DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
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.
Claim(s) 1, 3, 10-12, 14-15, 17 and 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Zhang et al (US 20230229906 A1), hereinafter Zhang in view of Hu et al (US 11406053 B1), hereinafter Hu.
-Regarding claim 1, Zhang discloses a computer-implemented method for performing causal discovery and generating an outcome of a treatment using one or more deep causal machine-learning models, comprising (Abstract; FIGS. 1-8): receiving a causal query for a causal outcome based on a treatment variable and a covariate variable (FIG. 1; [0025], “a plurality of variables … target variable … other variables (including intervened values and/or conditioned values) … The API 110 returns the result of the requested causal query (the estimated treatment effect” …”; FIG. 2; [0086], “a specific CATE query is received … treatment specified in the query … by inputting the relevant value of the conditioning variable …”; [0166]; FIGS. 3, 5); receiving, from a large generative model, a selection of a deep causal machine-learning model that is generated based on variable types associated with the treatment variable, the covariate variable, and the causal outcome (FIGS. 1-2, 6-8; [0006]-[0007]; [0045]; [0047]; [0050]-[0052]; [0151], “wherein the ML model is trained to be able to generate a respective simulated value of a selected variable … the target variable … intervened-on variable … graph distribution … determining an expectation … giving the estimated treatment effect”; [0152]; [0154] ); generating an embedding for the covariate variable from an image provided as input to the deep causal machine-learning model (FIG. 2; [0032], “generate a respective embedding ..”; [0033], “image”; [0042]-[0044]; FIG. 7; [0047]; FIG. 8; [0060]; [0152]); generating the causal outcome using the deep causal machine-learning model based on the treatment variable and the embedding (FIGS. 2, 4; FIGS. 6-8, [0056]; [0069]; [[0081]; 0154]); and providing a response based on the causal outcome in response to the causal query (FIG. 1; [0025], “In response … returns the result of the requested causal query (the estimated treatment effect) to the client computer 114 via the network 112”; [0168]).
Zhang does not disclose that the cause query is an agriculture-based causal query and the input image is an overhead agricultural image. A person of ordinary skill in the art would understand that Zhang has no limitation on type of causal query and image data. Zhang’s method can be applied to many areas including agriculture.
In the same field of endeavor, Hu teaches a method applying machine learning models to agricultural field data and historical data to derive representations of causality of one or more agronomic processes pertaining to the field (Hu: Abstract; FIGS. 1-10). Hu further teaches an agriculture-based causal query (Hu: FIG. 1, data management layer 140, repository; Col. 7, line 50, “including queries”; Col. 29, line 61, “query weather information …”) and overhead agricultural image (Hu: Col. 5, lines 49-52, “imagery data (for example, imagery and light spectrum information from an agricultural apparatus sensor, camera, computer, smartphone, tablet, unmanned aerial vehicle, planes or satellite)”; Col. 15, line 5, “satellite images”).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Zhang with the teaching of Hu by using agriculture-based causal query and overhead agricultural image in order to provide the method for agricultural applications.
-Regarding claim 3, Zhang in view of Hu teaches the method of claim 1.
Zhang discloses determining an image as a model input based on the deep causal machine-learning model selected by the large generative model (FIGS. 1-2; [0033]). Zhang does not disclose that the input image is an overhead agricultural image. A person of ordinary skill in the art would understand that Zhang has no limitation on type of causal query and image data. Zhang’s method can be applied to many areas including agriculture.
In the same field of endeavor, Hu teaches a method applying machine learning models to agricultural field data and historical data to derive representations of causality of one or more agronomic processes pertaining to the field (Hu: Abstract; FIGS. 1-10). Hu further teaches an agriculture-based causal query and determining an overhead agricultural image as a model input based on the deep causal machine-learning model selected by the large generative model (Hu: FIG. 1, data management layer 140, field data 104, repository 160; FIGS. 3, 5; Col. 7, line 50, “including queries”; Col. 29, line 61, “query weather information …”) and overhead agricultural image (Hu: Col. 5, lines 49-52, “imagery data (for example, imagery and light spectrum information from an agricultural apparatus sensor, camera, computer, smartphone, tablet, unmanned aerial vehicle, planes or satellite)”; Col. 15, line 5, “satellite images”).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Zhang with the teaching of Hu by using agriculture-based causal query and overhead agricultural image in order to provide the method for agricultural applications.
