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 of Claims
Claims 1-20 are currently pending in this application.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-7 are rejected under 35 U.S.C. 112(b), second paragraph, as failing to set forth the subject matter which the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the applicant regards as the invention.
Claim 1 recites “one or more matches” in line 5. It is unclear whether “the matches” in line 6, 8, and 15 are referring to the same “one or more matches”. Furthermore, claim 1, line 8 teaches “historical defects”. This is also unclear whether it refers to the same “historical defects” in line 5. Line 17 also recites “one or more root causes”, which is unclear whether it refers to “one or more possible root causes” from line 6.
Claim 5 line 6 recites “the ranking”, which is unclear whether it refers to “a ranked order” in line 5. Claims 5 and 6 both also have unclear wording surrounding “the matches”, and thus, are also rejected under 112(b).
Claims 2-7 depend on base claim 1, further rejecting them under 112(b) as well. Appropriate correction is required.
Claims 8-14 are rejected under 35 U.S.C. 112(b), second paragraph, as failing to set forth the subject matter which the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the applicant regards as the invention.
Claim 8’s line 6, 8, and 15 recite “the matches”. This is unclear whether it refers to “one or more matches” in line 6. Furthermore, line 8 recites “historical root cause”, which is unclear if it refers to “a historical root cause” from line 7. Also, line 7 and 14 recite “historical defects”, which is unclear whether it refers to the same “historical defects” in its line 6. Adding onto this, line 17, recites “one or more root causes”, which is unclear whether it refers to the same “one or more possible root causes” from line 7.
Claim 10, line 1, recites “the trained machine learning”, which is unclear whether it refers to “the trained machine learning model” from line claim 8, line 5.
Claim 12’s line 3 and claim 13’s line 2 both recite “the matches” and is unclear whether it refers to “one or more matches” in claim 8’s line 6.
Claim 14, line 3, recites “one or more images”. This is unclear whether it refers to the same “one or more images” in line 1.
Claim 9, 11, 12, and 13 also have unclear wording surrounding “historical defects”, and thus, are also rejected under 112(b).
Claims 9-14 depend on base claim 8, further rejecting them under 112(b) as well. Appropriate correction is required.
Claims 15-20 are rejected under 35 U.S.C. 112(b), second paragraph, as failing to set forth the subject matter which the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the applicant regards as the invention.
Claim 15, line 6, recites “the matches”. This is unclear whether it refers to “one or
more matches” of the same line. Furthermore, line 8 recites “historical root causes”. This is unclear whether it refers to the same “a historical root cause” in line 7. Line 9 and 14 recites “historical defects”, which is unclear whether it refers to the same “historical defects” in line 6.
Claims 16, 17, 18, and 19 also have unclear wording surrounding “historical defects”, and thus, are also rejected under 112(b).
Claim 20, line 2, recites “the matches”. This is unclear whether it refers to “one or
more matches” of claim 19, line 3. Also, line 3 recites “historical defects”, which is unclear whether it refers to the same “historical defects” in claim 19, line 3.
Claims 16-20 depend on base claim 15, further rejecting them under 112(b) as well. Appropriate correction is required.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claim(s) 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Chowdhury et. Al (US 20240144464 A1) in view of Balasubramanian et. Al (US 20220171995 A1).
