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 .
Status
This instant application No. 19/348,937 has claims 21-40 pending based on the preliminary amendment filed on February 3, 2026.
Priority / Filing Date
Applicant’s claim for priority of parent application No. 18/679,988 (now Pat. No. US 12455903) is acknowledged. The effective filing date for this instant application is May 31, 2024.
Abstract
The abstract of the disclosure is objected due to the use of implied language. Note that in the abstract, the language should be clear and concise and should not repeat information given in the title. It should avoid using phrases which can be implied, such as, “The disclosure concerns,” “The disclosure defined by this invention,” “The disclosure describes,” etc… See MPEP § 608.01(b).
Note that in the abstract, Applicant cites “Systems and methods are provided for implementing large-scale density-based clustering functionalities” on lines 1-2. This citation clearly provokes the use of implied language and repeats the title. Revision and/or correction are required (e.g., removal of the entire first sentence of the abstract).
Drawings
The drawings filed on October 3, 2025 are accepted for examination purposes.
Information Disclosure Statement
As required by M.P.E.P. 609(C), the Applicant’s submission of the Information Disclosure Statement filed on December 1, 2025 is acknowledged by the Examiner and the cited references have been considered in the examination of the claims now pending. As required by M.P.E.P. 609 C(2), a copy of the PTOL-1449 initialed and dated by the Examiner is attached to the instant Office action.
Claim Objections
Claim 26 is objected for citing “…selecting the first upper bound value comprises: generated a sampled dataset…” on lines 1-3. It is believed “…generating a sampled dataset…” should be the appropriate form. Revision and/or correction are required.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
Claims 24 and 33 are rejected under 35 U.S.C. 112(b) for failing to properly point out and distinctly claim the invention.
Regarding claims 24, and 33, each claim recites “a modified tenary search algorithm” which is vague and indefinite since the term “modified” is inherently relative and open-ended that renders the scope of the claim unclear. Each claim recites that a modified ternary search is performed but fails to recite the specific methodological steps, functional logic, or mathematical operations that constitute the modification. A POSITA would not be able to determine with reasonable certainty the boundaries of the claim, as it is unclear how the claimed search differs from a standard ternary search nor is it clear that specific algorithmic steps are required to satisfy the modified limitation. To overcome this rejection, Applicant should explicitly recite the specific algorithmic steps that define the modification to the ternary search as supported by the specification.
Double Patenting
The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the “right to exclude” granted by a patent and to prevent possible harassment by multiple assignees. A nonstatutory double patenting rejection is appropriate where the claims at issue are not identical, but at least one examined application claim is not patentably distinct from the reference claim(s) because the examined application claim is either anticipated by, or would have been obvious over, the reference claim(s). See, e.g., In re Berg, 140 F.3d 1428, 46 USPQ2d 1226 (Fed. Cir. 1998); In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); and In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) or 1.321(d) may be used to overcome an actual or provisional rejection based on a nonstatutory double patenting ground provided the reference application or patent either is shown to be commonly owned with this application, or claims an invention made as a result of activities undertaken within the scope of a joint research agreement. See MPEP § 717.02 for applications subject to examination under the first inventor to file provisions of the AIA as explained in MPEP § 2159. See MPEP §§ 706.02(l)(1) - 706.02(l)(3) for applications not subject to examination under the first inventor to file provisions of the AIA . A terminal disclaimer must be signed in compliance with 37 CFR 1.321(b).
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/forms/. The filing date of the application in which the form is filed determines what form (e.g., PTO/SB/25, PTO/SB/26, PTO/AIA /25, or PTO/AIA /26) should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to http://www.uspto.gov/patents/process/file/efs/guidance/eTD-info-I.jsp.
Claims 21-40 are rejected on the ground of nonstatutory double patenting over claims 1-20 of Pat. No. US 12455903.
Claims 21-40 of the instant application recite similar limitations and claims 1-20 of ‘903 as being compared in the table below. For the purpose of illustration, only claims 21-39 (system and method claims) of the instant application are compared to the claims of the patent (underlining are used to indicate conflict limitations).
Instant Application
Pat. No. US 12455903
Claim 21, 32, and 40
A method/system/device comprising: a processing system; and a memory comprising computer executable instructions that, when executed, perform operations comprising:
receiving a dataset including embeddings of at least one of words, objects in an image, or objects in a video;
selecting, for the dataset, at least one of a first upper bound value or a first lower bound value of a density-based clustering algorithm;
identifying an optimal neighborhood radius parameter value based on the at least one of the first upper bound value or the first lower bound value of a neighborhood radius parameter;
providing as output at least one of: the optimal neighborhood radius parameter value; or an optimal number of clusters within the dataset corresponding to the optimal neighborhood radius parameter value; and
based on the output, clustering the at least one of words, objects in the image, or objects in the video.