-Regarding claim 10, , Zhang in view of Hu teaches the method of claim 1. The combination further teaches wherein the deep causal machine-learning model selected by the large generative model is trained to determine the causal outcome based on multiple treatment variables or multiple covariate variables (Zhang: Abstract; [0007]; [0009], “estimate a treatment effect from one or more intervened-on variables on another … among the variables of said set”; FIG. 1; [0025]; FIG. 2).
-Regarding claim 11, Zhang in view of Hu teaches the method of claim 1.
Zhang does not disclose generating multiple causal outcomes.
In the same field of endeavor, Hu teaches a method applying machine learning models to agricultural field data and historical data to derive representations of causality of one or more agronomic processes pertaining to the field (Hu: Abstract; FIGS. 1-10). Hu further teaches generating multiple causal outcomes (Hu: FIGS. 1-2, 6-10; Col. 15, lines 10-24, “enables the grower to seek improved outcomes for the next year through fact-based conclusions … treatment effect estimation … ”) and overhead agricultural image (Hu: Col. 5, lines 49-52, “imagery data (for example, imagery and light spectrum information from an agricultural apparatus sensor, camera, computer, smartphone, tablet, unmanned aerial vehicle, planes or satellite)”; Col. 15, line 5, “satellite images”).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Zhang with the teaching of Hu by using agriculture-based causal query and overhead agricultural image in order to provide the method for agricultural applications.
-Regarding claim 12, Zhang in view of Hu teaches the method of claim 1. The combination further teaches wherein the deep causal machine-learning model selected by the large generative model is based on multiple covariate variables, treatment variables, or causal outcomes (Zhang: Abstract; [0007]; [0009], “estimate a treatment effect from one or more intervened-on variables on another … among the variables of said set”; FIG. 1; [0025]; FIG. 2) and the multiple covariate variables, treatment variables, or causal outcomes used by the deep causal machine-learning model are based on one or more hyperparameters provided to the deep causal machine-learning model (Zhang: FIGS. 2, 6-8; [0040], “comprises a selector custom-character”; [0042]; [0047]).
-Regarding claim 14, Zhang in view of Hu teaches the method of claim 1.
Zhang does not disclose wherein the input image is an overhead agricultural image which is a satellite image or includes remote sensing data. A person of ordinary skill in the art would understand that Zhang has no limitation on type of causal query and image data. Zhang’s method can be applied to many areas including agriculture.
In the same field of endeavor, Hu teaches a method applying machine learning models to agricultural field data and historical data to derive representations of causality of one or more agronomic processes pertaining to the field (Hu: Abstract; FIGS. 1-10). Hu further teaches an agriculture-based causal query (Hu: FIG. 1, data management layer 140, repository; Col. 7, line 50, “including queries”; Col. 29, line 61, “query weather information …”) and wherein the overhead agricultural image is a satellite image or includes remote sensing data (Hu: Col. 5, lines 49-52, “imagery data (for example, imagery and light spectrum information from an agricultural apparatus sensor, camera, computer, smartphone, tablet, unmanned aerial vehicle, planes or satellite)”; Col. 15, line 5, “satellite images”).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Zhang with the teaching of Hu by using agriculture-based causal query and overhead agricultural image in order to provide the method for agricultural applications.
-Regarding claim 15, Zhang discloses a computer-implemented method for generating a causal outcome using one or more deep causal machine-learning models, comprising (Abstract; FIGS. 1-8): generating an embedding for a covariate variable from an overhead agricultural image provided as input to a deep causal machine-learning model (FIG. 2; [0032], “generate a respective embedding ..”; [0033], “image”; [0042]-[0044]; FIG. 7; [0047]; FIG. 8; [0060]; [0152]); the deep causal machine-learning model generated is based on variable types associated with a treatment variable, the covariate variable, and a causal outcome (FIGS. 1-2, 6-8; [0006]-[0007]; [0045]; [0047]; [0050]-[0052]; [0151], “wherein the ML model is trained to be able to generate a respective simulated value of a selected variable … the target variable … intervened-on variable … graph distribution … determining an expectation … giving the estimated treatment effect”; [0152]; [0154] ); and; generating the causal outcome using the deep causal machine-learning model based on the treatment variable and the embedding (FIGS. 2, 4; FIGS. 6-8, [0056]; [0069]; [[0081]; 0154]); and providing a response based on the causal outcome in response to a causal query (FIG. 1; [0025], “In response … returns the result of the requested causal query (the estimated treatment effect) to the client computer 114 via the network 112”; [0168]) for the causal outcome based on the treatment variable and the covariate variable (FIG. 1; [0025], “a plurality of variables … target variable … other variables (including intervened values and/or conditioned values) … The API 110 returns the result of the requested causal query (the estimated treatment effect” …”; FIG. 2; [0086], “a specific CATE query is received … treatment specified in the query … by inputting the relevant value of the conditioning variable …”; [0166]; FIGS. 3, 5).