With respect to claim 1, Chowdhury teaches a method, comprising:
receiving, by a processing device, first information about one or more defects of a substrate processed in a substrate processing system using a process recipe (Chowdhury: “A method includes obtaining, by a processing device, data indicative of locations of defects of a substrate” [0028]; “may generate as output a process recipe predicted to produce a product of that design” [0086]);
processing, by the processing device, the first information using a trained machine learning model that outputs one or more matches to historical defects (Chowdhury: “A new image, to be provided to a machine learning model, may be generated. The machine learning model may be configured to receive data indicative of locations of defects of a substrate and generate as output one or more classifications of defect patterns associated with the substrate” [0024]) for the matches, each of the historical defects having a historical root cause, wherein one or more possible root causes of the one or more defects correspond to historical root causes associated with the matches to historical defects (Chowdhury: “At block 320, system 300 uses the trained model to receive defect data 354 and determines (e.g., extracts), from the output of the trained model, classification data 369. Classification data 369 may classify defects of a substrate. Classification data 369 may assign a category to a pattern of defects of a substrate…Classification data 369 may be correlated to and/or include predicted root causes of defects of a substrate” [0106]);
receiving, by the processing device, user input selecting a subset of the one or more matches (Chowdhury: “A user (e.g., a subject matter expert) may be presented with an image visualizing locations of defects of the substrate. The user may classify a pattern of the defects.” [0023]);
that outputs one or more updated matches to historical defects (Chowdhury: “Data indicative of defect locations may be provided to a trained machine learning model as input. The trained machine learning model may generate, as output, a classification of a pattern of defects of the substrate.” [0017]) for the matches; and
outputting, by the processing device, an indication of one or more maintenance operations associated with the one or more updated matches to resolve one or more root causes associated with at least one of the one or more updated matches (Chowdhury: “A library may exist that correlates patterns of defects on a substrate to possible causes, possible corrective actions, appropriate maintenance to perform, or the like.” [0022]).
Chowdhury does not teach score values, reprocessing the first information using the updated trained machine learning model, and updated score values.
However, it is known that Balasubramanian teaches a method comprising score values (“a weighted anomaly score based on the probably that the item was classified correctly” [0151]), updated score values for matches (“using the retrained anomaly detection machine learning model, wherein an indication of an anomaly and a score for the anomaly are provided”, claim 13), updating, by the processing device, the trained machine learning model based on the user input to produce an updated trained machine learning model (In some embodiments, the training data can include of images with defects with a mark over the location of the defects they want to identify (“e.g. location of stains on cans) and/or a specified region where the defects are expected (such as expected location of machine parts). A custom machine learning (ML) model is generated to compare subsequent images to the baseline and through that comparison, identifies further defects (e.g. more stains on the cans or missing machine parts).” [0036]; “The user 143 may then provide feedback to the console 106 to engage in a feedback re-training of the model 103 (the feedback loop indicted by the large rectangular box).” [0058]), and reprocessing the first information using the updated trained machine learning model (“The selected model 2521 is then applied to unlabeled data 2502 compute sample importance scores, ranking information, and/or predictions for the unlabeled set which are stored in a history 2507. A feedback mechanism 2509 from a user on selected samples from the unlabeled data and augments the training set for the next iteration based on the feedback” [0157]).
With respect to claim 2, the previously cited reference(s) teach the limitations of claim 1 which claim 2 depends on. Chowdhury teaches the method of claim 1, further comprising:
retrieving, by the processing device, from a data structure, second information associated with the one or more matches to historical defects, wherein retrieving of the second information is based on the first information, and wherein the second information is indicative of the one or more maintenance operations (“In some embodiments, a second machine learning model may be used to determine a corrective actions. For example, a series of substrate defect pattern classifications may be provided to a trained machine learning model, and the trained machine learning model may generate as output one or more recommended corrective actions.” [0127]; “the first data indicative of locations of defects of the first substrate comprises an image, wherein the image indicates defect locations of the first substrate; and the second data indicative of locations of the second substrate comprises a list of locations of defects of the second substrate”, claim 10).
With respect to claim 3, the previously cited reference(s) teach the limitations of claim 1 which claim 3 depends on. Balasubramanian teaches the method of claim 1, wherein updating the trained machine learning model based on the user input comprises:
providing, by the processing device, the user input to the trained machine learning model as training data, wherein the trained machine learning model is further trained based on the user input (“At circle 3, an initial training of the models 103 is performed using the training dataset and unlabeled data from the testing dataset is subjected to the trained models 103. At circle 4, a request is sent to a feedback device 145 (e.g., a user providing feedback via this device) to evaluate the results of the trained models 103 on the unlabeled testing dataset and feedback is received by the anomaly detection service/component 102.” [0053]; “The feedback training 109 is engaged based on the response to update the models 103” [0054]).
