The computer system 1400 further includes a network interface 1440 coupled to the I/O interface 1430. While FIG. 14 shows the computer system 1400 as a single computing device, in various examples the computer system 1400 can include one computing device or any number of computing devices configured to work together as a single computer system 1400.” Here, Niu establishes a computer system that includes one or more processors to perform all techniques of the embodiment, which can be seen as providing for execution of steps. Further, see Niu in column 7 lines 13-15 describing, “The knowledge base 106 stored in the storage service 116 can be a data structure or collection storing enterprise-specific data.” Niu establishes the knowledge base as a data space here. Further, see Niu in column 10 lines 25-34 describing “At circle (4), the user 118A (such as an admin) using the compute device 122 issues one or more requests (e.g., API calls) to the machine learning manager 138 of the contact center service 180 that indicate the user's 118A desire to train base ML models 121 into domain adapted ML models 162. For example, the model training system 120 uses the enterprise specific data uploaded by an admin such as user 118A (e.g., knowledge base 106 data, conversation data 158, and/or labeled training data 160) to tune the base ML models 121 to domain adapted ML models 162.” Here, Niu establishes use of compute devices, which are established to be referred to as a computer system, and a machine learning manger to train ML models. An example is then giving off using a model training system to use enterprise data, which can be seen as data and analytics, that was shared by a user from knowledge base data, in which the knowledge base is seen here as a data space platform.
Further, Niu teaches “determining a set of cross-class neighborhood similarities based on a set of distributions of data obtained by applying a model to data present in a dataset;”
See Niu in column 12 lines 44-53 where it describes “The model training system 120 clusters the documents using connected components of a k-nearest neighbor graph. In some implementations, clustering the documents may result in a skewed distribution of cluster sizes (e.g., the hubness phenomenon, in which a small number of documents are nearest neighbors to a disproportionality large number of documents). In these implementations, the model training system 120 utilizes cross-domain similarity local scaling (CSLS), or any other variant, to establish a range for the nearest neighbors.” Here Niu describes establishing a range for nearest neighbors which can be seen as determining a set of cross-class neighborhood similarities in an analogous system. The model training system utilizes a cross-domain similarity local scaling, which can be a cross-class neighborhood similarity, and the clustering of documents resulting in a skewed distribution can be seen as the similarities being based on a set of distribution of data. This data is obtained by applying a model here, as the data utilizes a model training system. Further, see Niu in column 12 lines 23-28 describing, “The model training system 120 may also determine anchor-positive pairs according to mutually reinforced pretraining. At each iteration of training, a machine learning model such as base ML model 121A mines anchor-positive pairs from a document and reinforces the learning at that iteration by training using the mined pairs.” Here Niu establishes that the model training system utilizes a model for the data.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base reference of Ding with the teachings of Niu by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and incorporate with Niu’s teachings of determining a set of cross-class neighborhood similarities.
One of ordinary skill in the art would be motivated to do so because by integrating Niu’s frameworks into the methods of Ding, which are both in relation to nearest neighbor graph and computing similarity, one of ordinary skill in the art would bring “domain-specific machine learning (ML) models to learn relevant portions of a conversation that trigger assistance and identify relevant assistance information (e.g., documents, articles, phone numbers, website addresses, physical addresses, contact information, etc.). Accurately triggering the need to provide assistance (e.g., when to provide assistance to an agent) and the identification of relevant assistance information (e.g., what assistance to provide to an agent) reduces computing resources required by a system and beneficially provides superior chat assistance in the form of relevant recommended subject matter at relevant time(s) in the conversation” (Niu, column 2 lines 19-30).
Claim(s) 2-4 are rejected under 35 U.S.C. 103 as being unpatentable over Ding C. et al., in view of Niu W. et al., and further in view of Zhao Z. et al., (International Application Publication WO2023018330A1) effectively filed on August 11, 2022, (hereafter Zhao).
Claim 2:
Regarding claim 2, Ding in view of Niu teaches the limitations of claim 1.
Neither Ding or Niu appears to teach “The method of claim 1, applying the model to data present in the dataset comprises: receiving data for the dataset in a tabular from via user input; and modeling each data point of the data that belongs to a class (C) as a mixture of Gaussian distributions (M).” However, Zhao in the same field of art teaches “The method of claim 1, applying the model to data present in the dataset comprises: receiving data for the dataset in a tabular from via user input;”
See Zhao in page 1 lines 25-33 describing “a method of federated learning of a model representing tabulated data is provided, wherein the method comprises receiving distribution information from each of a plurality of clients, wherein the distribution information indicates a statistical distribution of each of a plurality of variables of a tabular dataset associated with the corresponding client; associating a weight with each client, based on the distribution information; receiving, from each client, a set of parameters indicative of a model trained on the tabular dataset associated with the client;” Here, Zhao establishes receiving distribution information, which can be seen as data, from a plurality of clients, which can be seen as the user input. Zhao also establishes the information comes in a tabular dataset and a federated learning model is applied to the tabular data.
Further, Zhao teaches “and modeling each data point of the data that belongs to a class (C) as a mixture of Gaussian distributions (M).”
See Zhao in page 2 lines 11-12 describing “a server system for federated learning of a model representing tabulated data is provided”. Here, Zhao establishes a server for federated learning of a model, which implies the server utilizing a model. Further, see Zhao in page 9 lines 23-27 describing “In step 403, the server determines an encoding of that variable based on the set of parameters indicating the overall statistical distribution. In case of a discrete variable, for example, the encoding may comprise a one-hot encoding of all possible values that may occur among the clients. In case of a continuous variable, the encoding may comprise a variational Gaussian mixture model, for example.” Here, Zhao establishes a Gaussian mixture model being used for an encoding, as known in the art you utilize a Gaussian mixture model to model data belonging to a class. This encoding comprises distributions and the encoding models each data point of data that belongs to a class here as the parameters of the overall distribution can be seen as the class and the server determining the encoding of a variable based on the parameters can be seen as the modeling of each data point of data that belongs to the class.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base references of Ding and Niu with the teachings of Zhao by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and Niu’s teachings of determining a set of cross-class neighborhood similarities, and incorporate with Zhao’s teachings of receiving data for the dataset in a tabular, and modeling data belonging to a class as a mixture of Gaussian distributions.
One of ordinary skill in the art would be motivated to do so because by integrating Zhao’s frameworks into the methods of Ding and Niu, one of ordinary skill in the art would bring “domain-specific machine learning (ML) models to learn relevant portions of a conversation that trigger assistance and identify relevant assistance information (e.g., documents, articles, phone numbers, website addresses, physical addresses, contact information, etc.). Accurately triggering the need to provide assistance (e.g., when to provide assistance to an agent) and the identification of relevant assistance information (e.g., what assistance to provide to an agent) reduces computing resources required by a system and beneficially provides superior chat assistance in the form of relevant recommended subject matter at relevant time(s) in the conversation” (Niu, column 2 lines 19-30).