Claim 1
A system, comprising: a processing system; and a memory coupled to the processing system, the memory comprising computer executable instructions that, when executed by the processing system, causes the system to perform operations comprising:
receiving a dataset including embeddings of at least one of words, objects in an image, or objects in a video;
selecting, for the dataset, a first upper bound value and a first lower bound value of a neighborhood radius parameter of a density-based clustering algorithm;
identifying, using a modified ternary search algorithm based on near-unimodality of the neighborhood radius parameter as a variant of ternary search algorithm, an optimal neighborhood radius parameter value, based on the first upper bound value and the first lower bound value of the neighborhood radius parameter;
providing as output performing at least one of:
outputting the optimal neighborhood radius parameter value; or outputting an optimal number of clusters within the dataset corresponding to the optimal neighborhood radius parameter value; and
based on the output, clustering the at least one of words, objects in the image, or objects in the video associated with the dataset.
Claim 22
The system of claim 21, wherein selecting the at least one of the first upper bound value or the first lower bound value comprises selecting the first upper bound value and the first lower bound value.
Claim 1
…selecting, for the dataset, a first upper bound value and a first lower bound value of a neighborhood radius parameter of a density-based clustering algorithm…
Claim 23
The system of claim 21, wherein the at least one of the first upper bound value or the first lower bound value is associated with the neighborhood radius parameter of the density-based clustering algorithm.
Claim 1
…selecting, for the dataset, a first upper bound value and a first lower bound value of a neighborhood radius parameter of a density-based clustering algorithm…
Claim 24
The system of claim 21, wherein identifying the optimal neighborhood radius parameter value comprises using a modified ternary search algorithm to perform the identifying.
Claim 1
…identifying, using a modified ternary search algorithm based on near-unimodality of the neighborhood radius parameter as a variant of ternary search algorithm …
Claim 25
The system of claim 24, wherein the modified ternary search algorithm is based on near-unimodality of the neighborhood radius parameter as a variant of ternary search algorithm.
Claim 1
…identifying, using a modified ternary search algorithm based on near-unimodality of the neighborhood radius parameter as a variant of ternary search algorithm…
Claim 26
The system of claim 21, wherein selecting the first upper bound value comprises:
generated a sampled dataset by sampling the dataset at a sampling rate between 10% and 50%; and
identifying the first upper bound value based on a first initial upper bound value and a first initial lower bound value that are selected for the sampled dataset.
Claim 6
The system of claim 1, wherein the first upper bound value is selected by:
generating a second sampled dataset, by sampling the dataset at a second sampling rate, wherein the second sampling rate is selected to be a value between 10% and 50%; and
identifying, using the modified ternary search algorithm, the first upper bound value, based on a first initial upper bound value and a first initial lower bound value that are selected for the second sampled dataset…
Claim 27
The system of claim 26, wherein the first initial upper bound value is one of:
a value of 1, for cosine-based neighborhood radius parameter values;
a maximum neighborhood radius parameter value, for other bounded neighborhood radius parameter values; or
a value of a sum of difference values between a maximum function and a minimum function of a distance variable, for unbounded neighborhood radius parameter values.
Claim 6
…wherein the first initial upper bound value is one of:
a value of 1, for cosine-based neighborhood radius parameter values;
a maximum neighborhood radius parameter value, for other bounded neighborhood radius parameter values; or
a value of a sum of difference values between a maximum function and a minimum function of a distance variable, for unbounded neighborhood radius parameter values.
Claim 28
The system of claim 21, the operations further comprising:
providing, to a language model, the at least one of the optimal neighborhood radius parameter value or the optimal number of clusters within the dataset corresponding to the optimal neighborhood radius parameter value to train the language model to cluster natural language words.
See further Peng below for mapping and motivation to combine with the claims of ‘903.
Claim 29
The system of claim 21, the operations further comprising:
providing, to a computer vision system, the at least one of the optimal neighborhood radius parameter value or the optimal number of clusters within the dataset corresponding to the optimal neighborhood radius parameter value to train the computer vision system to cluster objects in objects or videos.
See further Lee below for mapping and motivation to combine with the claims of ‘903.
Claim 30
The system of claim 21, wherein identifying the optimal neighborhood radius parameter value comprises:
identifying a first middle left value and a first middle right value of a neighborhood radius parameter by dividing a first range of neighborhood radius parameter values between the first upper bound value and the first lower bound value into three sets of neighborhood radius parameter values.
