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 .
DETAILED ACTION
1. This Office Action is in response to the preliminary amendment filed on 06/21/2024.
Claims 1-7 and 9-16 have been amended.
Claim 8 has been canceled.
Claims 18-21 have been added.
Claims 1-7 and 9-21 are pending.
Information Disclosure Statement
2. The information disclosure statement (IDS) filed on 06/21/2024 complies with the provisions of M.P.E.P. 609. The examiner has considered it.
Claim Rejections - 35 USC § 101
3. 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.
4. Claims 1-7 and 9-21 are rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter.
At Step 1:
Independent claims 1 and 9 and are directed to a “method” and a “program product” and thus directed to a statutory category
At Step 2A, Prong One:
The claim recites the following limitations directed to an abstract idea:
• " obtaining a dataset comprising a plurality of datapoints corresponding to a
plurality of values of one or more dependent variables for a plurality of first features
over a time period" as drafted this recites certain methods of organizing human activity (Data Collection / Gathering: The Federal Circuit (in cases like Electric Power Group v. Alstom) has established that merely collecting, selecting, and storing information is a well-known, routine human activity).
• " determining, using machine learning and explainability, in the absence of
determining whether or not the plurality of datapoints has a drift, whether or not there
has been a change in respective one or more characteristics of a subset of the plurality
of first features in the plurality of datapoints having a largest contribution to a
variability of the datapoints in the plurality of datapoints based on a threshold from a
first time period to a second time period" as drafted this recites mathematical concepts (While framing the invention as an alternative approach (avoiding an explicit holistic drift check), it still focuses on calculating statistical variations, which can be categorized as a mathematical concept.).
• " initiating application of a drift policy on the plurality of datapoints based on
a result of the determination of whether or not there has been a change" as drafted this recites mathematical concepts and mental processes (Tracking datapoints, computing whether they deviate (drift), and triggering a secondary rule is an operational workflow that can be mapped out mathematically or performed conceptually by a human reviewer).
At Step 2A, Prong Two:
• The claim recites no additional elements. At most one might consider that a " a communications system " or “executed on at least one processor” as claimed might be considered to represent a computer-implemented system and method consistent with Fig. 1 even though the claim does not recite any computer. At most this would be a high-level recitation of a generic computer components and represents mere instructions to apply the abstract idea on a computer as in MPEP 2106.05(f), which does not provide integration into a practical application.
• Viewing the additional limitations together and the claim as a whole, nothing provides integration into a practical application.
At Step 2B:
• The conclusions for the mere implementation using a computer are carried over and does not provide significantly more.
• Looking at the claim as a whole does not change this conclusion and the claim is ineligible.
Dependent Claims 2-14
The limitations as recited in dependent claims 2, 11 and 17 recite, “wherein the obtained dataset is a second dataset, the plurality of datapoints is a second plurality of datapoints, the plurality of values is a second plurality of values, the subset is a second subset, the respective one or more characteristics are respective one or more second characteristics and the time period is the second time period, and wherein the method further comprises: obtaining a first dataset comprising a first plurality of datapoints corresponding to a first plurality of values of the one or more dependent variables for the plurality of first features over a first time period” which further describes certain methods of organizing human activity (USPTO guidelines and Federal Circuit precedent consistently treat generic claims to "obtaining," "gathering," "receiving," or "storing" data as falling under the judicial exception of an abstract idea (often categorized under methods of organizing human activity or mental processes)), and “determining, using machine learning and explainability: i) a first subset of the plurality of first features having a largest contribution to a variability of the datapoints in the first set of datapoints based on a threshold” which further describes mathematical concepts (Calculating data variance, measuring feature contribution (e.g., feature importance or Shapley values), and applying a numerical threshold are mathematical calculations and mental processes that can be performed with pen and paper or generic data manipulation), and “determining, using machine learning and explainability: ii) respective one or more first characteristics of the first subset of the plurality of first features, and wherein the determining of whether or not there has been a change is based on comparing the determined first subset and the respective one or more first characteristics with the second subset and the respective one or more second characteristics which further describes mathematical concepts / mental process (Under recent precedent (such as Recentive Analytics, Inc. v. Fox Corp.) and USPTO guidelines, invoking generic machine learning, feature selection, or data analysis without specifying a concrete improvement to the underlying computer or machine learning model is viewed as reciting a mental process or mathematical algorithm) ,in step 2A prong one.