Claim 3:
Regarding claim 3, Ding in view of Niu and further in view of Zhao teaches the limitations of claim 2.
Neither Ding or Niu appears to teach “The method of claim 2, wherein applying the model to data present in the dataset further comprises determining learned parameters for the set of distributions of data present in the dataset.” However, Zhao in the same field of art teaches “The method of claim 2, wherein applying the model to data present in the dataset further comprises determining learned parameters for the set of distributions of data present in the dataset.”
See Zhao in page 9 lines 16-21 describing “In step 402, the server determines a set of parameters indicating an overall statistical distribution of a variable among the plurality of variables. For example, the server may calculate the set of parameters indicating the overall statistical distribution of a variable among the plurality of variables, based on the distribution information received from each client, by combining the statistical distributions of that variable in each tabular dataset associated with each client.” Here, Zhao explicitly establishes determining parameters of a set of distribution of data. In the previous limitation Zhao established the server contains establishing a model to data. The parameters here are learned as they indicate an overall statistical distribution of a variable among a plurality, the system gains this by input which can be seen as being learned.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base references of Ding and Niu with the teachings of Zhao by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and Niu’s teachings of determining a set of cross-class neighborhood similarities, and incorporate with Zhao’s teachings of receiving data for the dataset in a tabular, and modeling data belonging to a class as a mixture of Gaussian distributions.
One of ordinary skill in the art would be motivated to do so because by integrating Zhao’s frameworks into the methods of Ding and Niu, one of ordinary skill in the art would bring “domain-specific machine learning (ML) models to learn relevant portions of a conversation that trigger assistance and identify relevant assistance information (e.g., documents, articles, phone numbers, website addresses, physical addresses, contact information, etc.). Accurately triggering the need to provide assistance (e.g., when to provide assistance to an agent) and the identification of relevant assistance information (e.g., what assistance to provide to an agent) reduces computing resources required by a system and beneficially provides superior chat assistance in the form of relevant recommended subject matter at relevant time(s) in the conversation” (Niu, column 2 lines 19-30).
Claim 4:
Regarding claim 4, Ding in view of Niu and further in view of Zhao teaches the limitations of claim 3.
Further, Ding teaches “The method of claim 3, wherein determining the set of cross-class neighborhood similarities comprises using the learned parameters to compute a value of the cross-class neighborhood similarities, wherein the nearest neighbor graph is built based on the value of the cross-class neighborhood similarities.”
See Ding in Similarity function section on page 3 describing “Suppose we have the training samples {Xj,yj}Ni=1. Then, we define a new similarity function between xi and xj as follows:
PNG
media_image6.png
124
772
media_image6.png
Greyscale
From (9), we know that the similarity functions for the samples with the same label and the ones with the different labels are different. Specifically, the previous ones are endowed with larger weights and the latter ones are endowed with smaller weights. Fig. 1 shows the curves of similarity function G(xj,xj) vs the Euclidean distance between xi and xj. When these two samples belong to the same class, the similarity rapidly decreases with the increase of their distance. If they belong to different classes, the similarity slowly decreases with the increase of their distance.” Here, Ding establishes a similarity function for computing a similarity between different samples over a set of samples. The similarity is computed for not just same class but different classes which can be seen as computing a value using the function for a cross-class neighborhood similarity. The labels mentioned for each of the samples can also be seen as the parameters learned. Further, see Ding in Construction of adjacent graphs section on page 4 describing “Now, we consider the construction of adjacent graphs according to the new similarity function (9). Our scheme is to select the farthest homogeneous neighbors for a sample to construct an intra-class graph Fw, and it's nearest heterogeneous neighbors to build an inter-class graph Fb.” Here, Ding establishes a construction or building of a graph based on nearest neighbors using the similarity function, which does give a value.
Claim(s) 5 is rejected under 35 U.S.C. 103 as being unpatentable over Ding C. et al., in view of Niu W. et al., further in view of Zhao Z. et al., and further in view of Fujiwara Y. et al. "Fast Similarity Computation for t-SNE", available at https://ieeexplore.ieee.org/document/9458668, published on June 22, 2021, (hereafter Fujiwara).
Claim 5:
Regarding claim 5, Ding in view of Niu and further in view of Zhao teaches the limitations of claim 4.
Neither Ding, Niu or Zhao appears to teach “The method of claim 4, wherein the value of the cross-class neighborhood similarities is computed using Monte Carlo simulations.” However, Fujiwara in the same field of art teaches “The method of claim 4, wherein the value of the cross-class neighborhood similarities is computed using Monte Carlo simulations.”
See Fujiwara in Related Work section on pages 6-7 describing “Monte Carlo-based MR is the state-of-the-art approach based on graph sampling to approximately compute the similarities given by the linear system of the graph Laplacian [29]. This approach consists of two phases: the simulation phase and the estimation phase. In the simulation phase, it iteratively simulates weighted random walks from data points to compute similarities. In the estimation phase, it estimates the similarity scores based on the simulated weighted random walks.” Here, Fujiwara explicitly establishes using Monte Carlo based approach to compute similarities, which in an analogous system can be viewed as cross-class neighborhood similarities, which has a simulation phase making the computation being made using Monte Carlo simulations.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base references of Ding, Niu, and Zhao with the teachings of Fujiwara by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, Niu’s teachings of determining a set of cross-class neighborhood similarities, and Zhao’s teachings of receiving data for the dataset in a tabular, and modeling data belonging to a class as a mixture of Gaussian distributions, and incorporate with Fujiwara’s teachings of computing similarities using Monte Carlo simulations.
One of ordinary skill in the art would be motivated to do so because by integrating Fujiwara’s frameworks into the methods of Ding, Niu and Zhao, one of ordinary skill in the art would bring “an approach to effectively use the sparseness of LDL decomposed matrices to compute a similarity matrix for landmark points efficiently” (Fujiwara, introduction section pages 1-2).
Claim(s) 7 is rejected under 35 U.S.C. 103 as being unpatentable over Ding C. et al., in view of Niu W. et al., and further in view of Yang J. et al., (US Patent Application Publication US20240257907A1) effectively filed on January 20, 2023, (hereafter Yang).
Claim 7:
Regarding claim 7, Ding in view of Niu teaches the limitations of claim 1.
Neither Ding or Niu appears to teach “The method of claim 1, wherein the model applied to the data present in the dataset is a Hierarchical Naïve Bayes model, wherein the Hierarchical Naïve Bayes model models the set of distributions of the data as a mixture of Gaussian distributions, and wherein mixing weights of the mixture are obtained by another mixture of categorical distributions.” However, Yang in the same field of art teaches “The method of claim 1, wherein the model applied to the data present in the dataset is a Hierarchical Naïve Bayes model, wherein the Hierarchical Naïve Bayes model models the set of distributions of the data as a mixture of Gaussian distributions, and wherein mixing weights of the mixture are obtained by another mixture of categorical distributions.”
