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 .
This action is responsive to the Amendment filed on 8/20/2025.
Claims 1, 2, 4-6, 8-15 are pending in the case. Claim(s) 13-15 are new.
Response to Arguments
Applicant's arguments and amendments with regards to the 35 U.S.C. § 112(b) rejection of claim(s) 1, 2, 4-6, 8-12, have been fully considered and are persuasive. The 35 U.S.C. § 112(b) rejection of claim(s) 1, 2, 4-6, 8-12 is respectfully withdrawn.
Applicant's arguments and amendments with regards to the 35 U.S.C. § 101 rejection of claim(s) 1, 2, 4-6, 8-12, have been fully considered and are persuasive. The 35 U.S.C. § 101 rejection of claim(s) 1, 2, 4-6, 8-12 is respectfully withdrawn.
Applicant's following arguments with respect to claim(s) 1, 2, 4-6, 8-12 under 35 U.S.C. § 103 have been considered but are moot because the new ground of rejection does not rely on any reference or portion of reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claims 2, 4-6, 8-15 are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention.
Claims 1 and 5, each recite “determine whether a device or component is a non-defective item or a defective item” ... “wherein the device includes medical devices and manufacturing components”
The original specification does not teach a device that is a manufacturing component, and therefore does not teach “wherein the device includes... manufacturing components” in this context.
The original specification teaches determining whether a device or component is a non-defective item or a defective item, wherein the device includes medical devices and “various devices or their components”. However, the specification is silent regarding the specifics of these “various devices or their components”.
Therefore the above noted limitations of claim(s) 1 and 5 do/does not have support in the original specification.
Claims 2, 4, 6, 8-15 merely recite additional functions performed by the inventions of claims 1 and 5. Accordingly, claims 2, 4, 6, 8-15are also rejected under 35 U.S.C. 112(a).
Claim 13 recites “a feature extraction component implemented as multiple convolutional neural network layers configured to extract spatial and temporal patterns from the sensor data; ...a classification component implemented as multiple fully-connected neural network layers configured to process the extracted patterns to generate classification scores; and ...a real-time data processing pipeline configured to: ...receive streaming sensor measurements at a rate of at least 100 Hz from multiple temperature, humidity and pressure sensors; ...normalize and filter the streaming sensor measurements to remove noise; and ...batch the filtered measurements for parallel processing through the neural network layers to generate defect determinations with sub-second latency”.
The original specification does not disclose the concepts of
convolutional neural network, multiple layers, extract spatial and temporal patterns, fully-connected layers, process the extracted patterns to generate classification scores, receive streaming sensor measurements, a rate of at least 100 Hz, normalize, filter, remove noise, batch, parallel processing or sub-second latency.
Therefore the above noted limitations of claim(s) 13 do/does not have support in the original specification.
Claim 14 recites “maintain a distributed database of historical sensor measurements, prediction rules, and defect determinations across multiple manufacturing facilities; ...analyze correlations between sensor measurement patterns and defect occurrences across the distributed database using automated feature extraction; ...automatically adjust prediction rule thresholds based on the analyzed correlations to optimize defect detection accuracy; and ...propagate the adjusted thresholds to prediction rule instances running across the multiple manufacturing facilities in real-time”
The original specification does not disclose the concepts of
distributed database across multiple manufacturing facilities, historical sensor measurements, storing prediction rules, storing defect determinations, propagate the adjusted thresholds in real-time, prediction rule instances running across the multiple manufacturing facilities.
Therefore the above noted limitations of claim(s) 14 do/does not have support in the original specification.
Claim 15 recites “a hardware acceleration unit comprising tensor processing circuits configured to: ...parallelize neural network computations for processing multiple sensor data streams simultaneously; ...perform real-time feature extraction from the sensor measurements using optimized matrix operations; and ...generate defect classification outputs within microsecond-scale latency requirements for high-speed manufacturing lines; ...wherein the hardware acceleration unit enables processing of sensor data at rates exceeding 10,000 measurements per second to provide real-time defect detection”.