The limitations as recited in dependent claims 3, 12 and 18 recite, “obtaining a predictive model of the one or more dependent variables, based on the obtained first dataset and the plurality of first features, and wherein the determining of the first subset and the respective one or more first characteristics, and the determining of the second subset and the respective one or more second characteristics are further based on the obtained predictive model” which further describes mathematical concepts (Building or "obtaining" a predictive model from datasets and features relies on underlying statistical correlations, equations, and mathematical relationships) and/or a mental process (Humans have long mentally or manually analyzed variables and datasets to make predictions; framing this as a software step without specifying how the computer improves its own internal functioning or operations can trigger an abstract idea rejection) , in step 2A prong one.
The limitations as recited in dependent claims 4, 13 and 19 recite, “wherein the initiating of the application of the drift policy comprises: flagging any datapoints wherein the node has determined there has been change and determining (308) whether or not there is drift using only the flagged datapoints” which further describes a mental process (Courts have frequently found that generic functions like collecting, analyzing, comparing, or flagging data are abstract ideas if they represent a method of organizing human activity, a mental process that can be performed in the human mind, or a generalized mathematical/logical comparison.) , in step 2A prong one.
The limitations as recited in dependent claims 5, 14 and 20 recite, “wherein drift is determined to have occurred in the flagged datapoints and wherein the initiating of the application of the drift policy further comprises at least one of: retraining the obtained predictive model based on the determined drift” which further describes a mental process (Broadly observing that data has changed and feeding new data into an algorithm is frequently analogized by patent examiners to a basic mental process or a generic mathematical/statistical routine) and “replacing the obtained predictive model by another predictive model based on the determined drift” which further describes a mental process (Under USPTO and Federal Circuit case law, the general idea of collecting data, detecting a change, and updating or selecting a model can be lumped into mental processes, mathematical calculations, or generic data manipulation. Simply applying a generic machine learning concept to new data or a new scenario without an unconventional technical solution often triggers a subject-matter eligibility rejection) and “reverting to an earlier version of the obtained predictive model based on the determined drift” which further describes a mental process (Courts and the USPTO often view the bare concept of monitoring data change ("drift") and deciding to roll back or switch software/model versions as a routine business or cognitive activity (akin to keeping or swapping a ledger or routine log) implemented on generic computer hardware), in step 2A prong one.
The limitations as recited in dependent claims 6 and 21 recite, “wherein the obtaining of the dataset comprising the plurality of datapoints is performed after an actuation executed based on the obtained predictive model” which further describes the concept is mere gathered data under prong 2 (insignificant extra solution activity— MPEP 2106.06g) and WURC under 2b (using gather data - MPEP 2106.05d).
The limitations as recited in dependent claims 7 and 16 recite, “wherein drift is determined to have occurred in the flagged datapoints and wherein the initiating of the application of the drift policy further comprises: sending an indication to another node operating in the communications system to indicate the detected drift” which further describes a mental process (Under Alice Step 1 and Step 2, simply stating that a generic "node" or "communications system" performs the sending is often viewed as generic computer automation of an abstract mental or organizational task) , in step 2A prong one.
Please see claim 1 for the limitations as recited in dependent claims 10.
The limitations as recited in dependent claim 15 recites, “wherein the obtaining of the dataset comprising the plurality of datapoints is configured to be performed after an actuation configured to be executed based on the predictive model configured to be obtained” which further describes the concept is mere gathered data under prong 2 (insignificant extra solution activity— MPEP 2106.06g) and WURC under 2b (using gather data - MPEP 2106.05d).