See Yang in paragraph [0109] describing “For instance, in some embodiments, the model 116 includes one or more gradient boosting models 116, one or more random forest models 116, one or more neural network (NN) models 116, one or more regression models, one or more Naïve Bayes models 116, one or more machine learning algorithms (MLA) 116, or a combination thereof. In some embodiments, an MLA or a NN is trained from a training data set that includes one or more features identified from a data set. MLAs include supervised algorithms (such as algorithms where the features/classifications in the data set are annotated) using linear regression, logistic regression, decision trees, classification and regression trees, Naïve Bayes, nearest neighbor clustering; unsupervised algorithms (such as algorithms where no features/classification in the data set are annotated a priori), such as means clustering, principal component analysis, random forest, adaptive boosting; and semi-supervised algorithms (such as algorithms where an incomplete number of features/classifications in the data set are annotated) using generative approach (such as a mixture of Gaussian distributions, mixture of multinomial distributions, hidden Markov models), low density separation, graph-based approaches (such as minimum cut, harmonic function, manifold regularization, etc.), heuristic approaches, or support vector machines” Here, Yang explicitly states a model being applied to train a data set, the model can be one or more Naïve Bayes models which implies a hierarchical Naïve Bayes model. The models machine learning algorithm is then said to comprise of using a generative approach such as a mixture of Gaussian distributions, the dataset is also said to comprise of annotated features/classifications, which can be seen as categories, uses the gaussian distribution mixture or another mixture mentioned.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base references of Ding and Niu with the teachings of Yang by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and Niu’s teachings of determining a set of cross-class neighborhood similarities, and incorporate with Yang’s teachings of using a Hierarchical Naïve bayes model to model sets of distributions of data as a mixture of gaussian distributions.
One of ordinary skill in the art would be motivated to do so because by integrating Yang’s frameworks into the methods of Ding and Niu, one of ordinary skill in the art would bring “systems and methods that integrate: 1) library design using various in silico detection methods to measure specifics of single mutations and rule out single deleterious mutations: 2) intelligent library construction that makes mutants with a wide range of mutation rates and allows the deep interactions of different mutations: 3) library screening that generates a diverse set of functional data for the mutants: 4) library mutation-function results that encode an effective learning to acquire a surrogate model without over-fitting: 5) optimal library design using a search model to guide the selection of single residue mutations and mutation rate range; and 6) construction and screening of the optimal library to obtain the protein candidates with improved properties. Using the disclosed methodology, desired properties of target proteins are realized in as few as two rounds of evolution” (Yang, paragraph [0012]).
Claim(s) 8 is rejected under 35 U.S.C. 103 as being unpatentable over Ding C. et al., in view of Niu W. et al., and further in view of Wang et al. " Classification in Networked Data with Heterophily", available at https://onlinelibrary.wiley.com/doi/10.1155/2013/236769, published on April 30, 2013, (hereafter Wang).
Claim 8:
Regarding claim 8, Ding in view of Niu teaches the limitations of claim 1.
Neither Ding or Niu appears to teach “The method of claim 1, wherein determining the set of distributions comprises computing a probability that a first node, belonging to a first class, has a nearest neighbor node, belonging to a second class.” However, Wang in the same field of art teaches “The method of claim 1, wherein determining the set of distributions comprises computing a probability that a first node, belonging to a first class, has a nearest neighbor node, belonging to a second class.”
See Wang in Method Based on Class Propagating Distributions section on page 2 describing “In the networks with heterophily, most of connected nodes have different classes. In this case, those classification methods based on the classes of the neighbor nodes lose their effectiveness. In this study, the scope on which we focus is expanded to the neighbors of their neighbor nodes and a probabilistic approach is utilized to calculate the class labels of unlabeled nodes. Let
PNG
media_image7.png
28
27
media_image7.png
Greyscale
denote the probability that the node
PNG
media_image8.png
25
21
media_image8.png
Greyscale
has the class
PNG
media_image9.png
24
19
media_image9.png
Greyscale
.. The vector
PNG
media_image10.png
36
120
media_image10.png
Greyscale
is the class distribution of
PNG
media_image11.png
23
20
media_image11.png
Greyscale
where M is the number of classes.
PNG
media_image12.png
32
111
media_image12.png
Greyscale
. The probabilities
PNG
media_image13.png
31
80
media_image13.png
Greyscale
can represent the class of the node
PNG
media_image8.png
25
21
media_image8.png
Greyscale
. In the network G, some nodes are labeled while others are unlabeled. For the labeled nodes, the class distribution is the vector in which only one element is 1 and the rest are 0. Calculating the classes of unlabeled nodes is to calculate the class distributions of unlabeled nodes.” Here, Wang establishes that connected nodes have different classes, implying that a node can belong to a first and second class. Wang, further establishes that the neighbor nodes are calculated using probability of having a class, implying computing a probability that a first node belonging to a first class, may have a nearest neighbor node that belongs to a different or second class.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base references of Ding and Niu with the teachings of Wang by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and Niu’s teachings of determining a set of cross-class neighborhood similarities, and incorporate with Wang’s teachings of computing a probability of a first node belonging to a first class, has a neighbor node belonging to a second class.
One of ordinary skill in the art would be motivated to do so because by integrating Wang’s frameworks into the methods of Ding and Niu, one of ordinary skill in the art would bring a “proposed algorithm [that] provides better performance on the networks with heterophily” (Yang, page 1 Introduction section).
Claim(s) 6, 11, and 13 are rejected under 35 U.S.C. 103 as being unpatentable over Ding C. et al., in view of Niu W. et al., and further in view of Goldner D. et al., (International Application Publication WO2021247922A1) effectively filed on June 3, 2021, (hereafter Goldner).
Claim 6:
Regarding claim 6, Ding in view of Niu teaches the limitations of claim 1.
Further, Niu teaches “The method of claim 1, further comprising: training a graph machine learning model based on the nearest neighbor graph;”
See Niu in column 12 lines 22-27 describing “The model training system 120 may also determine anchor-positive pairs according to mutually reinforced pretraining. At each iteration of training, a machine learning model such as base ML model 121A mines anchor-positive pairs from a document and reinforces the learning at that iteration by training using the mined pairs.” Here, Niu establishes a model training system utilizing a machine learning model and getting trained based pairs from documents. Further, see Niu in column 12 lines 22-27 describing “The model training system 120 clusters the documents using connected components of a k-nearest neighbor graph. In some implementations, clustering the documents may result in a skewed distribution of cluster sizes (e.g., the hubness phenomenon, in which a small number of documents are nearest neighbors to a disproportionality large number of documents). In these implementations, the model training system 120 utilizes cross-domain similarity local scaling (CSLS), or any other variant, to establish a range for the nearest neighbors. Subsequently, the base ML model 121A mines anchor-positive pairs from clusters of documents (as opposed to a single document).” Here, Niu establishes the model training system using the trained documents in relation to a nearest neighbor graph, implying that the training system is trained using the nearest neighbor graph and further establishing that the model utilized by the training system can be viewed as a graph learning model.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base reference of Ding with the teachings of Niu by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and incorporate with Niu’s teachings of determining a set of cross-class neighborhood similarities.