Zhang does not disclose that the cause query is an agriculture-based causal query and the input image is an overhead agricultural image. A person of ordinary skill in the art would understand that Zhang has no limitation on type of causal query and image data. Zhang’s method can be applied to many areas including agriculture.
In the same field of endeavor, Hu teaches a method applying machine learning models to agricultural field data and historical data to derive representations of causality of one or more agronomic processes pertaining to the field (Hu: Abstract; FIGS. 1-10). Hu further teaches an agriculture-based causal query (Hu: FIG. 1, data management layer 140, repository; Col. 7, line 50, “including queries”; Col. 29, line 61, “query weather information …”) and overhead agricultural image (Hu: Col. 5, lines 49-52, “imagery data (for example, imagery and light spectrum information from an agricultural apparatus sensor, camera, computer, smartphone, tablet, unmanned aerial vehicle, planes or satellite)”; Col. 15, line 5, “satellite images”).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Zhang with the teaching of Hu by using agriculture-based causal query and overhead agricultural image in order to provide the method for agricultural applications.
-Regarding claim 17, Zhang in view of Hu teaches the method of claim 15.
Zhang in view of Hu teaches training an embedding model that generates the embedding for the covariate variable from an image jointly with training the deep causal machine-learning model (Hu: FIG. 2; [0032], “generate a respective embedding …”; [0033], “image”; [0042]-[0044]; FIG. 7; [0047]; FIG. 8; [0060]; [0152]).
Zhang does not disclose that the cause query is an agriculture-based causal query and the input image is an overhead agricultural image. A person of ordinary skill in the art would understand that Zhang has no limitation on type of causal query and image data. Zhang’s method can be applied to many areas including agriculture.
In the same field of endeavor, Hu teaches a method applying machine learning models to agricultural field data and historical data to derive representations of causality of one or more agronomic processes pertaining to the field (Hu: Abstract; FIGS. 1-10). Hu further teaches an agriculture-based causal query (Hu: FIG. 1, data management layer 140, repository; Col. 7, line 50, “including queries”; Col. 29, line 61, “query weather information …”) and overhead agricultural image (Hu: Col. 5, lines 49-52, “imagery data (for example, imagery and light spectrum information from an agricultural apparatus sensor, camera, computer, smartphone, tablet, unmanned aerial vehicle, planes or satellite)”; Col. 15, line 5, “satellite images”).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Zhang with the teaching of Hu by using agriculture-based causal query and overhead agricultural image in order to provide the method for agricultural applications.
-Regarding claim 20, Zhang discloses a system for generating a causal outcome using one or more deep causal machine-learning models, comprising (Abstract; FIGS. 1-8): a processing system; and a computer memory comprising instructions that, when executed by the processing system, cause the system to perform operations of (FIG. 1; [0022]; [0169]-[0170]): receiving a causal query for a causal outcome based on a treatment variable and a covariate variable (FIG. 1; [0025], “a plurality of variables … target variable … other variables (including intervened values and/or conditioned values) … The API 110 returns the result of the requested causal query (the estimated treatment effect” …”; FIG. 2; [0086], “a specific CATE query is received … treatment specified in the query … by inputting the relevant value of the conditioning variable …”; [0166]; FIGS. 3, 5); receiving, from a large generative model, a selection of a deep causal machine-learning model that is generated based on variable types associated with the treatment variable, the covariate variable, and the causal outcome (FIGS. 1-2, 6-8; [0006]-[0007]; [0045]; [0047]; [0050]-[0052]; [0151], “wherein the ML model is trained to be able to generate a respective simulated value of a selected variable … the target variable … intervened-on variable … graph distribution … determining an expectation … giving the estimated treatment effect”; [0152]; [0154] ); generating an embedding for the covariate variable from an image provided as input to the deep causal machine-learning model (FIG. 2; [0032], “generate a respective embedding ..”; [0033], “image”; [0042]-[0044]; FIG. 7; [0047]; FIG. 8; [0060]; [0152]); generating the causal outcome using the deep causal machine-learning model based on the treatment variable and the embedding (FIGS. 2, 4; FIGS. 6-8, [0056]; [0069]; [[0081]; 0154]); and providing a response based on the causal outcome in response to the causal query (FIG. 1; [0025], “In response … returns the result of the requested causal query (the estimated treatment effect) to the client computer 114 via the network 112”; [0168]).