Claim 2
The system of claim 1, wherein the optimal neighborhood radius parameter value is identified by performing modified ternary search operations comprising:
dividing a first range of neighborhood radius parameter values between the first upper bound value and the first lower bound value into three equidistant sets of neighborhood radius parameter values, the three equidistant sets corresponding to a left region, a middle region, and a right region, with a first middle boundary between the left region and the middle region…
Claim 31
The system of claim 30, wherein the first middle left value and the first middle right value mark corresponding middle boundaries between adjacent equidistant sets of neighborhood radius parameter values.
Claim 2
…dividing a first range of neighborhood radius parameter values between the first upper bound value and the first lower bound value into three equidistant sets of neighborhood radius parameter values, the three equidistant sets corresponding to a left region, a middle region, and a right region, with a first middle boundary between the left region and the middle region…
Claim 33
The method of claim 32, wherein receiving the dataset comprises:
receiving, at a density-based clustering system, the dataset, wherein the density-based clustering system utilizes a density-based clustering algorithm, a modified ternary search algorithm, and a convergence algorithm.
Claim 8
…identifying, using a modified ternary search algorithm, an optimal neighborhood radius parameter value, based on the first upper bound value and the first lower bound value of the neighborhood radius parameter…
…identifying the optimal neighborhood radius parameter value, by using a convergence algorithm on neighborhood radius parameter values within the equidistant set of neighborhood radius parameter values that is identified in the last iteration…
Claim 34
The method of claim 33, wherein the density-based clustering algorithm includes a Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm or an Ordering Points To Identify Clustering Structure ("OPTICS") algorithm.
See [0224] of Peng for mapping and see below for motivation to combine with the claims of ‘903.
Claim 35
The method of claim 33, wherein the modified ternary search algorithm leverages near-unimodality as a variant of ternary search algorithm.
Claim 1
…identifying, using a modified ternary search algorithm based on near-unimodality of the neighborhood radius parameter as a variant of ternary search algorithm…
Claim 36
The method of claim 33, wherein the convergence algorithm is used on neighborhood radius parameter values within an equidistant set of neighborhood radius parameter values.
Claim 8
…identifying the optimal neighborhood radius parameter value, by using a convergence algorithm on neighborhood radius parameter values within the equidistant set of neighborhood radius parameter values that is identified in the last iteration…
Claim 37
The method of claim 32, further comprising:
generating a first sampled dataset by sampling the dataset at a first sampling rate between 10% and 90%,
wherein the optimal neighborhood radius parameter value is a maximum neighborhood radius parameter value that is calculated based on the first sampled dataset.
Claim 10
The computer-implemented method of claim 8, wherein the first sampling rate is selected to be a value between 10% and 90%.
Claim 11
The computer-implemented method of claim 8, wherein the optimal neighborhood radius parameter value is a maximum calculated neighborhood radius parameter value that is calculated based on the first sampled dataset.
Claim 38
The method of claim 37, further comprising:
generating a second sampled dataset by sampling the dataset at a second sampling rate between 10% and 50%,
wherein the optimal neighborhood radius parameter value is a maximum neighborhood radius parameter value that is calculated based on a first initial upper bound value and a first initial lower bound value selected for the second sampled dataset.
Claim 12
…The computer-implemented method of claim 8, wherein the first upper bound value is selected by:
generating a second sampled dataset, by sampling the dataset at a second sampling rate, wherein the second sampling rate is selected to be a value between 10% and 50%; and
identifying, using the modified ternary search algorithm, the first upper bound value, based on a first initial upper bound value and a first initial lower bound value that are selected for the second sampled dataset…
Claim 39
The method of claim 38, wherein the at least one of the first upper bound value or the first lower bound value is selected for the second sampled dataset.
Claim 12
…identifying, using the modified ternary search algorithm, the first upper bound value, based on a first initial upper bound value and a first initial lower bound value that are selected for the second sampled dataset…
Although the conflicting claims are not identical, they are not patentably distinct from each other because they are substantially similar in scope and they use the similar limitations to produce the same end result of large-scale density-based clustering.
It would have been obvious to a person with ordinary skills in the art at the time of the invention was made to modify the elements of claims 1-20 of ‘903 with any combination of the cited references below to arrive at the pending claims of the instant application for the purpose of dynamically and efficiently adapting clustering parameters to different or evolving datasets with a high degree of accuracy of predictable results based on selection of the optimal neighborhood radius parameters. Further, it would have been obvious to a person with ordinary skills in the art at the time of the invention was made to modify or to omit the additional elements of claims 1-20 of ‘903 to arrive at the pending claims of the instant application because the person would have realized that the remaining element would perform the same functions as before. “Omission of element and its function in combination is obvious expedient if the remaining elements perform same functions as before.” See In re Karlson (CCPA) 136 USPQ 184, decide Jan 16, 1963, Appl. No. 6857, U.S. Court of Customs and Patent Appeals.
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 21-31, and 40 are rejected under 35 U.S.C. 101 because the claimed invention is directed to nonstatutory subject matters.