The original specification does not disclose the concepts of hardware acceleration unit, tensor processing circuits, parallelize neural network computations, processing multiple sensor data streams simultaneously; perform real-time feature extraction, using optimized matrix operations, using matrix operations, microsecond-scale latency requirements, high-speed manufacturing lines, rates exceeding 10,000 measurements per second, provide real-time defect detection.
Therefore the above noted limitations of claim(s) 15 do/does not have support in the original specification.
Applicant is requested to make appropriate amendments to the claims or clearly point of the specific portions of paragraphs in the specification that support the claim limitations.
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.
Claims 1, 2, 4-6, 8-12 are rejected under 35 U.S.C. 103 as being unpatentable over Shimada (US 20120084236 A1), in view of Hetherington (US 20200302318 A1), Noone (US 20200166909 A1) and Chen (US 10929762 B1).
Regarding claim 1, Shimada teaches a prediction rule calibration system comprising: a processor coupled to a memory storing instructions for the processor to function as: a prediction rule simplification unit...; a branch condition search unit... and a threshold optimization unit ... (Shimada [93-104] device with processor executes functions to perform operations),
a prediction rule simplification unit that modifies a prediction rule of a learning model outputs a predicted value...by a tree structure using an evaluation metric and a restriction (Shimada [3] processes may be for machine learning model(s), Shimada [40, 41, 49, 50-52, 59, 60] based on rule stipulations (attribute and constraints), rule may be expanded to determine attribute set to apply condition to, model may use a decision tree to predict causes and decisions);
a branch condition search unit that updates a part of a modified branch condition for prediction rule based on calibration information expressing a request for a specific branch condition (Shimada [38,-0, 41, 49, 50-52, 59, 60] based on stipulations and attribute set, branching attribute is determined); and
a threshold optimization unit that updates a part of a threshold of the modified prediction rule based on the calibration information, wherein the threshold optimization unit iteratively changes the threshold of branch conditions to calculate a simple prediction rule with a high evaluation metric while satisfying the calibration information (Shimada [51-54] based on updates- constraints may be deleted, Shimada [46, 47] based on determined branching conditions and analysis of reducing entropy- branching attribute(s) may be updated to reduce entropy, Shimada [40, 41, 46, 47, 49, 50-54, (See 40-57 as a whole)] based on determined attribute and constraint, stipulations may be applied to nodes above or below (before and after, ) the determined attribute in the determined attribute branch, based on determined branching conditions and analysis of reducing entropy- determined or branching attribute(s) may be updated to reduce entropy, based on analysis of reducing entropy (which would increase the inverse metric- i.e. information content) for candidate list, attributes may be updated to reduce entropy, tree is evaluated recursively over various nodes)),
wherein a section that is updated in the branch condition search unit is limited to a branch condition specified by the calibration information, wherein, in the branch condition search unit, a process of extracting a variable having a predetermined correlation with a variable forming a branch condition before updating is performed, wherein the predetermined correlation is determined based on a correlation coefficient calculated between the variable forming the branch condition before updating and other explanatory variables, wherein the branch condition that includes the extracted variable is set as a candidate of the branch condition (Shimada [40, 41, 49, 50-54] based on determined attribute and constraint, stipulations may be applied to nodes above or below (before and after- limited branching conditions) the determined attribute (extracted variable) in the determined attribute branch, Shimada [43-47] based on determined branching conditions being the same (correlation coefficient of zero or one)- determined or branching attribute(s) may be added to candidate list, based on analysis of reducing entropy),
wherein the modified prediction rule is inserted into conditions requested by a user (Shimada [38, 39, 71] modified rules are output as decision tree).
Shimada does not specifically teach a predicted value of an object variable from a combination of values of variable learning data... , wherein the learning model processes sensor data that includes temperature, humidity, and pressure measurements as explanatory variables to output a binary determination of whether an item is defective or non-defective, wherein 0 indicates non-defective and 1 indicates defective; ...wherein the prediction rule simplification unit creates a simple prediction rule that simplifies the learning data using a tree structure where nodes express branch conditions that determine paths based on explanatory variable combinations, branch conditions have real number thresholds, and leaves correspond to prediction values; ...wherein the threshold optimization unit limits branch condition candidates based on correlation analysis for improved search efficiency; ...to determine whether a device or component is a non-defective item or a defective item using the sensor data, wherein the device includes medical devices and manufacturing components.
