CTNF 18/471,570 CTNF 95574 Notice of Pre-AIA or AIA Status 07-03-aia AIA 15-10-aia The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA. Claim Objections 07-29-01 AIA Claim s 1-2 and 4 are objected to because of the following informalities: In claim 1, line 5, “obtaining” should recite “obtain”. The first term in each remaining limitation recites the same minor informality. In claim 1, line 9, “each of dynamic causal graphs” should recite “each of the dynamic causal graphs”, and the term “Acrylic” should recite “Acyclic”. Claim 1, line 3 recites “at least one processor configured” and claim 2, line 2 recites “the one or more processors are configured”. Examiner suggests changing the limitation in claim 2 to “the at least one processor is configured” to keep the limitation consistent. Claim 4, lines 3-4 recites “by machine-learning algorithm.” This limitation is not grammatically correct. Examiner suggests amending this limitation to recite “by a machine-learning algorithm.” Appropriate correction is required. Claim Rejections - 35 USC § 112 07-30-02 AIA The following is a quotation of 35 U.S.C. 112(b): (b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention. 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. 07-34-01 Claims 1-7 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA), second paragraph, as being indefinite for failing to particularly point out and distinctly claim 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. In claim 1, the limitations “states” in line 7 and “each environment state” in line 8 render the claim indefinite because it is unclear whether states and environment states are the same or different. Examiner treats “states” in line 7 as “environment states”. Claims 2-7 are rejected for failing to cure the deficiencies of claim 1. 07-34-03 AIA The term “ similar ” in claim 6, line 4 is a relative term which renders the claim indefinite. The term “ similar ” is not defined by the claim, the specification does not provide a standard for ascertaining the requisite degree, and one of ordinary skill in the art would not be reasonably apprised of the scope of the invention. It is unclear what makes two time steps similar to each other, it is unclear what makes two time steps dissimilar from each other, and it is unclear what the boundary is between similar time steps and dissimilar time steps. Examiner treats “similar time steps” as “time steps” . In claim 7, the limitation “the causal graph” recited in line 2 renders the claim indefinite. It is unclear if this causal graph is one of the “dynamic causal graphs” recited in claim 1, line 8. Examiner treats “the causal graph is generated by optimizing the causal graph” as “each of the dynamic causal graphs is generated by optimizing the dynamic causal graph” . Claim Rejections - 35 USC § 101 07-04-01 AIA 07-04 35 U.S.C. 101 reads as follows: Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title. Claims 1-7 rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. Step 1: Claims 1-7 each recites a system, which is one of the four statutory categories of patent eligible subject matter. Claim 1 Step 2A Prong 1: Training the model by computing actions to be taken with respect to states is a mathematical calculation. Specification paragraphs [0042]-[0043] discloses computing actions, and [0070]-[0073] discloses a learning objective for a policy, which the model generates. Generating dynamic causal graphs for each environment state, wherein each of dynamic causal graphs is a Directed Acrylic Graph is a judgement mental process which can reasonably be performed in the human mind with the aid of pencil and paper. Specification paragraph [0027] states that Fig. 1 depicts a causal graph in the top right section of the figure. A person can draw the causal graph on a piece of paper. Encoding discovered causal relationships by updating state variable embeddings is a mathematical calculation. Specification paragraphs [0053]-[0055] and [0062]-[0065] include equations (4) and (8) which encode a trajectory of states and generates an embedding of state variables. The claim recites an abstract idea. Step 2A Prong 2: At least one memory storing instructions, and at least one processor configured to access the at least one memory and execute the instructions amount to generic computer components for applying the abstract ideas on a generic computer under MPEP 2106.05(f). Obtaining demonstrations of a target task from experts for training a model to generate a learned policy amounts to insignificant pre-solution activity under MPEP 2106.05(g). Outputting the learned policy amounts to insignificant post-solution activity under MPEP 2106.05(g). The additional elements as disclosed above, alone or in combination, do not integrate the abstract ideas into a practical application as they are mere insignificant extra solution activities as disclosed in combination with generic computer functions that are implemented to perform the abstract ideas disclosed above. The claim is directed to an abstract idea. Step 2B: At least one memory storing instructions, and at least one processor configured to access the at least one memory and execute the instructions amount to generic computer components for applying the abstract ideas on a generic computer under MPEP 2106.05(f). Obtaining demonstrations of a target task from experts for training a model to generate a learned policy is analogous to retrieving information