One of ordinary skill in the art would be motivated to do so because by integrating Niu’s frameworks into the methods of Ding, which are both in relation to nearest neighbor graph and computing similarity, one of ordinary skill in the art would bring “domain-specific machine learning (ML) models to learn relevant portions of a conversation that trigger assistance and identify relevant assistance information (e.g., documents, articles, phone numbers, website addresses, physical addresses, contact information, etc.). Accurately triggering the need to provide assistance (e.g., when to provide assistance to an agent) and the identification of relevant assistance information (e.g., what assistance to provide to an agent) reduces computing resources required by a system and beneficially provides superior chat assistance in the form of relevant recommended subject matter at relevant time(s) in the conversation” (Niu, column 2 lines 19-30).
Neither Ding or Niu appears to teach “and performing predictive tasks using the trained graph machine learning model.”
However, Goldner in an analogous system teaches “and performing predictive tasks using the trained graph machine learning model.”
See Goldner in paragraph [0034] describing “In some implementations, the system 100 can collect and analyze periodically (e.g., hourly, daily, weekly, monthly, etc.) the user health data. The system 100 can identify (e.g., forecast) a health event (e.g., high or low blood pressure, high or low blood glucose levels, risk of cardiovascular disease, etc.) of the user based on the health data. The system 100 can select (or generate) a self-care mode with an action of sequence of actions (e.g., exercise, diet, sleep, reduce stress, etc.) for the user to perform to mitigate the risk of the health event or avoid the health event.” Here, Goldner establishes performing a predictive task of identification of a health risk using a system. Further, see Goldner in paragraph [0023] describing “the system 100 can include one or more analyzing devices” Here, Goldner establishes that the system can include analyzing devices. Further, see Goldner in paragraph [0037] describing “In some embodiments, the analyzing devices 102 is configured to analyze the input data and generate the output data using one or more machine learning models 122. The machine learning models 122 can include supervised learning models, unsupervised learning models, semi-supervised learning models, and/or reinforcement learning models generated by one or more modeling engines 112. Examples of machine learning models suitable for use with the present technology include, but are not limited to: regression algorithms (e.g., ordinary least squares regression, linear regression, logistic regression, stepwise regression, multivariate adaptive regression splines, locally estimated scatterplot smoothing), instance-based algorithms (e.g., k-nearest neighbor, learning vector quantization, self-organizing map, locally weighted learning, support vector machines), ” Here, Goldner establishes that the system which can include analyzing devices, has machine learning models, these models have algorithms which includes k-nearest neighbor which can be a graph training model.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base references of Ding and Niu with the teachings of Goldner by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and Niu’s teachings of determining a set of cross-class neighborhood similarities, and incorporate with Goldner’s teachings of performing predictive tasks using a machine learning model.
One of ordinary skill in the art would be motivated to do so because by integrating Goldner’s frameworks into the methods of Ding and Niu, one of ordinary skill in the art would bring “improved systems and methods for biomonitoring and/or providing personalized healthcare recommendations or information for the treatment of diabetes and other chronic conditions.” (Goldner, paragraph [0004]).
Claim 11:
Regarding claim 11, Ding in view of Niu teaches the limitations of claim 1.
Further, Ding teaches “The method of claim 1, wherein: the nearest neighbor graph is built based on…the dataset;”
See Ding in Construction of adjacent graphs section on pages 3-4 and Figure 2 describing “Now, we consider the construction of adjacent graphs according to the new similarity function (9). Our scheme is to select the farthest homogeneous neighbors for a sample to construct an intra-class graph Fw, and it's nearest heterogeneous neighbors to build an inter-class graph Fb. The reason can be illustrated by Fig. 2. In Fig. 2(a), there are three classes denoted by solid square, circle and solid triangle. For the hollow circle point, we select the farthest neighbor in the solid circle points, and the nearest neighbors in the solid square and triangle points as shown in Fig. 2(b). Fig. 2(c) ideally gives their images in the subspace. We expect that the farthest homogeneous could be attracted to around the sample and the nearest heterogeneous neighbors could be pushed way from the sample.
PNG
media_image4.png
167
352
media_image4.png
Greyscale
” Here, Ding establishes a subspace of points based on nearest neighbors which can be seen as a nearest neighbor graph. Further, see in Construction of adjacent graphs section on page 4 describing “We try to maximize the difference between the nearest inter-class distance and the farthest intra-class distance so as to make the distance between the same classes is nearer and the distance between the different classes is farther in the projection sub-space.” Here Ding establishes a difference between the nearest intra-class distance and farthest intra-class distance which can be seen as a selection of a first cross-class similarity. Ding then establishes that this difference is projected in a sub-space which as established earlier can be seen as a nearest neighbor graph, the projection can be seen as building the nearest neighbor graph. The selections as established in previous limitations is done over a set of similarities which can be seen as a dataset.
Further, Niu teaches “a graph machine learning model is trained using the nearest neighbor graph;”
See Niu in column 12 lines 22-27 describing “The model training system 120 may also determine anchor-positive pairs according to mutually reinforced pretraining. At each iteration of training, a machine learning model such as base ML model 121A mines anchor-positive pairs from a document and reinforces the learning at that iteration by training using the mined pairs.” Here, Niu establishes a model training system utilizing a machine learning model and getting trained based pairs from documents. Further, see Niu in column 12 lines 22-27 describing “The model training system 120 clusters the documents using connected components of a k-nearest neighbor graph. In some implementations, clustering the documents may result in a skewed distribution of cluster sizes (e.g., the hubness phenomenon, in which a small number of documents are nearest neighbors to a disproportionality large number of documents). In these implementations, the model training system 120 utilizes cross-domain similarity local scaling (CSLS), or any other variant, to establish a range for the nearest neighbors. Subsequently, the base ML model 121A mines anchor-positive pairs from clusters of documents (as opposed to a single document).” Here, Niu establishes the model training system using the trained documents in relation to a nearest neighbor graph, implying that the training system is trained using the nearest neighbor graph and further establishing that the model utilized by the training system can be viewed as a graph learning model.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base reference of Ding with the teachings of Niu by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and incorporate with Niu’s teachings of determining a set of cross-class neighborhood similarities.
One of ordinary skill in the art would be motivated to do so because by integrating Niu’s frameworks into the methods of Ding, which are both in relation to nearest neighbor graph and computing similarity, one of ordinary skill in the art would bring “domain-specific machine learning (ML) models to learn relevant portions of a conversation that trigger assistance and identify relevant assistance information (e.g., documents, articles, phone numbers, website addresses, physical addresses, contact information, etc.). Accurately triggering the need to provide assistance (e.g., when to provide assistance to an agent) and the identification of relevant assistance information (e.g., what assistance to provide to an agent) reduces computing resources required by a system and beneficially provides superior chat assistance in the form of relevant recommended subject matter at relevant time(s) in the conversation” (Niu, column 2 lines 19-30).