With respect to claim 4, the previously cited reference(s) teach the limitations of claim 1 which claim 4 depends on. Chowdhury and Balasubramanian combined teaches the method of claim 1, wherein the trained machine learning model comprises:
a first portion to output the one or more matches to historical defects based on the first information (Chowdhury: “The method further includes obtaining, as output from the trained machine learning model, a classification of the locations of the defects”, abstract); and
a second portion to assign the score values, based at least in part on the user input, for the one or more matches to historical defects (Balasubramanian: “The model 103 is applied to the source data and generates a result which is provided to a console 106 to present to a user 143. Examples of how a result may be presented are detailed later, but may include, but are not limited to: the image evaluated, a score for the image” [0057]).
With respect to claim 5, the previously cited reference(s) teach the limitations of claim 1 which claim 5 depends on. Chowdhury and Balasubramanian combined teaches the method of claim 1, further comprising:
ranking, by the processing device, the one or more matches to historical defects based on the score values for the matches (Balasubramanian: “In some embodiments, the importance ranking score is a weighted anomaly score based on the probably that the item was classified correctly. A high probability suggests a correct classification. The higher the importance score, the more important an item is to review (e.g., the most anomalous). Other rankings include the confidence scores.” [0151]); and
providing, by the processing device, information associated with the one or more matches to historical defects for display on a graphical user interface (GUI) (Chowdhury: “Predictive component 114 may receive/obtain data from model 190. Predictive component 114 may use data obtained from model 190 to generate predictive data 168. In some embodiments, predictive component 114 provides predictive data 168 to client device 120, and client device 120 causes a corrective action via corrective action component 122 in view of predictive data 168. A corrective action performed by client device 120 may include displaying predictive data 168 for a user.” [0043]; Chowdhury: “Corrective action component 122 may receive user input (e.g., via a Graphical User Interface (GUI) displayed via the client device 120)” [0042]) in a ranked order based on the ranking (Balasubramanian: “In some embodiments, the items are displayed by ranking. In some embodiments, a user can select how ranking scores are used to order (e.g., higher ranking are shown first, etc.)” [0153]).
With respect to claim 6, the previously cited reference(s) teach the limitations of claim 5 which claim 6 depends on. Balasubramanian teaches the method of claim 5, wherein the user input comprises a ranking of one or more images associated with the matches to historical defects, and wherein the score values are determined based at least in part on the ranking (“An importance ranking score is computed for each item of the unlabeled dataset based on the result of applying the trained model at 2404. The ranking of the training set items is made in terms of how informative (when re-annotated) those items might be.” [0150]; “In some embodiments, a user can select how ranking scores are used to order (e.g., higher ranking are shown first, etc.).” [0153]).
With respect to claim 7, the previously cited reference(s) teach the limitations of claim 1 which claim 7 depends on. Chowdhury teaches the method of claim 1, wherein the first information comprises one or more images of the one or more defects of the substrate processed in the substrate processing system (“A method includes obtaining, by a processing device, data indicative of locations of defects of a substrate. The method further includes generating an image indicating the locations of the defects”, abstract), and wherein the trained machine learning model outputs one or more images of the historical defects (“A model may receive an image (e.g., a block diagram of a target substrate design) and generate a related image (e.g., a realistic image of a simulate substrate)” [0086]).
Regarding claims 8 – 14, they are directed to a system or apparatuses to implement the method of steps set forth in claims 1– 7. Chowdhury et. Al and Balasubramanian et. Al teaches the claimed method of steps in claims 1– 7. Therefore, Chowdhury et. Al and Balasubramanian et. Al teaches the system or apparatuses, to implement the claimed method of steps, in claims 8 – 14.