Claim Rejections - 35 USC § 101
5. 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.
6. Claims 9 and 17-21 are rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter.
Claims 9 and 17-21 are rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter because claim 15 directs to “computer-readable storage medium” in which, when using the broadest reasonable interpretation would include non-statutory subject matter. That is, language such as “physical”, “tangible” and “storage” do not make an otherwise non-statutory computer-readable medium claim statutory, since a data signal per se is considered a physical and tangible medium that temporarily stores data (it’s a transitory storage medium, but it’s still a storage medium).
Examiner’s Note
7. Data drift (According to Google): “Data drift is a change in the statistical properties of a machine learning model's input data in a production environment compared to the data it was trained on. When real-world conditions, user behavior, or system inputs evolve, the data diverges from historical baselines, causing the model's accuracy to degrade over time. Core Types of Data Drift: (1) Covariate Shift (Feature Drift): The distribution of input features changes over time, while the underlying relationship between the inputs and the target output stays the same. (2) Prior Probability Shift (Label Shift): The distribution of the target variable changes in production. (3) Concept Drift: The statistical relationship between the input features and the target variable shifts, meaning the fundamental "ground truth" or logic has changed.”
A data set (According to Google): “A data set is an organized collection of data values—usually stored in tables, CSV files, JSON objects, or databases—used for analysis, statistics, or machine learning. Common Data Set Examples by Type: (1) Numerical (Quantitative) Data: Contains numbers used for math or calculations, such as home square footage, daily temperatures, or bank account balances. (2) Categorical (Qualitative) Data: Represents discrete groups or labels, such as customer eye color, product categories, or true/false survey responses. (3) Time-Series Data: Records values sequentially over time, such as daily stock prices, monthly retail sales, or hourly weather tracking. (4) Image / Unstructured Data: Consists of visual or text files, such as medical X-rays for disease detection, handwritten digit scans, or customer service chat logs. Popular Real-World Data Sets for Practice: (1) LEGO Sets Data: Tracks LEGO sets released from 1970 to 2022 across 18,459 records. (2) Bank Customer Churn: Contains account information and churn metrics for 10,000 bank customers. (3) World Economic Indicators: Combines economic development data from the World Bank and United Nations. (4) US Census Data: Offers nationwide demographic, age, income, and housing statistics.”
Machine learning and explainability (According to Google): “Machine learning is a subfield of artificial intelligence that lets computers learn patterns from data and make predictions or decisions without being explicitly programmed with hardcoded rules. Explainability (often called Explainable AI or XAI) is the set of methods and tools that help humans understand, evaluate, and trust how a machine learning model reaches a specific output or decision. Machine Learning Core Concepts: How it works: Instead of a programmer writing step-by-step instructions, the computer analyzes training data, recognizes underlying relationships, and adjusts its parameters to reduce errors. Main types: (1) Supervised Learning: Trains on labeled data (e.g., images marked as "cat" or "dog"). (2) Unsupervised Learning: Finds hidden patterns or groupings in unlabeled data. (3) Reinforcement Learning: Learns through trial, error, and rewards.”
Is the phrase, "determining a statistical distance based on the reference distribution and the current distribution", implying dataset? (According to Google): “Yes, the phrase implies the underlying presence of datasets or data samples, even if it is described in terms of abstract probability distributions. Why a Dataset is Implied: (1) Empirical Derivation: True mathematical probability distributions are often theoretical. In practice, a "reference distribution" (such as a training set snapshot or a baseline time window) and a "current distribution" (such as live production data or a recent sliding window) are built and estimated directly from datasets. (2) This specific phrasing ("reference distribution" vs. "current distribution") is a standard term in machine learning observability and data drift detection. You cannot compute a distance between a current state and a baseline without collecting or observing data points (samples) from both time periods or cohorts. (3) Samples vs. Distributions: While the math treats them as probability measures or random variables, the distributions themselves are empirical approximations derived from data records (tabular data, logs, feature values, or predictions).”