Neither Ding or Niu appears to teach “the data present in the dataset comprises electronic health records corresponding to a plurality of patients, wherein the electronic health records comprise heart rate, oxygen saturation, weight, height, glucose, temperature associated with each patient in the plurality of patients; …the electronic health records present in…; and a clinical risk is predicted for a patient using the trained graph machine learning model.”
However, Goldner in an analogous system teaches “the data present in the dataset comprises electronic health records corresponding to a plurality of patients, wherein the electronic health records comprise heart rate, oxygen saturation, weight, height, glucose, temperature associated with each patient in the plurality of patients;”
See Goldner in paragraph [0080] describing “the health data (e.g., the new data and/or the user history 124) can include any data relevant to the user’s health state. Examples of health data include, but are not limited to, any of the following: blood pressure data (e.g., current and/or previous measurements of systolic and/or diastolic blood pressure), blood glucose data (e.g., current and/or previous blood glucose measurements, current and/or previous HbA1 c data values), heart rate data, food data (e.g., number of meals; timing of meals; number of calories; amount of carbohydrates, fats, sugars, etc.), medical history data (e.g., current and/or previous weight, height, BMI, age, sleeping patterns, medical conditions, cholesterol levels, diabetes type, family history, user health history, diagnoses, blood pressure, etc.), activity data (e.g., time and/or duration of activity; activity type such as walking, running, swimming; strenuousness of the activity such as low, moderate, high; etc.), personal data (e.g., name, gender, demographics, social network information, etc.), medication data (e.g., timing and/or dosages of medications such as insulin, prescription and/or non-prescription medications taken), and/or any other suitable data (e.g., basal energy consumption, oxygen consumption) or combination thereof. The health data can be obtained automatically by one or more biosensors (e.g., the biosensor(s) 104a of FIG. 1 ), input manually by the user (e.g., via the user devices 104), retrieved from a database and/or server (e.g., database 106), or any other suitable collection technique.” Here, Goldner establishes data received electronically because it uses biosensors or a user input retrieved over a server comprising of heart rate, oxygen saturation with levels of oxygen consumptions, glucose, height, weight and temperature. All this data comes from any user making the data come from a plurality of users.
Further, Goldner teaches “…the electronic health records present in…”
See Goldner in paragraph [0080] describing “The health data can be obtained automatically by one or more biosensors (e.g., the biosensor(s) 104a of FIG. 1 ), input manually by the user (e.g., via the user devices 104), retrieved from a database and/or server (e.g., database 106), or any other suitable collection technique.” Here, Goldner establishes health data records received electronically because it uses biosensors or a user input retrieved over a server.
Further, Goldner teaches “and a clinical risk is predicted for a patient using the trained graph machine learning model”
See Goldner in paragraph [0034] describing “In some implementations, the system 100 can collect and analyze periodically (e.g., hourly, daily, weekly, monthly, etc.) the user health data. The system 100 can identify (e.g., forecast) a health event (e.g., high or low blood pressure, high or low blood glucose levels, risk of cardiovascular disease, etc.) of the user based on the health data.” Here, Goldner explicitly establishes predicting a clinical risk for a patient with the identification of a health event due to risk. This risk is precited using the system. Further, see Goldner in paragraph [0037] describing “In some embodiments, the analyzing devices 102 is configured to analyze the input data and generate the output data using one or more machine learning models 122. The machine learning models 122 can include supervised learning models, unsupervised learning models, semi-supervised learning models, and/or reinforcement learning models generated by one or more modeling engines 112. Examples of machine learning models suitable for use with the present technology include, but are not limited to: regression algorithms (e.g., ordinary least squares regression, linear regression, logistic regression, stepwise regression, multivariate adaptive regression splines, locally estimated scatterplot smoothing), instance-based algorithms (e.g., k-nearest neighbor, learning vector quantization, self-organizing map, locally weighted learning, support vector machines), ” Here, Goldner establishes that the system which can include analyzing devices, has machine learning models, these models have algorithms which includes k-nearest neighbor which can be a graph training model.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base references of Ding and Niu with the teachings of Goldner by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and Niu’s teachings of determining a set of cross-class neighborhood similarities, and incorporate with Goldner’s teachings of performing predictive tasks using a machine learning model.
One of ordinary skill in the art would be motivated to do so because by integrating Goldner’s frameworks into the methods of Ding and Niu, one of ordinary skill in the art would bring “improved systems and methods for biomonitoring and/or providing personalized healthcare recommendations or information for the treatment of diabetes and other chronic conditions.” (Goldner, paragraph [0004]).
Claim 13:
Regarding claim 13, Ding in view of Niu teaches the limitations of claim 1.
Further, Ding teaches “The method of claim 1, wherein: the nearest neighbor graph is built based on…the dataset;”
See Ding in Construction of adjacent graphs section on pages 3-4 and Figure 2 describing “Now, we consider the construction of adjacent graphs according to the new similarity function (9). Our scheme is to select the farthest homogeneous neighbors for a sample to construct an intra-class graph Fw, and it's nearest heterogeneous neighbors to build an inter-class graph Fb. The reason can be illustrated by Fig. 2. In Fig. 2(a), there are three classes denoted by solid square, circle and solid triangle. For the hollow circle point, we select the farthest neighbor in the solid circle points, and the nearest neighbors in the solid square and triangle points as shown in Fig. 2(b). Fig. 2(c) ideally gives their images in the subspace. We expect that the farthest homogeneous could be attracted to around the sample and the nearest heterogeneous neighbors could be pushed way from the sample.
PNG
media_image4.png
167
352
media_image4.png
Greyscale
” Here, Ding establishes a subspace of points based on nearest neighbors which can be seen as a nearest neighbor graph. Further, see in Construction of adjacent graphs section on page 4 describing “We try to maximize the difference between the nearest inter-class distance and the farthest intra-class distance so as to make the distance between the same classes is nearer and the distance between the different classes is farther in the projection sub-space.” Here Ding establishes a difference between the nearest intra-class distance and farthest intra-class distance which can be seen as a selection of a first cross-class similarity. Ding then establishes that this difference is projected in a sub-space which as established earlier can be seen as a nearest neighbor graph, the projection can be seen as building the nearest neighbor graph. The selections as established in previous limitations is done over a set of similarities which can be seen as a dataset.