However Hetherington teaches a prediction rule simplification unit that modifies a prediction rule of a learning model that outputs a predicted value of an object variable from a combination of values of variable learning data by a tree structure using an evaluation metric and a restriction.... wherein the learning model processes ... data that includes ... explanatory variables to output a binary determination of whether an item is defective or non-defective (Hetherington [5, 21, 24] machine learning (ML) model uses input data including features which may be explanatory variable, and model outputs a prediction based on input data, Hetherington [20-22, 33, 35-37, 76, 107] prediction is based on using tree structure of input features, optimization metrics (evaluation metric) and thresholds (restriction) may be specified, prediction may be a binary prediction of classification, classification may be positive or negative and may be for defect);
wherein the prediction rule simplification unit creates a simple prediction rule that simplifies the learning data using a tree structure where nodes express branch conditions that determine paths based on explanatory variable combinations, branch conditions have real number thresholds, and leaves correspond to prediction values (Hetherington [40, 60, 72, 80-83] decision tree is formed using input data features, tree may be optimized, decision tree may be mapped to rules using tree path traversal, Hetherington [47, 60, 62, 63, 100, 101, 118] rules may be optimized, Hetherington Fig. 3, [29, 37-39] tree has nodes with branch conditions (split values for features which may be numeric), and prediction values at leaves (positive or negative)),
wherein the threshold optimization unit iteratively changes the threshold of branch conditions to calculate a simple prediction rule with an optimized evaluation metric ..., wherein the threshold optimization unit limits branch condition candidates based on correlation analysis for improved search efficiency (Hetherington [62, 72, 75] rule thresholds and branch conditions may be iteratively updated for optimization of searching to tree and or rules, Hetherington [21, 25, 40-42, 60-62, 104] optimization may be based on feature correlation (importance) with prediction).
It would have been obvious to one of an ordinary skill in the art before the effective filing date of the claimed invention, to have incorporated the concept taught by Hetherington of a prediction rule simplification unit that modifies a prediction rule of a learning model that outputs a predicted value of an object variable from a combination of values of variable learning data by a tree structure using an evaluation metric and a restriction.... wherein the learning model processes ... data that includes ... explanatory variables to output a binary determination of whether an item is defective or non-defective; wherein the prediction rule simplification unit creates a simple prediction rule that simplifies the learning data using a tree structure where nodes express branch conditions that determine paths based on explanatory variable combinations, branch conditions have real number thresholds, and leaves correspond to prediction values; wherein the threshold optimization unit iteratively changes the threshold of branch conditions to calculate a simple prediction rule with an optimized evaluation metric ..., wherein the threshold optimization unit limits branch condition candidates based on correlation analysis for improved search efficiency,
into the invention suggested by Shimada, since both inventions are directed towards decision trees that may generate a predicted value, and incorporating the teaching of Hetherington into the invention suggested by Shimada would provide the added advantage of allowing rules to be optimized and therefore optimizing searching of the tree and rules- by using path traversal in rule optimization, and the combination would perform with a reasonable expectation of success (Hetherington Fig. 3 [5, 21, 24, 20-22, 33, 35-37, 76, 107, 40, 60, 72, 80-83, 47, 60, 62, 63, 100, 101, 118, 29, 37-39, 62, 72, 75, 21, 25, 40-42, 60-62, 104]).
Shimada and Hetherington do not specifically teach wherein the learning model processes sensor data that includes temperature, humidity, and pressure measurements ..., wherein 0 indicates non-defective and 1 indicates defective; to determine whether a device or component is a non-defective item or a defective item using the sensor data, wherein the device includes medical devices and manufacturing components.
However Noone teaches conditions ... to determine whether a device or component is a non-defective item or a defective item using ... sensor data, wherein the device includes medical devices and manufacturing components (Noone [39-41, 245, 272] decision tree may have branching conditions to determine whether there is a defect in manufacturing process, defect may be determined based in sensor information and for medical equipment and manufactured objects).