With respect to claim 8, Chowdhury teaches a system, comprising memory and a processing device coupled to the memory, wherein the processing device is configured to (“In a further aspect, the computer system 600 may include a processing device 602, a volatile memory 604 (e.g., Random Access Memory (RAM)), a non-volatile memory 606 (e.g., Read-Only Memory (ROM) or Electrically-Erasable Programmable ROM (EEPROM)), and a data storage device 618, which may communicate with each other via a bus 608” [0153]):
receive first information about one or more defects of a substrate processed in a substrate processing system using a process recipe (“A method includes obtaining, by a processing device, data indicative of locations of defects of a substrate, abstract; A model may receive an indication of a target product design and may generate as output a process recipe predicted to produce a product of that design.” [0086])
process the first information using a trained machine learning model that outputs one or more matches to historical defects and score values for the matches (Balasubramanian: “a weighted anomaly score based on the probably that the item was classified correctly” [0151]), each of the historical defects having a historical root cause, wherein one or more possible root causes of the one or more defects correspond to historical root causes associated with the matches to historical defects (“At block 320, system 300 uses the trained model to receive defect data 354 and determines (e.g., extracts), from the output of the trained model, classification data 369. Classification data 369 may classify defects of a substrate… Classification data 369 may be correlated to and/or include predicted root causes of defects of a substrate.” [0106]);
receive user input selecting a subset of the one or more matches (“A user (e.g., a subject matter expert) may be presented with an image visualizing locations of defects of the substrate. The user may classify a pattern of the defects.” [0023]);
update the trained machine learning model based on the user input to produce an updated trained machine learning model (“In some aspects of the disclosure, a method includes providing, as training input to a machine learning model, a first plurality of images.” [0005]; “Selected model 308 may be retrained based on this data. Retraining may generate an updated machine learning model. Retraining may include providing retraining input and/or retraining target output to selected model 308.” [0107]);
the trained machine learning model that outputs one or more updated matches to historical defects for the matches (“Data indicative of defect locations may be provided to a trained machine learning model as input. The trained machine learning model may generate, as output, a classification of a pattern of defects of the substrate.” [0017]); and
output an indication of one or more maintenance operations associated with the one or more updated matches to resolve one or more root causes associated with at least one of the one or more updated matches (Chowdhury: “A library may exist that correlates patterns of defects on a substrate to possible causes, possible corrective actions, appropriate maintenance to perform, or the like.” [0022]).
Chowdhury does not teach a system reprocessing the first information using the updated trained machine learning model, and updated score values.
However, it is known that Balasubramanian teaches a system (fig.1) comprising updated score values for matches (“using the retrained anomaly detection machine learning model, wherein an indication of an anomaly and a score for the anomaly are provided”, claim 13), and reprocessing the first information using the updated trained machine learning model (“The selected model 2521 is then applied to unlabeled data 2502 compute sample importance scores, ranking information, and/or predictions for the unlabeled set which are stored in a history 2507. A feedback mechanism 2509 from a user on selected samples from the unlabeled data and augments the training set for the next iteration based on the feedback” [0157]).
With respect to claim 9, the previously cited reference(s) teach the limitations of claim 8 which claim 9 depends on. Chowdhury teaches the system of claim 8, wherein the processing device is further configured to: retrieve, from a data structure, second information associated with the one or more matches to historical defects, wherein retrieving of the second information is based on the first information, and wherein the second information is indicative of the one or more maintenance operations (“In some embodiments, a second machine learning model may be used to determine a corrective actions. For example, a series of substrate defect pattern classifications may be provided to a trained machine learning model, and the trained machine learning model may generate as output one or more recommended corrective actions.” [0127]; “the first data indicative of locations of defects of the first substrate comprises an image, wherein the image indicates defect locations of the first substrate; and the second data indicative of locations of the second substrate comprises a list of locations of defects of the second substrate”, claim 10).
With respect to claim 10, the previously cited reference(s) teach the limitations of claim 8 which claim 10 depends on. Balasubramanian teaches the system of claim 8, wherein to update the trained machine learning, the processing device is configured to:
provide the user input to the trained machine learning model as training data, wherein the trained machine learning model is further trained based on the user input (“At circle 3, an initial training of the models 103 is performed using the training dataset and unlabeled data from the testing dataset is subjected to the trained models 103. At circle 4, a request is sent to a feedback device 145 (e.g., a user providing feedback via this device) to evaluate the results of the trained models 103 on the unlabeled testing dataset and feedback is received by the anomaly detection service/component 102.” [0053]; “The feedback training 109 is engaged based on the response to update the models 103” [0054]).