Zhang does not disclose that the cause query is an agriculture-based causal query and the input image is an overhead agricultural image. A person of ordinary skill in the art would understand that Zhang has no limitation on type of causal query and image data. Zhang’s method can be applied to many areas including agriculture.
In the same field of endeavor, Hu teaches a method applying machine learning models to agricultural field data and historical data to derive representations of causality of one or more agronomic processes pertaining to the field (Hu: Abstract; FIGS. 1-10). Hu further teaches an agriculture-based causal query (Hu: FIG. 1, data management layer 140, repository; Col. 7, line 50, “including queries”; Col. 29, line 61, “query weather information …”) and overhead agricultural image (Hu: Col. 5, lines 49-52, “imagery data (for example, imagery and light spectrum information from an agricultural apparatus sensor, camera, computer, smartphone, tablet, unmanned aerial vehicle, planes or satellite)”; Col. 15, line 5, “satellite images”).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Zhang with the teaching of Hu by using agriculture-based causal query and overhead agricultural image in order to provide the method for agricultural applications.
Claim(s) 2 and 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Zhang et al (US 20230229906 A1), hereinafter Zhang in view of Hu et al (US 11406053 B1), hereinafter Hu, and further in view of Muthusamy et al (US 20250181989 A1), hereinafter Muthusamy.
-Regarding claims 2 and 18, Zhang in view of Hu teaches the method of claim 1 and the method of claim 15.
Zhang in view of Hu does not teach providing a query with a prompt to select the deep causal machine-learning model. A person of ordinary skill in the art would understand that this is a common practice in the machine learning filed.
However, Muthusamy is an analogous art pertinent to the problem to be solved in this application and teaches a method selecting from the relevant training data partitioned by the set of groups in response to receiving a query from a user and using the top relevant training samples for the user to generate a prompt of a machine learning model (Muthusamy: Abstract; FIGS. 1-8). Muthusamy further teaches providing a query with a prompt to select the deep causal machine-learning model (Muthusamy: FIGS. 1, 4A-4B, 7; [0002]; [0004]; [0019]; [0024]-[0025]; [0047]).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to modify the teaching of Zhang in view of Hu with the teaching of Muthusamy by providing a query with a prompt to select the deep causal machine-learning model in order to fast and accurate determine a machine learning model for causal outcome.
Claim(s) 4-6 is/are rejected under 35 U.S.C. 103 as being unpatentable over Zhang et al (US 20230229906 A1), hereinafter Zhang in view of Hu et al (US 11406053 B1), hereinafter Hu, and further in view of Ferreira Moreno et al (US 20230016157 A1), hereinafter Ferreira.
-Regarding claims 4, Zhang in view of Hu teaches the method of claim 1.
Zhang in view of Hu does not teach that the selected model is a multimodal deep machine-learning model.
However, Ferreira is an analogous art pertinent to the problem to be solved in this application and teaches a method for mapping and combining the application of machine learning models to answer queries according to semantic specification (Ferreira: Abstract; FIGS. 1-9). Ferreira further teaches that the selected model is a multimodal deep machine-learning model (Ferreira: Abstract; [0017]; [0022], “process queries to extract keywords and meaning, and map to a contextualization of machine learning models; select, sort and apply selected machine learning models on selected multimodal data and conceptual symbolic representation; dynamically index multimodal data fragments and machine learning models in query time”; FIG. 5; [0062]-[0063]; FIG.2; [0037]).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to modify the teaching of Zhang in view of Hu with the teaching of Ferreira by selecting a multimodal deep causal machine-learning model in order to input/output digital content that includes a combination of two or more of text, images, video or audio to answer the causal query.
-Regarding claims 5, Zhang in view of Hu, and further in view Ferreira teaches the method of claim 4.
Zhang in view of Hu does not teach unstructured data. A person of ordinary skill in the art would understand that text, images, videos are unstructured data.
However, Ferreira is an analogous art pertinent to the problem to be solved in this application and teaches a method for mapping and combining the application of machine learning models to answer queries according to semantic specification (Ferreira: Abstract; FIGS. 1-9). Ferreira further teaches providing unstructured data (Ferreira: FIG. 1; [0028], “by running one or more selected machine learning models on unstructured data”).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to modify the teaching of Zhang in view of Hu with the teaching of Ferreira by selecting a multimodal deep causal machine-learning model and using unstructured data in order to input/output digital content that includes a combination of two or more of text, images, video or audio to answer the causal query.