Regarding claims 21, and 40, each claim recites a system or a device comprising components such as a processing system and memory coupled to the processing system. However, these claimed component can be interpreted by a person of ordinary skills in the art as software modules that carry out the claimed functions (e.g., a virtual machine and/or virtual memory). Furthermore, in accordance with Applicant’s disclosure ([0062] of instant specification), Applicant states the recited system and/or each of the components can be “an entirely software implementation” consisting of data structures and computer programs, which impart functionality when employed as a computer component. As such, the claims each is not limited to statutory subject matter and is therefore non-statutory.
Applicant is suggested to include at least one hardware component (e.g. hardware processor and/or hardware memory) to overcome the issue raised.
Claims 22-31 fail to resolve the deficiencies of claim 21 since they only further limit the scope of claim 21. Hence, claims 22-31 are also rejected under 35 U.S.C. 101.
The claimed invention in claims 21-40 are directed to a judicial exception (i.e., an abstract idea) without significantly more (Applicant is noted that even if claims 21-31 and 40 are amended to pass step 1 of the “abstract idea” analysis, such claims are still ineligible under steps 2A and 2B of the analysis).
a. Claims 21, 32, and 40 recite, in part, elements that are directed to an abstract idea (“Courts have examined claims that required the use of a computer and still found that the underlying, patent-ineligible invention could be performed via pen and paper or in a person’s mind.” Versata Dev. Group v. SAP Am., Inc., 793 F.3d 1306, 1335, 115 USPQ2d 1681, 1702 (Fed. Cir. 2015)). Each claim recites the steps of receiving a dataset, selecting upper/lower bounds, identifying an optimal neighborhood radius parameter value, and clustering data based on the output. These steps describe mathematical concepts (e.g., performing algorithms to identify optimal numerical parameters) and mental processes (e.g., evaluating data constraints and grouping data). Thus, the claims are directed to an abstract idea per step 2A – prong 1 of the abstract idea analysis.
Per step 2A – prong 2, the claims do not include additional elements that integrate the abstract idea into a practical application. The claims merely apply the mathematical clustering operations to generic data, i.e., embedding of at least one of words, objects in an image, or objects in a video. Applying an abstract idea to a particular technological environment (e.g., data analysis) is a generic field-of-use restriction. The outputs are merely the results of the calculation, and reciting a generic processing system and memory to execute the instructions does not improve the functioning of the computer itself, nor does it effect a transformation of a particular article to a different state of thing.
Per step 2B, the claims do not recite an inventive concept that amounts to significantly more than the abstract idea itself. The additional elements recited in the claims, both individually and as an ordered combination, amount to no more than a generic computer (i.e., processing system, memory) performing generic computer functions of receiving, selecting, identifying, outputting. Executing mathematical algorithms on generic hardware is well-understood, routing, and conventional (WURC) activity in the art.
b. Claims 22-27, 30-31, and 33-39 add specific algorithmic techniques, equations, and data sampling parameters to the base claims.
Per step 2A – prong 2: claims 22-23, and 30-31 introduce concepts like selecting bounds, dividing ranges into three sets, and marking equidistant boundaries. Claims 24-25, and 33-36 specify algorithms such as a modified ternary search, DBSCAN, OPTICS, and hear-unimodality. Claims 26-27, and 37-39 recite generating sampled datasets at specific rates (e.g., 10%, 50%, 90%). These additional limitations merely narrow the scope of the abstract idea by providing more specific mathematical formulas and data-gathering techniques. Narrowing an abstract idea does not integrate it into a practical application. The claims still fail to improve a technological process or the computer itself.
Per step 2B, taken individually and as an ordered combination, these claims lack an inventive concept. Utilizing specific known algorithms such as DBSCAN, OPTICS, ternary searches, and statistical sampling such as 10% or 90% are WURC methods for data scientists organizing datasets. The claims merely instruct the practitioners to perform conventional mathematical operations of a generic processing system.
c. Claims 28-29 recite using the identified optimal neighborhood radius parameter or clustered datasets to train downstream models.
Per step 2A – prong 2: claim 28 recites using the output to training the language model to cluster natural language words. Claim 29 recites using the output to train a computer vision system to cluster objects in objects or videos. While these claims recite a practical end-use, they do not integrate the abstract idea into a practical application under the MPEP framework. Merely linking an abstract mathematical clustering process to a particular technological environment (e.g., training a language model or computer vision system) constitutes extra-solution activity and a field-of-use restriction. The claims do not detail how the computer’s functionality is fundamentally improved beyond merely providing it with more accurately calculated data.