With respect to claim 11, the previously cited reference(s) teach the limitations of claim 8 which claim 11 depends on. Chowdhury and Balasubramanian combined teaches the system of claim 8, wherein the trained machine learning model comprises:
a first portion to output the one or more matches to historical defects based on the first information (Chowdhury: “Data indicative of defect locations may be provided to a trained machine learning model as input. The trained machine learning model may generate, as output, a classification of a pattern of defects of the substrate.” [0017]); and
a second portion to assign the score values, based at least in part on the user input, for the one or more matches to historical defects (Balasubramanian: “The model 103 is applied to the source data and generates a result which is provided to a console 106 to present to a user 143. Examples of how a result may be presented are detailed later, but may include, but are not limited to: the image evaluated, a score for the image” [0057]).
With respect to claim 12, the previously cited reference(s) teach the limitations of claim 8 which claim 12 depends on. Chowdhury and Balasubramanian combined teaches the system of claim 8, wherein the processing device is further configured to:
rank the one or more matches to historical defects based on the score values for the matches (Balasubramanian: “In some embodiments, the importance ranking score is a weighted anomaly score based on the probably that the item was classified correctly. A high probability suggests a correct classification. The higher the importance score, the more important an item is to review (e.g., the most anomalous). Other rankings include the confidence scores.” [0151]); and
provide information associated with the one or more matches to historical defects for display on a graphical user interface (GUI) (Chowdhury: “Predictive component 114 may receive/obtain data from model 190. Predictive component 114 may use data obtained from model 190 to generate predictive data 168. In some embodiments, predictive component 114 provides predictive data 168 to client device 120, and client device 120 causes a corrective action via corrective action component 122 in view of predictive data 168. A corrective action performed by client device 120 may include displaying predictive data 168 for a user.” [0043]; Chowdhury: “(e.g., via a Graphical User Interface (GUI) displayed via the client device 120)” [0042]) in a ranked order based on the ranking (Balasubramanian: “In some embodiments, the items are displayed by ranking. In some embodiments, a user can select how ranking scores are used to order (e.g., higher ranking are shown first, etc.)” [0153]).
With respect to claim 13, the previously cited reference(s) teach the limitations of claim 12 which claim 13 depends on. Balasubramanian teaches the system of claim 12, wherein the user input comprises a ranking of one or more images associated with the matches to historical defects, and wherein the score values are determined based at least in part on the ranking (“An importance ranking score is computed for each item of the unlabeled dataset based on the result of applying the trained model at 2404. The ranking of the training set items is made in terms of how informative (when re-annotated) those items might be.” [0150]; “In some embodiments, a user can select how ranking scores are used to order (e.g., higher ranking are shown first, etc.).” [0153]).
With respect to claim 14, the previously cited reference(s) teach the limitations of claim 8 which claim 14 depends on. Chowdhury teaches the system of claim 8, wherein the first information comprises one or more images of the one or more defects of the substrate processed in the substrate processing system (“A method includes obtaining, by a processing device, data indicative of locations of defects of a substrate. The method further includes generating an image indicating the locations of the defects,” abstract), and wherein the trained machine learning model outputs one or more images of the historical defects (“A model may receive an image (e.g., a block diagram of a target substrate design) and generate a related image (e.g., a realistic image of a simulate substrate)” [0086]).
Regarding claims 15 – 20, they are directed to a non-transitory machine-readable storage medium to implement the method of steps set forth in claims 1– 7. Chowdhury et. Al teaches a non-transitory machine-readable storage medium (fig. 1) and Balasubramanian et. Al teaches a non-transitory computer-readable storage media in [0033]. Therefore, Chowdhury et. Al and Balasubramanian et. Al teaches the system or apparatuses, to implement the claimed element, in claims 15 – 20.