Is the phrase, "determining a local feature importance parameter for each feature associated with a prediction made by the machine learning model", implying datapoint? (According to Google): “Yes, the phrase directly implies a specific data point (or single instance) rather than the entire dataset. Why It Implies a Data Point: (1) "A prediction made by the machine learning model": A model generates a prediction for one specific instance, observation, or row of data (a data point) at a time. (2) "Local feature importance": In machine learning interpretability (such as SHAP values or LIME), local importance means explaining why the model made that specific prediction for an individual data point. By contrast, global feature importance looks at how features behave across the whole dataset. (3) "For each feature associated with a prediction": It evaluates how much each input variable contributed to that single, localized output value.”
Is executing something similar to actuation? (According to Google): “Yes, executing is highly similar to actuation, especially when looking at how systems process information to cause real-world effects. Both terms describe the final "doing" phase of a process, but they come from different fields.”
Lopatecki et al, US 20230186144, [Lopatecki: Title and Abstract (“Data drift impact in a machine learning model” ; “determining a drift impact score in a machine learning model”)] [Lopatecki: Abstract (“determining a statistical distance based on the reference distribution and the current distribution”)] [Lopatecki: Abstract and paragraph 29 (“determining a local feature importance parameter for each feature associated with a prediction made by the machine learning model” ; “moving time window”)] explainability [Lopatecki: Paragraphs 27 and 70 (“a metric called drift impact score that can be used by explainability tools” ; “Cohort explainability can explain why a model is not performing as well for a particular subset of its inputs.”)] [Lopatecki: Paragraphs 36 and 39 (“Setting of thresholds can be recommended based on common finance industry benchmarks or on days/hours of previous samples of PSI for that feature/prediction/actual. Industry benchmarks of 0.1-0.25 in finance can catch moves of around 10% between bins”, i.e., “0.25” = ‘a largest contribution to a variability of the datapoints’ and “the stability index is below the 0.15 limit that was set”, i.e., upper threshold)] [Lopatecki: Paragraph 38 (“If the PSI falls below a well-defined threshold, the change in the model input can be investigated and may indicate a model performance issue”, i.e., upper threshold)] [Lopatecki: Paragraph 57 (“pinpointing specific cohorts of predictions that have the biggest difference from their actuals”)] [Lopatecki: Paragraph 76 (“goal of output drift can be to detect large changes in the way the model is working relative to training”)].
Erlandson et al, US 20190147357, [Erlandson: Paragraph 22 (“When the operational data changes sufficiently relative to the original training data set, the predictive performance (aka “inference”) of the predictive learning model degrades because the operational data is from regions of the larger feature space that the predictive learning model never encountered through the training data set. This phenomenon is sometimes referred to as “learning model drift,” although in fact it is the operational data, not the predictive learning model, that is drifting.”)] [Erlandson: Paragraph 23 (“Because operational data that has drifted from the training data degrades the performance of the predictive learning model, and therefore erodes the business value of the predictive learning model in operation, it is desirable to detect the drift of the operational data as it occurs. Detecting the drift of the operational data may be useful, for example, to determine when a predictive learning model should be retrained on current operational data”)].
Claim Rejections - 35 USC § 102
8. 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 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.