Per step 2B, the additional elements of training a language model or a computer vision system are recited at a high level of generality. Utilizing mathematically clustered data to train machine learning models is a WURC activity in the field of AI/ML. The claims lack specific features demonstrating a non-conventional combination of elements that amounts to significantly more than the abstract idea of clustering data itself.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 21-23, 28, 32, and 40 are rejected under 35 U.S.C. 103 as being unpatentable over Peng et al. (Pub. No. US 2019/0065576, published on February 28, 2019; hereinafter Peng) in view of Das et al. (Pub. No. US 2025/0005571, at least effectively filed on May 22, 2024; hereinafter Das).
Regarding claims 21, 32, and 40, Peng clearly shows and discloses a method (Abstract); a system comprising: a processing system; and a memory comprising computer executable instructions that, when executed, perform operations of the method; and a device comprising: a processing system; and a memory comprising computer executable instructions that, when executed, perform operations of the method (Figure 2) comprising:
receiving a dataset including embeddings of at least one of words, objects in an image, or objects in a video (Turning to the embodiment illustrated in FIG. 1B, question and answer pairs are exported from online question and answer forums 500 and stored in system memory 110 in arrays of pairs, or arrays having multiple lines of pairs. Each line in the array comprises two strings of characters in two fields: a question string of characters in a question field and an answer string of characters in an answer field, [0059]. Turning to FIGS. 1A and 1B, the system comprises a vectorizer 140 that vectorizes extracted questions into multidimensional vectors 530. One example of a vectorizer is Word2vec™, which produces word embeddings from labeled question, [0104]);
selecting, for the dataset, at least one of a first upper bound value or a first lower bound value (To determine the value empirically, 96 questions are randomly sampled from the community QA data directly and query the Solr system for questions that share keywords with them. The average distance between the 1,591 pairs of different meanings is 0.32, which is significantly lower than the 0.70 average in the sample of extracted data. Meanwhile, the average distance between the 423 pairs of very similar meaning is 0.12, [0231]. DBSCAN (Ester et al., 1996) is used to cluster the questions. DBSCAN is one of the density-based clustering algorithms, which assigns closely packed data points to clusters and mark remote points as outliers, [0224]. It is clear that distance thresholds are evaluated by finding a lower bound average distance of 0.12 for pairs of similar meanings and an upper bound average distance of 0.32 for pairs of different meaning);
identifying an optimal neighborhood radius parameter value based on the at least one of the first upper bound value or the first lower bound value of a neighborhood radius parameter (∈ should be set to a value such that any two questions are similar in meaning if and only if the distance between the two is lower than ∈. If ∈ is set too large, every data point would be assigned to a single cluster and a lot of noise will result; if ∈ is set too small, few data points will be assigned to clusters and the system will have few data to work with, [0230]. The ∈ is chosen to be 0.2, which is approximately the average of 0.12 and 0.32, [0231]. It is clear that the optimal neighborhood radius threshold is identified by averaging the upper and lower boundary distances);
providing as output at least one of:
the optimal neighborhood radius parameter value (The ∈ is chosen to be 0.2, which is approximately the average of 0.12 and 0.32, [0231]-[0235]. It is clear that this determined ∈ parameter value of 0.2 is outputted to the DBSCAN algorithm to successfully assign the closely packed data points to clusters); or
an optimal number of clusters within the dataset corresponding to the optimal neighborhood radius parameter value; and
based on the output, clustering the at least one of words, objects in the image, or objects in the video (If the distance between two sentence vectors are within a threshold value (“ϵ”), then the two sentence vectors belong to the same cluster, [0118]-[0119]. Returning to the sentence vectors of the labelled question sentences, these sentence vectors are assigned into clusters by the clusterer 150 in the multidimensional vector space if each and every sentence vector in a cluster are separated by each and every other sentence vector by a distance or cosine similarity value less than the threshold value ϵ calculated above using the random sample question set. To do so, the clusterer 150 starts with a starting sentence vector and identifies all sentence vectors with relative distance or cosine similarity value less than the threshold value ϵ, [0120]).
Das then alternatively or additionally discloses:
providing as output at least one of:
the optimal neighborhood radius parameter value; or
an optimal number of clusters within the dataset corresponding to the optimal neighborhood radius parameter value ((ii) determining, with the at least one processor, a number of clusters into which the plurality of node embeddings is to be clustered; (iii)…a node associated with that node embedding is within k-hops in the graph of each other node associated with each other node embedding in that cluster, [0026]); and
based on the output, clustering the at least one of words, objects in the image, or objects in the video ((iii) clustering, with the at least one processor, based on distances between pairs of node embeddings in the plurality of node embeddings, the plurality of node embeddings into the number of clusters until, for each node embedding in each cluster, [0026]).