With respect to claim 15, Chowdhury teaches a non-transitory machine-readable storage medium storing instructions which, when executed, cause a processing device to perform operations comprising (“In a further aspect, the computer system 600 may include a processing device 602, a volatile memory 604 (e.g., Random Access Memory (RAM)), a non-volatile memory 606 (e.g., Read-Only Memory (ROM) or Electrically-Erasable Programmable ROM (EEPROM)), and a data storage device 618, which may communicate with each other via a bus 608” [0153]):
receiving first information about one or more defects of a substrate processed in a substrate processing system using a process recipe (“A method includes obtaining, by a processing device, data indicative of locations of defects of a substrate” [0028]; “may generate as output a process recipe predicted to produce a product of that design” [0086]);
processing the first information using a trained machine learning model that outputs one or more matches to historical defects and score values for the matches, each of the historical defects having a historical root cause, wherein one or more possible root causes of the one or more defects correspond to historical root causes associated with the matches to historical defects (“At block 320, system 300 uses the trained model to receive defect data 354 and determines (e.g., extracts), from the output of the trained model, classification data 369. Classification data 369 may classify defects of a substrate. Classification data 369 may assign a category to a pattern of defects of a substrate…Classification data 369 may be correlated to and/or include predicted root causes of defects of a substrate” [0106]);
receiving user input selecting a subset of the one or more matches (“A user (e.g., a subject matter expert) may be presented with an image visualizing locations of defects of the substrate. The user may classify a pattern of the defects.” [0023]);
updating the trained machine learning model based on the user input to produce an updated trained machine learning model (“In some aspects of the disclosure, a method includes providing, as training input to a machine learning model, a first plurality of images. [0005]; Selected model 308 may be retrained based on this data. Retraining may generate an updated machine learning model. Retraining may include providing retraining input and/or retraining target output to selected model 308.” [0107]);
outputs one or more updated matches to historical defects (“Data indicative of defect locations may be provided to a trained machine learning model as input. The trained machine learning model may generate, as output, a classification of a pattern of defects of the substrate.” [0017]); and
outputting an indication of one or more maintenance operations associated with the one or more updated matches to resolve one or more root causes associated with at least one of the one or more updated matches (“A library may exist that correlates patterns of defects on a substrate to possible causes, possible corrective actions, appropriate maintenance to perform, or the like.” [0022]).
Chowdhury does not teach a processing system that performs operations comprising score values for matches, reprocessing the first information using the updated trained machine learning model, and updated score values.
However, it is known that Balasubramanian teaches a process system (figs.1-4) comprising score values for matches (“using the retrained anomaly detection machine learning model, wherein an indication of an anomaly and a score for the anomaly are provided”, claim 13), reprocessing the first information using the updated trained machine learning model (“The selected model 2521 is then applied to unlabeled data 2502 compute sample importance scores, ranking information, and/or predictions for the unlabeled set which are stored in a history 2507. A feedback mechanism 2509 from a user on selected samples from the unlabeled data and augments the training set for the next iteration based on the feedback” [0157]), and updated scores (“using the retrained anomaly detection machine learning model, wherein an indication of an anomaly and a score for the anomaly are provided”, claim 13).
With respect to claim 16, the previously cited reference(s) teach the limitations of claim 15 which claim 16 depends on. Chowdhury teaches the non-transitory machine-readable storage medium of claim 15, wherein the processing device is to perform operations comprising:
retrieving, from a data structure, second information associated with the one or more matches to historical defects, wherein retrieving of the second information is based on the first information, and wherein the second information is indicative of the one or more maintenance operations (“In some embodiments, a second machine learning model may be used to determine a corrective actions. For example, a series of substrate defect pattern classifications may be provided to a trained machine learning model, and the trained machine learning model may generate as output one or more recommended corrective actions.” [0127]; “the first data indicative of locations of defects of the first substrate comprises an image, wherein the image indicates defect locations of the first substrate; and the second data indicative of locations of the second substrate comprises a list of locations of defects of the second substrate”, claim 10).
With respect to claim 17, the previously cited reference(s) teach the limitations of claim 15 which claim 17 depends on. Balasubramanian teaches the non-transitory machine-readable storage medium of claim 15, wherein updating the trained machine learning model based on the user input comprises:
providing the user input to the trained machine learning model as training data, wherein the trained machine learning model is further trained based on the user input (“At circle 3, an initial training of the models 103 is performed using the training dataset and unlabeled data from the testing dataset is subjected to the trained models 103. At circle 4, a request is sent to a feedback device 145 (e.g., a user providing feedback via this device) to evaluate the results of the trained models 103 on the unlabeled testing dataset and feedback is received by the anomaly detection service/component 102.” [0053]; “The feedback training 109 is engaged based on the response to update the models 103” [0054]).