Further, Niu teaches “a graph machine learning model is trained using the nearest neighbor graph;”
See Niu in column 12 lines 22-27 describing “The model training system 120 may also determine anchor-positive pairs according to mutually reinforced pretraining. At each iteration of training, a machine learning model such as base ML model 121A mines anchor-positive pairs from a document and reinforces the learning at that iteration by training using the mined pairs.” Here, Niu establishes a model training system utilizing a machine learning model and getting trained based pairs from documents. Further, see Niu in column 12 lines 22-27 describing “The model training system 120 clusters the documents using connected components of a k-nearest neighbor graph. In some implementations, clustering the documents may result in a skewed distribution of cluster sizes (e.g., the hubness phenomenon, in which a small number of documents are nearest neighbors to a disproportionality large number of documents). In these implementations, the model training system 120 utilizes cross-domain similarity local scaling (CSLS), or any other variant, to establish a range for the nearest neighbors. Subsequently, the base ML model 121A mines anchor-positive pairs from clusters of documents (as opposed to a single document).” Here, Niu establishes the model training system using the trained documents in relation to a nearest neighbor graph, implying that the training system is trained using the nearest neighbor graph and further establishing that the model utilized by the training system can be viewed as a graph learning model.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base reference of Ding with the teachings of Niu by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and incorporate with Niu’s teachings of determining a set of cross-class neighborhood similarities.
One of ordinary skill in the art would be motivated to do so because by integrating Niu’s frameworks into the methods of Ding, which are both in relation to nearest neighbor graph and computing similarity, one of ordinary skill in the art would bring “domain-specific machine learning (ML) models to learn relevant portions of a conversation that trigger assistance and identify relevant assistance information (e.g., documents, articles, phone numbers, website addresses, physical addresses, contact information, etc.). Accurately triggering the need to provide assistance (e.g., when to provide assistance to an agent) and the identification of relevant assistance information (e.g., what assistance to provide to an agent) reduces computing resources required by a system and beneficially provides superior chat assistance in the form of relevant recommended subject matter at relevant time(s) in the conversation” (Niu, column 2 lines 19-30).
Neither Ding or Niu appears to teach “the data present in the dataset comprises soil data corresponding to a plurality of areas, wherein the soil data comprises humidity, temperature, and performance metrics related to different areas;…the soil data present in…; and a quality of an input soil type is predicted based on the nearest neighbor graph.”
However, Goldner in an analogous system teaches “the data present in the dataset comprises soil data corresponding to a plurality of areas, wherein the soil data comprises humidity, temperature, and performance metrics related to different areas;”
See Goldner in paragraph [0026] describing “The input data for the system 100 can include health-related information, contextual information, and/or any other information relevant to the patient’s health state.” Here, Goldner establishes data which can comprise of contextual information. Further, see Goldner in paragraph [0026] describing “Contextual information can include user location (e.g., GPS coordinates, elevation data), environmental conditions (e.g., air pressure, humidity, temperature, air quality, etc.), and/or combinations thereof.” Here, Goldner establishes the contextual information can include environmental conditions, which can be soil data, this information comprising humidity, temperature, and other metrics which can be seen as performance metrics.
Further, Goldner teaches “…the soil data present in…”
See Goldner in paragraph [0026] describing “The input data for the system 100 can include health-related information, contextual information, and/or any other information relevant to the patient’s health state.” Here, Goldner establishes data which can comprise of contextual information. Further, see Goldner in paragraph [0026] describing “Contextual information can include user location (e.g., GPS coordinates, elevation data), environmental conditions (e.g., air pressure, humidity, temperature, air quality, etc.), and/or combinations thereof.” Here, Goldner establishes the contextual information can include environmental conditions, which can be soil data.
Further, Goldner teaches “and a quality of an input soil type is predicted based on the nearest neighbor graph”
See Goldner in paragraph [0132] describing “The auxiliary data 1516 can be selected by the system 1410 to modify the adaptive support machine-learning model based on received indication of the response. The auxiliary data 1416 can include predictions (e.g., short-term predictions, long-term predictions, forecasted events, etc.), environment data (e.g., weather data, temperature data, etc.), or the like. The auxiliary data 1516 can be inputted to models to generate output data based on non-user specific parameters.” Here, Goldner establishes that an auxiliary data can be selected by a system which includes a prediction of environmental data which for example, can be quality of a soil type. Goldner also establishes that this data can be input into a model. As established in a previous limitation, in ana analogous system the model used is trained sing a nearest neighbor graph.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base references of Ding and Niu with the teachings of Goldner by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and Niu’s teachings of determining a set of cross-class neighborhood similarities, and incorporate with Goldner’s teachings of performing predictive tasks using a machine learning model.
One of ordinary skill in the art would be motivated to do so because by integrating Goldner’s frameworks into the methods of Ding and Niu, one of ordinary skill in the art would bring “improved systems and methods for biomonitoring and/or providing personalized healthcare recommendations or information for the treatment of diabetes and other chronic conditions.” (Goldner, paragraph [0004]).
Claim(s) 10 is rejected under 35 U.S.C. 103 as being unpatentable over Ding C. et al., in view of Niu W. et al., and further in view of Sadayappan P. et al. "Nearest-Neighbor Mapping of Finite Element Graphs onto Processor Meshes", available at https://ieeexplore.ieee.org/document/5009494, published on December 31, 1987, (hereafter Sadayappan).
Claim 10:
Regarding claim 10, Ding in view of Niu teaches the limitations of claim 1.
Neither Ding or Niu appears to teach “The method of claim 1, wherein the nearest neighbor graph comprises: a selected node at a center of hypercube, wherein the hypercube comprises an edge that is optimized; and a set of neighbors of the selected node within the hypercube based on the edge of the hypercube, wherein the hypercube is formed based on the first parameter.” However, Sadayappan in an analogous system teaches “The method of claim 1, wherein the nearest neighbor graph comprises: a selected node at a center of hypercube, wherein the hypercube comprises an edge that is optimized;”
See Sadayappan in Performance evaluation section on page 14 describing “Since a hypercube interconnection is a superset of many two dimensional mesh configurations, by using a subset of the physical links of the 16 processor hypercube, a 4 x 4 processor mesh was obtained.” Here, Sadayappan establishes a hypercube which is formed with processors. Further, see Sadayappan in The Nearest-Neighbor Approach to mapping section on page 4 describing “Our approach to mapping uses a two-step procedure. 1) Creation of an initial nearest-neighbor partition, where the vertices of the finite element graph are partitioned into clusters and the clusters allocated to processors so that any two vertices sharing an edge are either mapped to the very same processor, or to neighbor processors. 2) Iterative boundary refinement involving incremental modification of the initial partition of step 1) above, so that the nearest-neighbor property is always maintained, but better balancing of computational load among the processors is achieved.” Here, Sadayappan establishes an edge that is mapped to a processor or neighbor processors in an optimal way, with the clusters being allocated to processors so that any two vertices that share an edge are mapped to the very same processor or neighbor processor. Further, see Sadayappan in The Nearest-Neighbor Approach to mapping section on page 4 describing “The boundary refinement procedure may be understood by reference to Fig. 3. Fig. 3(a) shows an initial nearest-neighbor partition that maps the finite element graph onto a regular mesh of four processors (2 x 2). Vertices contained by common finite element represent processes that need communicate; such vertex pairs can be seen to be assigned to either the same processor or to immediate neighbor processors. The initial partition, however, does not distribute the computational load uniformly among the processors-processors 1, 2, 3, and 4 are allocated 6, 12, 11, and 11 processes, respectively. One way of balancing the computational loads of the processors is to reassign some of the assigned nodes among the processors; for example, transferring two processes from P2 to P1, and one process from P3 to P4 will do so. If this reassignment of processes is done in such a way that the nearest-neighbor property is still maintained, then the final mapping will be a load-balanced, nearest-neighbor mapping. This is shown in Fig. 3(c) for the chosen example.” Here, Sadayappan establishes an initial nearest neighbor partition which can be seen as a selected node here that is put on a mesh of processors, which is said to compromise the hypercube. Further, see Sadayappan in Figure 3, “
PNG
media_image14.png
335
468
media_image14.png
Greyscale
”. Here we can see the initial partition over a graph with a center point. The partition is centered to the best it can on this graph and since the partition is said to be placed on a mesh of processors, be seen as the selected node, and a hypercube comprises of multiple processors, the center of this processor can be seen as the center of the hypercube.