It would have been obvious to an ordinary person skilled in the art at the time of the invention that was effectively filed to incorporate the teachings of Das with the teachings of Peng for the purpose of dynamically and efficiently adapting clustering parameters to different or evolving datasets with a high degree of accuracy of predictable results based on selection of the optimal neighborhood radius parameters.
Regarding claim 22, Peng further discloses selecting the at least one of the first upper bound value or the first lower bound value comprises selecting the first upper bound value and the first lower bound value (To determine the value empirically, 96 questions are randomly sampled from the community QA data directly and query the Solr system for questions that share keywords with them. The average distance between the 1,591 pairs of different meanings is 0.32, which is significantly lower than the 0.70 average in the sample of extracted data. Meanwhile, the average distance between the 423 pairs of very similar meaning is 0.12, [0231]. It is clear that distance thresholds are evaluated by finding a lower bound average distance of 0.12 for pairs of similar meanings and an upper bound average distance of 0.32 for pairs of different meaning).
Regarding claim 23, Peng further discloses the at least one of the first upper bound value or the first lower bound value is associated with the neighborhood radius parameter of the density-based clustering algorithm (To determine the value empirically, 96 questions are randomly sampled from the community QA data directly and query the Solr system for questions that share keywords with them. The average distance between the 1,591 pairs of different meanings is 0.32, which is significantly lower than the 0.70 average in the sample of extracted data. Meanwhile, the average distance between the 423 pairs of very similar meaning is 0.12, [0231]. DBSCAN (Ester et al., 1996) is used to cluster the questions. DBSCAN is one of the density-based clustering algorithms, which assigns closely packed data points to clusters and mark remote points as outliers, [0224]. It is clear that distance thresholds are evaluated by finding a lower bound average distance of 0.12 for pairs of similar meanings and an upper bound average distance of 0.32 for pairs of different meaning).
Regarding claim 28, Peng further discloses providing, to a language model, the at least one of the optimal neighborhood radius parameter value or the optimal number of clusters within the dataset corresponding to the optimal neighborhood radius parameter value to train the language model to cluster natural language words (These models are shallow, two-layer neural networks that are trained to reconstruct linguistic contexts of words. Word2vec takes a text corpus as input and produces the word vectors as output. It first constructs a vocabulary from the training text data and then learns vector representation of words. The resulting word vector file can be used as features in various natural language processing and machine learning applications, [0104]).
Claims 24, 30-31, and 33-34 are rejected under 35 U.S.C. 103 as being unpatentable over Peng in view of Das and further in view of Leipold (Pub. No. US 2025/0139476, filed on October 30, 2023).
Regarding claim 24, Leipold then discloses identifying the optimal neighborhood radius parameter value comprises using a modified ternary search algorithm to perform the identifying (the search algorithm may be a linear search algorithm, a sentinel linear search algorithm, a binary search algorithm, a meta binary search algorithm, a ternary search algorithm, a jump search algorithm, an interpolation search algorithm, an exponential search algorithm, a Fibonacci search algorithm, a breadth first search algorithm, a depth first search algorithm, Dijkstra's algorithm, the Floyd Warshall algorithm, Prim's algorithm, Kruskal's algorithm, and/or any other algorithm used to determine neighboring nodes, [0034]-[0035]).
It would have been obvious to an ordinary person skilled in the art at the time of the invention that was effectively filed to incorporate the teachings of Leipold with the teachings of Peng, as modified by Das, for the purpose of enhancing similarity matchings of datasets based on a threshold number of neighbor nodes associated with an initial dataset that have higher correlations between pairs of nodes.
Regarding claim 30, Leipold further discloses identifying the optimal neighborhood radius parameter value comprises: identifying a first middle left value and a first middle right value of a neighborhood radius parameter by dividing a first range of neighborhood radius parameter values between the first upper bound value and the first lower bound value into three sets of neighborhood radius parameter values (the search algorithm may be a ternary search algorithm and/or any other algorithm used to determine neighboring nodes, [0034]-[0035]. It is clear that dividing the search range into three sets is the inherent mathematical definition of ternary search algorithm. Optimizing a parameter within a bounded search space via ternary search fundamentally requires calculating two midpoints which necessitates dividing the continuous search space into exactly three distinct segments).
Regarding claim 31, Leipod further discloses the first middle left value and the first middle right value mark corresponding middle boundaries between adjacent equidistant sets of neighborhood radius parameter values (the search algorithm may be a ternary search algorithm and/or any other algorithm used to determine neighboring nodes, [0034]-[0035]. Because the total search space is divided by a constant denominator of 3 to calculate the respective midpoints, the resulting three sets are mathematically guaranteed to be equal in size or equidistant and adjacent to one another within the search space).