-Regarding claims 6, Zhang in view of Hu, and further in view Ferreira teaches the method of claim 4.
Zhang does not disclose providing structured data.
In the same field of endeavor, Hu teaches a method applying machine learning models to agricultural field data and historical data to derive representations of causality of one or more agronomic processes pertaining to the field (Hu: Abstract; FIGS. 1-10). Hu further teaches an agriculture-based causal query (Hu: FIG. 1, data management layer 140, repository; Col. 7, line 50, “including queries”; Col. 29, line 61, “query weather information …”) and providing structured agriculture data with the overhead agricultural image (Hu: FIG. 1; Col. 5, lines 49-52, “imagery data (for example, imagery and light spectrum information from an agricultural apparatus sensor, camera, computer, smartphone, tablet, unmanned aerial vehicle, planes or satellite)”; Col. 7, lines 52-62, “Repository 160 may comprise a database … any other structured collection of records or data”; Col. 15, line 5, “satellite images”).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Zhang with the teaching of Hu by using agriculture-based causal query and overhead agricultural image in order to provide the method for agricultural applications.
Claim(s) 13 and 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Zhang et al (US 20230229906 A1), hereinafter Zhang in view of Hu et al (US 11406053 B1), hereinafter Hu, and further in view of Baldua et al (US 12298975 B2), hereinafter Baldua.
-Regarding claims 13, Zhang in view of Hu teaches the method of claim 15.
Zhang in view of Hu does not teach generating a plain language answer prompt for the large generative model to generate a plain language response based on the causal outcome and the agriculture-based causal query, and providing the plain language response in response to the agriculture-based causal query.
However, Baldua is an analogous art pertinent to the problem to be solved in this application and teaches a method for configuring at least one prompt to cause a large language model to translate a received query term into a set of functions (Baldua: Abstract; FIGS. 1-8). Baldua further teaches generating a plain language answer prompt for the large generative model (Baldua: Abstract; FIGS. 1A-3, 7A-7B; Col. 4, lines 26-45, “dynamically configure a prompt to include instructions to cause one or more generative artificial intelligence models (e.g., one or more large language models) to generate and output a plan for executing a query …”) and to generate a plain language response based on the causal outcome and the agriculture-based causal query, and providing the plain language response in response to the agriculture-based causal query (Baldua: FIGS. 1A-3, 7A-7B; Col. 4, lines 53-55, “A generative language model is a particular type of GAI model that is capable of generating new text in response to model input”).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to modify the teaching of Zhang in view of Hu with the teaching of Baldua by providing a query with a prompt to select the deep causal machine-learning model in order to fast and accurate determine a machine learning model for causal outcome, and selecting a multimodal causal deep machine-learning model in order to input/output digital content that includes a combination of two or more of text, images, video or audio to answer the causal query.
-Regarding claims 19, Zhang in view of Hu teaches the method of claim 15.
Zhang in view of Hu does not teach providing a query with a prompt to select the deep causal machine-learning model. Zhang in view of Hu does not teach that the selected model is a multimodal deep machine-learning model.
However, Baldua is an analogous art pertinent to the problem to be solved in this application and teaches a method for configuring at least one prompt to cause a large language model to translate a received query term into a set of functions (Baldua: Abstract; FIGS. 1-8). Baldua further teaches providing a query with a prompt to select the deep causal machine-learning model (Baldua: Abstract; FIGS. 1A-3, 7A-7B; Col. 4, lines 26-45, “dynamically configure a prompt to include instructions to cause one or more generative artificial intelligence models (e.g., one or more large language models) to generate and output a plan for executing a query …”) and teaches that the selected model is a multimodal deep causal machine-learning model (Baldua: Col. 13, lines 40-64, “one or more types of neural network-based machine learning model architectures include or are based on one or more multimodal neural networks capable of outputting different modalities … a multimodal neural network implemented in the dynamic query …”).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to modify the teaching of Zhang in view of Hu with the teaching of Baldua by providing a query with a prompt to select the deep causal machine-learning model in order to fast and accurate determine a machine learning model for causal outcome, and selecting a multimodal causal deep machine-learning model in order to input/output digital content that includes a combination of two or more of text, images, video or audio to answer the causal query.