Further, Sadayappan teaches “and a set of neighbors of the selected node within the hypercube based on the edge of the hypercube, wherein the hypercube is formed based on the first parameter”
See Sadayappan in The Nearest-Neighbor Approach to mapping section on page 4 describing “The boundary refinement procedure may be understood by reference to Fig. 3. Fig. 3(a) shows an initial nearest-neighbor partition that maps the finite element graph onto a regular mesh of four processors (2 x 2). Vertices contained by common finite element represent processes that need communicate; such vertex pairs can be seen to be assigned to either the same processor or to immediate neighbor processors. The initial partition, however, does not distribute the computational load uniformly among the processors-processors 1, 2, 3, and 4 are allocated 6, 12, 11, and 11 processes, respectively. One way of balancing the computational loads of the processors is to reassign some of the assigned nodes among the processors; for example, transferring two processes from P2 to P1, and one process from P3 to P4 will do so. If this reassignment of processes is done in such a way that the nearest-neighbor property is still maintained, then the final mapping will be a load-balanced, nearest-neighbor mapping. This is shown in Fig. 3(c) for the chosen example.” Here, Sadayappan establishes an initial nearest neighbor partition which can be seen as a selected node as mentioned before and is understood to comprise of set of neighbors as known in the art. The partition is placed on a mesh of processors. Further, see Sadayappan in Performance evaluation section on page 14 describing “Experimentally measuring the obtained speedup upon executing a parallel program with the task interaction graph specified by the sample finite element graphs on a 16 processor Intel iPSC hypercube parallel computer. The parameters for the estimation above were obtained empirically through measurements on the same parallel computer system. Since a hypercube interconnection is a superset of many two dimensional mesh configurations, by using a subset of the physical links of the 16 processor hypercube, a 4 x 4 processor mesh was obtained.” Here, Sadayappan establishes parameters that are obtained through measurements on a computer system which is a hypercube. A hypercube is established to consists of mesh configurations and is obtained by such and is based on parameters we can see this as a hypercube being formed by parameters. Although a first parameter was never established, regardless a first parameter is needed in the parameter(s) inherently to also do this formation.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base references of Ding and Niu with the teachings of Sadayappan by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and Niu’s teachings of determining a set of cross-class neighborhood similarities, and incorporate with Sadayappan’s teachings of nearest neighbor graph comprising a hypercube.
One of ordinary skill in the art would be motivated to do so because by integrating Sadayappan’s frameworks into the methods of Ding and Niu, one of ordinary skill in the art would bring “a two-step procedure: 1) generation of an initial "nearest-neighbor" mapping, and 2) "boundary refinement" to improve load balancing.” (Sadayappan, page 1 Introduction section).
Claim(s) 12 is rejected under 35 U.S.C. 103 as being unpatentable over Ding C. et al., in view of Niu W. et al., and further in view of Schaeffer J. et al., (US Patent Application Publication US20220044826 A1) effectively filed on February 10, 2022, (hereafter Schaeffer).
Claim 12:
Regarding claim 12, Ding in view of Niu teaches the limitations of claim 1.
Further, Ding teaches “The method of claim 1, wherein: the nearest neighbor graph is built based on…the dataset;”
See Ding in Construction of adjacent graphs section on pages 3-4 and Figure 2 describing “Now, we consider the construction of adjacent graphs according to the new similarity function (9). Our scheme is to select the farthest homogeneous neighbors for a sample to construct an intra-class graph Fw, and it's nearest heterogeneous neighbors to build an inter-class graph Fb. The reason can be illustrated by Fig. 2. In Fig. 2(a), there are three classes denoted by solid square, circle and solid triangle. For the hollow circle point, we select the farthest neighbor in the solid circle points, and the nearest neighbors in the solid square and triangle points as shown in Fig. 2(b). Fig. 2(c) ideally gives their images in the subspace. We expect that the farthest homogeneous could be attracted to around the sample and the nearest heterogeneous neighbors could be pushed way from the sample.
PNG
media_image4.png
167
352
media_image4.png
Greyscale
” Here, Ding establishes a subspace of points based on nearest neighbors which can be seen as a nearest neighbor graph. Further, see in Construction of adjacent graphs section on page 4 describing “We try to maximize the difference between the nearest inter-class distance and the farthest intra-class distance so as to make the distance between the same classes is nearer and the distance between the different classes is farther in the projection sub-space.” Here Ding establishes a difference between the nearest intra-class distance and farthest intra-class distance which can be seen as a selection of a first cross-class similarity. Ding then establishes that this difference is projected in a sub-space which as established earlier can be seen as a nearest neighbor graph, the projection can be seen as building the nearest neighbor graph. The selections as established in previous limitations is done over a set of similarities which can be seen as a dataset.
Further, Niu teaches “a graph machine learning model is trained using the nearest neighbor graph;”
See Niu in column 12 lines 22-27 describing “The model training system 120 may also determine anchor-positive pairs according to mutually reinforced pretraining. At each iteration of training, a machine learning model such as base ML model 121A mines anchor-positive pairs from a document and reinforces the learning at that iteration by training using the mined pairs.” Here, Niu establishes a model training system utilizing a machine learning model and getting trained based pairs from documents. Further, see Niu in column 12 lines 22-27 describing “The model training system 120 clusters the documents using connected components of a k-nearest neighbor graph. In some implementations, clustering the documents may result in a skewed distribution of cluster sizes (e.g., the hubness phenomenon, in which a small number of documents are nearest neighbors to a disproportionality large number of documents). In these implementations, the model training system 120 utilizes cross-domain similarity local scaling (CSLS), or any other variant, to establish a range for the nearest neighbors. Subsequently, the base ML model 121A mines anchor-positive pairs from clusters of documents (as opposed to a single document).” Here, Niu establishes the model training system using the trained documents in relation to a nearest neighbor graph, implying that the training system is trained using the nearest neighbor graph and further establishing that the model utilized by the training system can be viewed as a graph learning model.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base reference of Ding with the teachings of Niu by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and incorporate with Niu’s teachings of determining a set of cross-class neighborhood similarities.