9. The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
10. Claims 1-7 and 9-21 are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Lopatecki et al (US 20230186144).
Claim 1:
Lopatecki suggests a computer-implemented method, performed by a node
method being for handling drift in data [Lopatecki: Title and Abstract (“Data drift impact in a machine learning model” ; “determining a drift impact score in a machine learning model”)], the node [Lopatecki: Abstract (“determining a statistical distance based on the reference distribution and the current distribution”)] comprising a plurality of datapoints corresponding to a plurality of values of one or more dependent variables for a plurality of first features over a time period[[,]] [Lopatecki: Abstract and paragraph 29 (“determining a local feature importance parameter for each feature associated with a prediction made by the machine learning model” ; “moving time window”)]. Lopatecki suggests [[-]] determining [Lopatecki: Paragraphs 27 and 70 (“a metric called drift impact score that can be used by explainability tools” ; “Cohort explainability can explain why a model is not performing as well for a particular subset of its inputs.”)], in the absence of determining whether or not the plurality of datapoints has a drift [Lopatecki: Abstract (“determining a drift impact score in a machine learning model”)], whether or not there has been a change in respective one or more characteristics of a subset of the plurality of first features in the plurality of datapoints having a largest contribution to a variability of the datapoints in the plurality of datapoints based on a threshold from a first time period to a second time period[[,] [Lopatecki: Paragraphs 36 and 39 (“Setting of thresholds can be recommended based on common finance industry benchmarks or on days/hours of previous samples of PSI for that feature/prediction/actual. Industry benchmarks of 0.1-0.25 in finance can catch moves of around 10% between bins”, i.e., “0.25” = ‘a largest contribution to a variability of the datapoints’ and “the stability index is below the 0.15 limit that was set”, i.e., upper threshold)] [Lopatecki: Paragraph 38 (“If the PSI falls below a well-defined threshold, the change in the model input can be investigated and may indicate a model performance issue”, i.e., upper threshold)] [Lopatecki: Paragraph 57 (“pinpointing specific cohorts of predictions that have the biggest difference from their actuals”)]. Lopatecki suggests [[-]] initiating [Lopatecki: Paragraph 76 (“goal of output drift can be to detect large changes in the way the model is working relative to training”)].
Claim 2:
Lopatecki suggests wherein the obtained dataset is a second dataset, the plurality of datapoints is a second plurality of datapoints, the plurality of values is a second plurality of values, the subset is a second subset, the respective one or more characteristics are respective one or more second characteristics and the time period is the second time period, and wherein the method further comprises: [[-]] obtaining
the plurality of first features over a first time period[[,]] [Lopatecki: Abstract and paragraph 29 (“determining a local feature importance parameter for each feature associated with a prediction made by the machine learning model” ; “moving time window”)] [Lopatecki: Paragraphs 27 and 70 (“a metric called drift impact score that can be used by explainability tools” ; “Cohort explainability can explain why a model is not performing as well for a particular subset of its inputs.”)]. Lopatecki suggests [[-]] determining [Lopatecki: Paragraphs 36 and 39 (“Setting of thresholds can be recommended based on common finance industry benchmarks or on days/hours of previous samples of PSI for that feature/prediction/actual. Industry benchmarks of 0.1-0.25 in finance can catch moves of around 10% between bins”, i.e., “0.25” = ‘a largest contribution to a variability of the datapoints’ and “the stability index is below the 0.15 limit that was set”, i.e., upper threshold)] [Lopatecki: Paragraph 38 (“If the PSI falls below a well-defined threshold, the change in the model input can be investigated and may indicate a model performance issue”, i.e., upper threshold)] [Lopatecki: Paragraph 57 (“pinpointing specific cohorts of predictions that have the biggest difference from their actuals”)].
Claim 3:
Lopatecki suggests obtaining
determining (305) of the second subset and the respective one or more second characteristics are further based on the obtained predictive model [Lopatecki: Paragraphs 58 and 76 (“Using performance metrics with distributions changes can enable identification of slices of predictions that may be bringing down overall performance of the model. For example, as shown in FIG. 10, the overall model accuracy is 71%. Using statistical distance measures described previously, significant movement can be observed in 2 bins”)].
Claim 4:
Lopatecki suggests wherein the initiating
been change, and [[-]] determining (308) whether or not there is drift using only the flagged datapoints [Lopatecki: Abstract (“Tracking changes in data and understanding how these changes affect model performance”)] [Lopatecki: Paragraph 28 (“what changes in the machine learning model are to be observed and caught and it can be saved in an external or internal storage”)] [Lopatecki: Paragraph 30 (“a current distribution of the machine learning model can be obtained by monitoring changes to the various features (model inputs, model outputs, actuals, etc.)”)] [Lopatecki: Paragraph 35 (“any change in the distribution will add to the PSI total—whether the change is positive or negative”)].