It would have been obvious to one of an ordinary skill in the art before the effective filing date of the claimed invention, to have incorporated the concept taught by Noone of conditions ... to determine whether a device or component is a non-defective item or a defective item using ... sensor data, wherein the device includes medical devices and manufacturing components, into the invention suggested by Shimada and Hetherington,
so that the modified prediction rule ... inserted into conditions requested by a user of Shimada and Lillo,
results in modified prediction rule is inserted into conditions requested by a user to determine whether a device or component is a non-defective item or a defective item using ... sensor data, wherein the device includes medical devices and manufacturing components;
since both inventions are directed towards decision trees that may generate a predicted value, and incorporating the teaching of Noone into the invention suggested by Shimada and Hetherington would provide the added advantage of allowing modification of prediction rules used to determine defects in a variety of devices, and the combination would perform with a reasonable expectation of success (Noone [39-41, 245, 272]).
Shimada, Hetherington and Noone do not specifically teach wherein the learning model processes sensor data that includes temperature, humidity, and pressure measurements ... to output a binary determination of whether an item is defective or non-defective, wherein 0 indicates non-defective and 1 indicates defective.
However Chen teaches wherein the learning model processes sensor data that includes temperature, humidity, and pressure measurements... to output a binary determination of whether an item is defective or non-defective, ...wherein the determination may be indicated as 0 or 1... (Chen Col 4, line 48- Col 8, line 8, sensor data may include temperature, humidity, and pressure measurements, Chen Background, Col 26, line 55- Col 27, line 8, machine learning model may be used to determine whether or not item is defective, Chen Col 22, lines 18-34, model may use a tree, Chen ).
It would have been obvious to one of an ordinary skill in the art before the effective filing date of the claimed invention, to have incorporated the concept taught by Chen of wherein the learning model processes sensor data that includes temperature, humidity, and pressure measurements to output a binary determination of whether an item is defective or non-defective, wherein 0 indicates non-defective and 1 indicates defective, into the invention suggested by Shimada, Hetherington and Noone, so that
the data that includes ... explanatory variables of Shimada, Hetherington and Noone,
is sensor data that includes temperature, humidity, and pressure measurements as taught by Chen,
and
the binary determination of whether an item is defective or non-defective of Shimada, Hetherington and Noone,
is indicated as 0 or 1 as taught by Chen,
to result in wherein 0 indicates non-defective and 1 indicates defective;
since both inventions are directed towards using sensor data to make a determination as to predict whether an item is defective, and incorporating the teaching of Chen into the invention suggested by Shimada, Hetherington and Noone would provide the added advantage of allowing specific sensor information (i.e. temperature, humidity, and pressure measurements to predict whether an item is defective, and the combination would perform with a reasonable expectation of success (Chen Col 4, line 48- Col 8, line 8, Background, Col 26, line 55- Col 27, line 8, Col 22, lines 18-34).
Regarding claim 2, Shimada, Hetherington, Noone and Chen teach the invention as claimed in claim 1 above. Shimada further teaches wherein, in the branch condition search unit and the threshold optimization unit, a section in which the modified prediction rule is updated is limited to a lower subtree of a branch condition specified by the calibration information (Shimada [40, 41, 49, 50-54] based on determined attribute and constraint, stipulations may be applied to nodes above or below (before and after) the determined attribute in the determined attribute branch).
Regarding claim 4 Shimada, Hetherington, Noone and Chen teach the invention as claimed in claim 1 above. Shimada further teaches wherein, in the branch condition search unit and the threshold optimization unit, a section in which the modified prediction rule is updated is limited to a lower subtree of a branch condition specified by the calibration information (Shimada [40, 41, 49, 50-54] based on determined attribute and constraint, stipulations may be applied to nodes above or below (before and after) the determined attribute in the determined attribute branch).
Claim 5 is directed towards a method performing steps similar in scope to the instructions executed by the system of claim 1, and is rejected under the same rationale.
Claim(s) 6, is/are dependent on claim 5 above, is/are directed towards a method performing steps similar in scope to the instructions executed by the system of claim(s) 2, and is/are rejected under the same rationale.