Regarding claim 33, Peng and Leipold further discloses receiving the dataset comprises: receiving, at a density-based clustering system, the dataset, wherein the density-based clustering system utilizes a density-based clustering algorithm (DBSCAN (Ester et al., 1996) is used to cluster the questions. DBSCAN is one of the density-based clustering algorithms, which assigns closely packed data points to clusters and mark remote points as outliers, [0224] of Peng), a modified ternary search algorithm, and a convergence algorithm (the search algorithm may be a linear search algorithm, a sentinel linear search algorithm, a binary search algorithm, a meta binary search algorithm, a ternary search algorithm, a jump search algorithm, an interpolation search algorithm, an exponential search algorithm, a Fibonacci search algorithm, a breadth first search algorithm, a depth first search algorithm, Dijkstra's algorithm, the Floyd Warshall algorithm, Prim's algorithm, Kruskal's algorithm, and/or any other algorithm used to determine neighboring nodes, [0034]-[0035] of Leipold).
Regarding claim 34, Peng further discloses the density-based clustering algorithm includes a Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm (DBSCAN (Ester et al., 1996) is used to cluster the questions. DBSCAN is one of the density-based clustering algorithms, which assigns closely packed data points to clusters and mark remote points as outliers, [0224]) or an Ordering Points To Identify Clustering Structure ("OPTICS") algorithm.
Claims 26-27, and 37-39 are rejected under 35 U.S.C. 103 as being unpatentable over Peng in view of Das and further in view of Shinagawa et al. (Pub. No. US 2024/0005503, published on January 4, 2024; hereinafter Shinagawa).
Regarding claim 26, Shinagawa then discloses selecting the first upper bound value comprises:
generated a sampled dataset by sampling the dataset at a sampling rate between 10% and 50% (the pixels or voxels are sampled in a sparse and/or random manner. “Sparse” is to be understood as, when having regard to the total number of pixels or voxels making up the reference and/or target medical image, only few pixels or voxels are being used in sparse sampling. In particular, “sparse” is to say that less than 50% or less than 20% or even less than 10% of the total number of pixels or voxels of the reference and/or target medical image are sampled, [0040]); and
identifying the first upper bound value based on a first initial upper bound value and a first initial lower bound value that are selected for the sampled dataset (a neighborhood region 600 within the target medical image 400 is selected. The neighborhood region 600 includes the coordinates x′, y′, z′. The neighborhood region 600 may, for example, include less than 50%, less than 10% or even less than 3% of the total number of pixels in the target medical image 400. Then, a descriptor 601 is generated based on pixels sampled from the neighborhood region 600. Any of the above-mentioned techniques such as sparse sampling and/or sampling with a sampling rate per unit length which decreases with the distance from the initial corresponding location x′, y′, z′ may be used, [0106]).
It would have been obvious to an ordinary person skilled in the art at the time of the invention that was effectively filed to incorporate the teachings of Shinagawa with the teachings of Peng, as modified by Das, for the purpose of predictably reducing processing time and memory consumption while still generating a statistically representative upper and lower bound for the dataset based on sparse sampling of the vectors associated with the dataset.
Regarding claim 27, Shinagawa further discloses the first initial upper bound value is one of:
a value of 1, for cosine-based neighborhood radius parameter values;
a maximum neighborhood radius parameter value, for other bounded neighborhood radius parameter values (the candidate locations are chosen such that the distance between each candidate location and the location x, y, z (taken from the reference image 500) does not exceed a predefined threshold value, [0110]); or
a value of a sum of difference values between a maximum function and a minimum function of a distance variable, for unbounded neighborhood radius parameter values.
Regarding claim 37, Shinagawa then discloses: generating a first sampled dataset by sampling the dataset at a first sampling rate between 10% and 90%, wherein the optimal neighborhood radius parameter value is a maximum neighborhood radius parameter value that is calculated based on the first sampled dataset (the pixels or voxels are sampled in a sparse and/or random manner. “Sparse” is to be understood as, when having regard to the total number of pixels or voxels making up the reference and/or target medical image, only few pixels or voxels are being used in sparse sampling. In particular, “sparse” is to say that less than 50% or less than 20% or even less than 10% of the total number of pixels or voxels of the reference and/or target medical image are sampled, [0040]).
Regarding claim 38, Shinagawa further discloses generating a second sampled dataset by sampling the dataset at a second sampling rate between 10% and 50% (the pixels or voxels are sampled in a sparse and/or random manner. “Sparse” is to be understood as, when having regard to the total number of pixels or voxels making up the reference and/or target medical image, only few pixels or voxels are being used in sparse sampling. In particular, “sparse” is to say that less than 50% or less than 20% or even less than 10% of the total number of pixels or voxels of the reference and/or target medical image are sampled, [0040]); wherein the optimal neighborhood radius parameter value is a maximum neighborhood radius parameter value that is calculated based on a first initial upper bound value and a first initial lower bound value selected for the second sampled dataset (a neighborhood region 600 within the target medical image 400 is selected. The neighborhood region 600 includes the coordinates x′, y′, z′. The neighborhood region 600 may, for example, include less than 50%, less than 10% or even less than 3% of the total number of pixels in the target medical image 400. Then, a descriptor 601 is generated based on pixels sampled from the neighborhood region 600. Any of the above-mentioned techniques such as sparse sampling and/or sampling with a sampling rate per unit length which decreases with the distance from the initial corresponding location x′, y′, z′ may be used, [0106]).