Claim 5:
Lopatecki suggests wherein drift is determined to have occurred in the flagged datapoints and wherein the initiating
on the determined drift, and [[-]] reverting [Lopatecki: Paragraph 53 (“It can lead to model retraining if there are significant impacts to the model performance.”)] [Lopatecki: Paragraph 55 (“statistical distance checks can be set up on the prediction distributions for different model versions.”)].
Claim 6:
Lopatecki suggests wherein the obtaining [Lopatecki: Paragraph 53 (“It can lead to model retraining if there are significant impacts to the model performance.”)] [Lopatecki: Abstract and paragraph 6 (“obtaining a current distribution of the machine learning model; determining a statistical distance based on the reference distribution and the current distribution; determining a local feature importance parameter for each feature associated with a prediction made by the machine learning model”)].
Claim 7:
Lopatecki suggests wherein drift is determined to have occurred in the flagged datapoints and wherein the initiating
communications system [Lopatecki: Paragraph 55 (“for canary model deployment, statistical distance checks can be set up on the prediction distributions for different model versions. While AB testing two different models in production with each model receiving a certain amount of traffic or back testing a model on historical data, comparing the prediction distribution can give insight into how one model performs over another.”, i.e., “two different models” = ‘another node’)].
Claim 9:
Claim 9 is essentially the same as claim 1 except that it sets forth the claimed invention as a program product rather than a method and rejected under the same reasons as applied above.
Claim 10:
Claim 10 is essentially the same as claim 1 except that it sets forth the claimed invention as a node/system rather than a method and rejected under the same reasons as applied above.
Claim 11:
Claim 11 is essentially the same as claim 2 except that it sets forth the claimed invention as a node/system rather than a method and rejected under the same reasons as applied above.
Claim 12:
Claim 12 is essentially the same as claim 3 except that it sets forth the claimed invention as a node/system rather than a method and rejected under the same reasons as applied above.
Claim 13:
Claim 13 is essentially the same as claim 4 except that it sets forth the claimed invention as a node/system rather than a method and rejected under the same reasons as applied above.
Claim 14:
Claim 14 is essentially the same as claim 5 except that it sets forth the claimed invention as a node/system rather than a method and rejected under the same reasons as applied above.
Claim 15:
Claim 15 is essentially the same as claim 6 except that it sets forth the claimed invention as a node/system rather than a method and rejected under the same reasons as applied above.
Claim 16:
Claim 16 is essentially the same as claim 7 except that it sets forth the claimed invention as a node/system rather than a method and rejected under the same reasons as applied above.
Claim 17:
Claim 17 is essentially the same as claim 2 except that it sets forth the claimed invention as a program product rather than a method and rejected under the same reasons as applied above.
Claim 18:
Claim 18 is essentially the same as claim 3 except that it sets forth the claimed invention as a program product rather than a method and rejected under the same reasons as applied above.
Claim 19:
Claim 19 is essentially the same as claim 4 except that it sets forth the claimed invention as a program product rather than a method and rejected under the same reasons as applied above.
Claim 20:
Claim 20 is essentially the same as claim 5 except that it sets forth the claimed invention as a program product rather than a method and rejected under the same reasons as applied above.
Claim 21:
Claim 21 is essentially the same as claim 6 except that it sets forth the claimed invention as a program product rather than a method and rejected under the same reasons as applied above.
11. Any inquiry concerning this communication or earlier communications from the examiner should be directed to [Hung D. Le], whose telephone number is [571-270-1404]. The examiner can normally be communicated on [Monday to Friday: 9:00 A.M. to 5:00 P.M.].
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Apu Mofiz can be reached on [571-272-4080]. 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, contact [800-786-9199 (IN USA OR CANADA) or 571-272-1000].
Hung Le
10/14/2026
/HUNG D LE/Primary Examiner, Art Unit 2161