With respect to claim 18, the previously cited reference(s) teach the limitations of claim 15 which claim 18 depends on. Chowdhury and Balasubramanian combined teaches the non-transitory machine-readable storage medium of claim 15, wherein the trained machine learning model comprises:
a first portion to output the one or more matches to historical defects based on the first information (Chowdhury: “The method further includes providing the image indicating the locations of the defects to a trained machine learning model. The method further includes obtaining, as output from the trained machine learning model, a classification of the locations of the defects”, abstract); and
a second portion to assign the score values, based at least in part on the user input, for the one or more matches to historical defects (Balasubramanian: “The model 103 is applied to the source data and generates a result which is provided to a console 106 to present to a user 143. Examples of how a result may be presented are detailed later, but may include, but are not limited to: the image evaluated, a score for the image” [0057]).
With respect to claim 19, the previously cited reference(s) teach the limitations of claim 15 which claim 19 depends on. Chowdhury and Balasubramanian combined teaches the non-transitory machine-readable storage medium of claim 15, wherein the processing device is to perform operations further comprising:
ranking the one or more matches to historical defects based on the score values for the matches (Balasubramanian: “An importance ranking score is computed for each item of the unlabeled dataset based on the result of applying the trained model at 2404. The ranking of the training set items is made in terms of how informative (when re-annotated) those items might be.” [0150]; “In some embodiments, a user can select how ranking scores are used to order (e.g., higher ranking are shown first, etc.).” [0153]); and
providing information associated with the one or more matches to historical defects for display on a graphical user interface (GUI) (Chowdhury: “Predictive component 114 may receive/obtain data from model 190. Predictive component 114 may use data obtained from model 190 to generate predictive data 168. In some embodiments, predictive component 114 provides predictive data 168 to client device 120, and client device 120 causes a corrective action via corrective action component 122 in view of predictive data 168. A corrective action performed by client device 120 may include displaying predictive data 168 for a user.” [0043]; Chowdhury: “Corrective action component 122 may receive user input (e.g., via a Graphical User Interface (GUI) displayed via the client device 120)” [0042]) in a ranked order based on the ranking, wherein the user input comprises a ranking of one or more images associated with the matches to historical defects (Balasubramanian: “This GUI shows a plurality of images that are sorted by number of objects first and then ranking. In some embodiments, a user may select individual ones of the images to evaluate under the detailed view” [0104]), and wherein the score values are determined based at least in part on the ranking (Balasubramanian: “An importance ranking score is computed for each item of the unlabeled dataset based on the result of applying the trained model at 2404. The ranking of the training set items is made in terms of how informative (when re-annotated) those items might be.” [0150]; “In some embodiments, a user can select how ranking scores are used to order (e.g., higher ranking are shown first, etc.).” [0153]).
With respect to claim 20, the previously cited reference(s) teach the limitations of claim 19 which claim 20 depends on. Balasubramanian teaches the non-transitory machine-readable storage medium of claim 19, wherein the user input comprises a ranking of one or more images associated with the matches to historical defects (“This GUI shows a plurality of images that are sorted by number of objects first and then ranking. In some embodiments, a user may select individual ones of the images to evaluate under the detailed view” [0104]), and wherein the score values are determined based at least in part on the ranking (“An importance ranking score is computed for each item of the unlabeled dataset based on the result of applying the trained model at 2404. The ranking of the training set items is made in terms of how informative (when re-annotated) those items might be.” [0150]; “In some embodiments, a user can select how ranking scores are used to order (e.g., higher ranking are shown first, etc.).” [0153]).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to NGUYEN P NGUYEN whose telephone number is (571)270-7351. The examiner can normally be reached Monday-Thursday 6:30-15:30, Friday 6:30-10:30.
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, KENNETH LO, can be reached at (571) 272-9774. 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.
/NGUYEN PHUONG NGUYEN/Examiner, Art Unit 2116
/CHAD G ERDMAN/Primary Examiner, Art Unit 2116