Regarding claim 8, Shimada, Hetherington, Noone and Chen teach the invention as claimed in claim 5 above. Shimada further teaches wherein, in the modifying of the prediction rule and the updating of the part of the threshold, a section in which the modified prediction rule is updated is limited to a lower subtree of a branch condition specified by the calibration information (Shimada [40, 41, 49, 50-54] based on determined attribute and constraint, stipulations may be applied to nodes above or below (before and after) the determined attribute in the determined attribute branch).
Regarding claim 9, Shimada, Hetherington, Noone and Chen teach the invention as claimed in claim 1 above. Claim 1 further teaches wherein the sensor data includes temperature, humidity, and pressure measurements.
Shimada does not specifically teach ...sensor measurements acquired during manufacture of the medical device, and wherein the learning model generates a binary defect determination output by processing the sensor measurements through multiple neural network layers to extract features for defect classification
However Noone teaches sensor measurements acquired during manufacture of the medical device (Noone 3, 32] sensor measurements may be acquired during fabrication), and
wherein the learning model generates a binary defect determination output by processing the sensor measurements through multiple neural network layers to extract features for defect classification (Noone [3, 39, 250-252, 219, 222-224] determination is made whether or not fabricated item meets specified quality, determination may be made with multiple neural network layers that extract features for defect classification).
Regarding claim 10, Shimada, Hetherington, Noone and Chen teach the invention as claimed in claim 1 above. Shimada does not specifically teach wherein the learning model is converted into a simplified prediction rule, and wherein the calibration information comprises expert knowledge inputted as requests for prediction values or specific prediction rules that modify thresholds and branch conditions of the simplified prediction rule.
However Hetherington teaches wherein the learning model is converted into a simplified prediction rule (Hetherington [20-23] model may be trained and include the simplified prediction rule(s)).
Further, Noone teaches wherein the calibration information comprises expert knowledge inputted as requests for prediction values or specific prediction rules that modify thresholds and branch conditions of the simplified prediction rule (Noone [245] model may be decision tree-based expert system that uses knowledge base with if-then rules).
Regarding claim 11 Shimada, Hetherington, Noone and Chen teach the invention as claimed in claim 1 above. Shimada further teaches wherein the threshold optimization unit iteratively modifies thresholds of branch conditions in the prediction rule to achieve an optimized evaluation metric while satisfying the calibration information (Shimada [51-54] based on updates- constraints may be deleted, Shimada [46, 47] based on determined branching conditions and analysis of reducing entropy- branching attribute(s) may be updated to reduce entropy, Shimada [40, 41, 46, 47, 49, 50-54, (See 40-57 as a whole)] based on determined attribute and constraint, stipulations may be applied to nodes above or below (before and after, ) the determined attribute in the determined attribute branch, based on determined branching conditions and analysis of reducing entropy- determined or branching attribute(s) may be updated to reduce entropy, based on analysis of reducing entropy (which would increase the inverse metric- i.e. information content) for candidate list, attributes may be updated to reduce entropy, tree is evaluated recursively over various nodes)).
Regarding claim 12, Shimada, Hetherington, Noone and Chen teach the invention as claimed in claim 1 above. Shimada further teaches wherein the branch condition search unit limits branch condition candidates to those including variables having the predetermined correlation with the variable forming the branch condition before updating, where the correlation is determined based on a correlation coefficient between the variable forming the branch condition before updating and other explanatory variables (Shimada [40, 41, 49, 50-54] based on determined attribute and constraint, stipulations may be applied to nodes above or below (before and after- limited branching conditions) the determined attribute (extracted variable) in the determined attribute branch, Shimada [43-47] based on determined branching conditions being the same (correlation coefficient of zero or one)- determined or branching attribute(s) may be added to candidate list, based on analysis of reducing entropy).
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SANCHITA ROY whose telephone number is (571)272-5310. The examiner can normally be reached Monday-Friday 12-8.
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.
SANCHITA . ROY
Primary Examiner
Art Unit 2146
/SANCHITA ROY/Primary Examiner, Art Unit 2146
Prosecution Insights