Allowable Subject Matter
Claims 7-9 and 16 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
Response to Arguments
Applicant's arguments filed on 02/23/2026 have been fully considered but they are not persuasive. Applicant argues that Zhang does not describe, teach, or suggest "receiving, from a large generative model, a selection of a deep causal machine-learning model that is generated based on variable types associated with the treatment variable, the covariate variable, and the causal outcome." as recited by currently amended independent claim 1 (Remarks: page 9, last paragraph) because Zhang does not teach or suggest “(i) receiving a selection of a deep causal machine- learning model from a large generative model, or (ii) that the large generative model selects the deep causal machine-learning model based on variable types associated with the treatment variable, the covariate variable, and the causal outcome, as recited by the independent claims” (Remarks: page 10, 3rd paragraph). The examiner respectfully disagrees with the above arguments.
In response to applicant's argument that Zhang does not teach or suggest “receiving a selection of a deep causal machine-learning model from a large generative model”, please note that there is no formal definition about “large generative model”. The applicant’s specification only describes that a "large generative model" (LGM) is a large artificial intelligence system that uses deep learning and a large number of parameters ([0027]) and LGM can also use a recurrent neural network (RNN) architecture, a long short-term memory (LSTM) model architecture, a convolutional neural network (CNN) architecture, or another architecture type ([0029]). Thus, any deep leaning machine learning model can be considered as a "large generative model” according to the description of the specification. Zhang discloses a deep machine learning (ML) architecture (Zhang: FIGS. 1-2, 6-8
PNG
media_image1.png
421
525
media_image1.png
Greyscale
) for treatment effect estimation. The ML model 104 comprises a plurality of encoders
g
e
i
and decoders
g
d
i
for a plurality of variables
x
i
. The ML model 104 further comprises a selector 1, a combiner 202 and a demultiplexer 204. The selector also receives a value of the index i for a selected target variable
x
i
. For the currently selected variable
x
i
, the respective embeddings
e
P
a
(
i
)
are generated by the respective encoders
g
e
P
a
(
i
)
of the parents Pa(i) of the node i (variable
x
i
) in the currently sampled graph G. The combiner 202 combines the selected embeddings
e
P
a
(
i
)
into a combined embedding
e
c
. Thus, the ML model 104 is a combination of a plurality of respective encoders and decoders, and each of the encoders
g
e
i
and
g
d
i
is a respective neural network (Zhang: [0045]). However, the final ML model for treatment effect estimation is based the selected target variables which determine the corresponding encoders and decoders. For different selected variables, the corresponding encoders and decoders are selected. Therefore, different ML models are selected. The target variable can be either specified from a request or generated given the intervened and/or conditioned values of the one or more other variables (Zhang: [0025]; [0151]).
In response to applicant's argument that Zhang does not teach or suggest “large generative model selects the deep causal machine-learning model based on variable types associated with the treatment variable, the covariate variable, and the causal outcome”, Zhang discloses receiving, from a large generative model, a selection of a deep causal machine-learning model that is generated based on variable types associated with the treatment variable, the covariate variable, and the causal outcome (FIGS. 1-2, 6-8; [0006]-[0007]; [0045]; [0047]; [0050]-[0052]; [0151], “wherein the ML model is trained to be able to generate a respective simulated value of a selected variable from among said set based on the sampled causal graph … using the ML model to estimate a treatment effect from one or more intervened-on variables on another, target variable from among the variables of said set …intervened-on variable … sampling graph distribution … determining an expectation … giving the estimated treatment effect”; [0152]; [0154], “comprises an inference network disposed between an unobserved one of the variables of said set and one or more observable ones of the variables of said set”; see also this office action, pages 2-3 or Non-Final Rejection: page 3). As discussed above, the selected ML model is based on the selected variables associated with intervened and/or conditioned values of the one or more other variable, and casual graph. The machine learning model can be designed to perform causal discovery using observational data or both observational and interventional data and a causal graph describes the causal relationships between these variables (Zhang: [0007]). An example of casual graph is shown in Zhang’s FIG. 5 which is associated with treatment variable, observation variable, observed effect variable, unobserved cause variable, and effect variable.
Conclusion
THIS ACTION IS MADE FINAL. 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 XIAO LIU whose telephone number is (571)272-4539. The examiner can normally be reached Monday-Thursday and Alternate Fridays 8:30-4:30.
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, Jennifer Mehmood can be reached at (571) 272-2976. 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.
/XIAO LIU/Primary Examiner, Art Unit 2664
Prosecution Insights