One of ordinary skill in the art would be motivated to do so because by integrating Niu’s frameworks into the methods of Ding, which are both in relation to nearest neighbor graph and computing similarity, one of ordinary skill in the art would bring “domain-specific machine learning (ML) models to learn relevant portions of a conversation that trigger assistance and identify relevant assistance information (e.g., documents, articles, phone numbers, website addresses, physical addresses, contact information, etc.). Accurately triggering the need to provide assistance (e.g., when to provide assistance to an agent) and the identification of relevant assistance information (e.g., what assistance to provide to an agent) reduces computing resources required by a system and beneficially provides superior chat assistance in the form of relevant recommended subject matter at relevant time(s) in the conversation” (Niu, column 2 lines 19-30).
Neither Ding or Niu appears to teach “the data present in the dataset comprises genomic activity information corresponding to a plurality of patients, wherein the genomic activity of each patient identifies a response of the respective patient to a drug; …the genomic activity information present in…; and a suitability of a patient for a drug trial is predicted using the graph machine learning model”
However, Schaeffer in an analogous system teaches “the data present in the dataset comprises genomic activity information corresponding to a plurality of patients, wherein the genomic activity of each patient identifies a response of the respective patient to a drug;”
Further, see Schaeffer in paragraph [0175] describing “With regard to the predictive model, in various embodiments a plurality of data is obtained or received for a plurality of patients, covering a period of time (e.g. a time span covering each of the patients' medical history from the time of their diagnosis until the current time or a time of death, medical history may also begin before diagnosis).” Here, Schaeffer establishes a data for a plurality of patients with history in regards to time. Further, see Schaeffer in paragraph [0186] describing “The prior features may include various features related to a patient's medical condition and/or treatment. In various embodiments the prior features may include temporal/time-based events or features, structural or biological features, or molecular/genetic features, among other categories. In particular embodiments the prior features may include one or more of: time since starting a particular medication; time since taking a particular medication; time since last progressive therapy outcome (e.g. patient response to drug); time since metastasis; largest tumor size to date/last recorded tumor size; most severe effect of identified SNP (e.g. low effect, high effect); or RNA features (e.g. expression level per gene/transcript).” Here, Schaeffer establishes that the data includes features which have genomic information and activity like genetic features and progressive therapy outcome, and also includes patients response to drugs explicitly.
Further, Schaeffer teaches “…the genomic activity information present in…”
See Schaeffer in paragraph [0096] describing “When applying patient similarity analytics to a patient's health record, such as through a laboratory report user interface, a number of predefined/predetermined selection criteria may be extracted or referenced from one or more genomic test results. Genomic test results may exist in a number of different formats. In one example, those formats may include a specific assay that was performed, such as a whole genome sequencing, a limited genome sequencing tailored to the particular disease state, a tissue sequencing, tumor sequencing, tumor-normal sequencing, a liquid biopsy sequencing, a cell-free DNA sequencing, DNA sequencing, RNA/transcriptome sequencing, next-generation sequencing panels, or other assays for identifying genomic alterations, fusions, or molecular biomarkers within a patient's genome.” Here, Schaeffer establishes a patient’s health record comprising of genomic test results and activity.
Further, Schaeffer teaches “and a suitability of a patient for a drug trial is predicted using the graph machine learning model”
See Schaeffer in paragraph [0028] describing “Artificial intelligence models referenced herein may be gradient boosting models, random forest models, neural networks (NN), regression models, Naive Bayes models, or machine learning algorithms (MLA). A MLA or a NN may be trained from a training data set. In an exemplary prediction profile, a training data set may include imaging, pathology, clinical, and/or molecular reports and details of a patient, such as those curated from an EHR or genetic sequencing reports. MLAs include supervised algorithms (such as algorithms where the features/classifications in the data set are annotated) using linear regression, logistic regression, decision trees, classification and regression trees, Naïve Bayes, nearest neighbor clustering; unsupervised algorithms (such as algorithms where no features/classification in the data set are annotated) using Apriori, means clustering, principal component analysis, random forest, adaptive boosting; and semi-supervised algorithms (such as algorithms where an incomplete number of features/classifications in the data set are annotated) using generative approach (such as a mixture of Gaussian distributions, mixture of multinomial distributions, hidden Markov models), low density separation, graph-based approaches (such as mincut, harmonic function, manifold regularization), heuristic approaches, or support vector machines.” Here, Schaeffer establishes a machine learning model that can use a graph based approach in the embodiment. Further, see Schaeffer in paragraph [0218] describing “At this point a plurality of sets of predictions for the plurality of patients may be generated based on the plurality of prior features and the plurality of forward features, and a prediction model may be generated based on the sets of predictions using machine learning. In some embodiments the prediction model may be generated using gradient boosting.” Here, Schaffer establishes a prediction for a patient based on prior features using a prediction model, which can be a graph machine learning model as established prior. Further, see Schaeffer in paragraph [0186] describing “The prior features may include various features related to a patient's medical condition and/or treatment. In various embodiments the prior features may include temporal/time-based events or features, structural or biological features, or molecular/genetic features, among other categories. In particular embodiments the prior features may include one or more of: time since starting a particular medication; time since taking a particular medication; time since last progressive therapy outcome (e.g. patient response to drug); time since metastasis; largest tumor size to date/last recorded tumor size; most severe effect of identified SNP (e.g. low effect, high effect); or RNA features (e.g. expression level per gene/transcript)..” Here, Schaffer establishes the prior features in relation to a patient’s medical condition or treatment and explicitly states a patients response to drug as an example. The prediction given by the model could then be a suitability of a patient for a drug trial because the example of the response to a drug is based on a temporal/time-based event and features.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the base references of Ding and Niu with the teachings of Schaeffer by using Ding’s teachings of selecting a first cross-class neighborhood similarity from a set of cross-class neighborhood similarities and building a nearest neighbor graph, and Niu’s teachings of determining a set of cross-class neighborhood similarities, and incorporate with Schaeffer’s teachings of performing predictive tasks based on patient information using a machine learning model.
One of ordinary skill in the art would be motivated to do so because by integrating Schaeffer’s frameworks into the methods of Ding and Niu, one of ordinary skill in the art would bring a “system described herein [that] facilitates the discovery of insights of therapeutic significance, through the automated analysis of patterns occurring in patient clinical, molecular, phenotypic, and response data, and enabling further exploration via a fully integrated, reactive user interface.” (Schaeffer, paragraph [0006]).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to HASSAN R SESAY whose telephone number is (571)272-8493. The examiner can normally be reached Monday-Friday 8am-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, Usmaan Saeed can be reached at (571) 272-4046. 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.
/HASSAN RAMADAN SESAY/Examiner, Art Unit 2146
/USMAAN SAEED/Supervisory Patent Examiner, Art Unit 2146
Prosecution Insights