Regarding claim 39, Shinagawa further discloses the at least one of the first upper bound value or the first lower bound value is selected for the second sampled dataset (the at least one marker is associated with a region in the target and/or reference image, all pixels in said region having a value above, below or between a value defined prior to step d), [0058]).
Claim 29 is rejected under 35 U.S.C. 103 as being unpatentable over Peng in view of Das and further in view of Lee et al. (Pub. No. US 2024/0185612, filed on December 6, 2022; hereinafter Lee).
Regarding claim 29, Lee then discloses providing, to a computer vision system, the at least one of the optimal neighborhood radius parameter value or the optimal number of clusters within the dataset corresponding to the optimal neighborhood radius parameter value to train the computer vision system to cluster objects in objects or videos (clustering embeddings of sufficient similarity. The clustering step is a preprocessing step that will allow pairwise connections to be very quickly drawn within independent clusters. This can be done incrementally to improve computational efficiency. According to various embodiments of the present disclosure, the method may include using the DBSCAN algorithm to compute the clusters. For example, clustering may use a maximum distance (eps) of 0.99 and a minimum of 2 samples out of a test set of ˜13,000 embeddings, which can result in an excellent adjusted mutual information score (AMI) (e.g., 0.793) with an average cluster size of 15.24 for 842 clusters. Optimal parameters will change for different models, [0075]. The sensor data may include the sensor data generated by one or more forward-facing cameras (e.g., a center or near-center mounted camera(s)), such as a wide-view camera, a surround camera, a stereo camera, and/or a long-range or mid-range camera. This sensor data may be useful for computer vision and/or perception when navigating, [0087]).
It would have been obvious to an ordinary person skilled in the art at the time of the invention that was effectively filed to incorporate the teachings of Lee with the teachings of Peng, as modified by Das, for the purpose of dynamically detecting and tracking objects in an autonomous computer vision environment based on similarity scores of the image embeddings of detected objects.
Allowable Subject Matter
Claims 25, and 35-36 are allowable over the prior art.
Relevant Prior Art
The following references are deemed relevant to the claims:
Wu et al. (Pub. No. US 2022/0083885) teaches a method for recognizing multi-dimensional anomalous urban traffic events based on a ternary Gaussian mixture model includes: reading a data sample of urban road traffic events; randomly dividing the data sample into a first subsample and a second subsample; performing modeling based on the first subsample by using the ternary Gaussian mixture model to obtain a second ternary Gaussian mixture model to calculate a distribution probability p of any sample point; clustering the second subsample, recognizing an outlier in the second subsample, and labeling the outlier and a normal point to obtain a labeled subsample; calculating the labeled subsample to obtain the distribution probability p corresponding to each sample point in the labeled subsample.
Buyukkayhan et al. (Pub. No. US 9998484) teaches an optimized DBSCAN clustering algorithm wherein indices are maintained on several dimensions or features that permit construction of a coarse-grained, expanded neighborhood for each software module. The optimized DBSCAN clustering algorithm only considers points in the expanded neighborhood and computes distances from a software module to those points to determine the neighborhood. The optimized DBSCAN clustering algorithm can scale for use with hundreds of thousands of software modules.
Shimakura (Pub. No. US 2010/0228754) teaches simultaneously search a plurality of combinations of elements in a search, and programmably rearrange the search elements and the search conditions (match, mismatch, range, and out-of-range). This allows to execute various search processes in the same search engine, and therefore improvement of the performance and functionality of the system is expected. In the match search and mismatch search, a match/mismatch is determined by a bit-by-bit comparison. In the range search and out-of-range search, it is determined for each region (4 bits) or each connected region formed by connecting a plurality of neighboring regions whether a corresponding region falls between a preset upper limit and lower limit.
Contact Information
Any inquiry concerning this communication or earlier communications from the Examiner should be directed to Son Hoang whose telephone number is (571) 270-1752. The Examiner can normally be reached on Monday – Friday (7:00 AM – 4:00 PM).
If attempts to reach the Examiner by telephone are unsuccessful, the Examiner’s supervisor, Sherief Badawi can be reached on (571) 272-9782. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/SON T HOANG/
Primary Examiner, Art Unit 2169 July 31, 2026
Prosecution Insights