from memory, which is a well-understood, routine, convention activity recognized by the courts under MPEP 2106.05(d)(II). Outputting the learned policy is analogous to presenting offers and gathering statistics, which is a well-understood, routine, convention activity recognized by the courts under MPEP 2106.05(d)(II). The additional elements as disclosed above, in combination with the abstract ideas, are not sufficient to amount to significantly more than the abstract ideas as they are well-understood, routine and conventional activities as disclosed in combination with generic computer functions that are implemented to perform the abstract ideas disclosed above. The claim is not patent eligible. Claim 2 incorporates the rejection of claim 1. Step 2A Prong 1: The abstract ideas of claim 1 are incorporated. Conduct an imitation learning task and a state regression task by employing the updated state variable embeddings as evidence are mathematical calculations. Specification paragraphs [0072]-[0073] disclose an imitation learning objective function, and specification paragraphs [0077]-[0079] disclose an objective function for state variables. Step 2A Prong 2 and Step 2B: The one or more processors amount to generic computer components for applying the abstract ideas on a generic computer under MPEP 2106.05(f). The claim is not patent eligible. Claim 3 incorporates the rejection of claim 2. Step 2A Prong 1: The abstract ideas of claim 2 are incorporated. The state regression task is used to provide auxiliary signals for learning causal edges among state variables is a mathematical calculation. Specification paragraphs [0077]-[0081] disclose applying the objective function L res from equation (12) to the final objective function in equation (13). Step 2A Prong 2 and Step 2B: The claim does not recite any additional elements which, alone or in combination, would integrate the abstract ideas into a practical application or which, in combination with the abstract ideas, would be sufficient to amount to significantly more than the abstract ideas. The claim is not patent eligible. Claim 4 incorporates the rejection of claim 2. Step 2A Prong 1: The abstract ideas of claim 2 are incorporated. The limitation “for the imitation learning task, the learned policy is adversarially trained with a discrimination model to discriminate between predicted actions and demonstrations from expert by machine-learning algorithm” is a mathematical calculation. Specification paragraphs [0070]-[0073] discloses an adversarial learning objective function for a policy. Step 2A Prong 2 and Step 2B: The claim does not recite any additional elements which, alone or in combination, would integrate the abstract ideas into a practical application or which, in combination with the abstract ideas, would be sufficient to amount to significantly more than the abstract ideas. The claim is not patent eligible. Claim 5 incorporates the rejection of claim 1. Step 2A Prong 1: The abstract ideas of claim 1 are incorporated. The state variable embeddings are updated with propagated messages from variables it depends on by employing an edge- aware update layer is a mathematical calculation. Specification paragraphs [0036], [0063]-[0068] discloses equations for updating the state variable embeddings. Step 2A Prong 2 and Step 2B: The claim does not recite any additional elements which, alone or in combination, would integrate the abstract ideas into a practical application or which, in combination with the abstract ideas, would be sufficient to amount to significantly more than the abstract ideas. The claim is not patent eligible. Claim 6 incorporates the rejection of claim 1. Step 2A Prong 1: The abstract ideas of claim 1 are incorporated. A sparsity constraint and an acyclicity constraint are employed to optimize the dynamic causal graphs, and a template selection regularization loss is employed to enable consistency in template selection across similar time steps are mathematical calculations. Specification paragraphs [0048]-[0052] and [0060]-[0061] disclose equations for the constraints and a template selection regularization loss. Step 2A Prong 2 and Step 2B: The claim does not recite any additional elements which, alone or in combination, would integrate the abstract ideas into a practical application or which, in combination with the abstract ideas, would be sufficient to amount to significantly more than the abstract ideas. The claim is not patent eligible. Claim 7 incorporates the rejection of claim 1. Step 2A Prong 1: The abstract ideas of claim 1 are incorporated. The causal graph is generated by optimizing the causal graph based on constraints is a mathematical calculation. Specification paragraphs [0048]-[0052] disclose equations for the constraints. Step 2A Prong 2 and Step 2B: The claim does not recite any additional elements which, alone or in combination, would integrate the abstract ideas into a practical application or which, in combination with the abstract ideas, would be sufficient to amount to significantly more than the abstract ideas. The claim is not patent eligible. Claim Rejections - 35 USC § 102 07-06 AIA 15-10-15 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. 07-07-aia AIA 07-07 The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action: A person shall be entitled to a patent unless – 07-08-aia AIA (a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention. 07-15 AIA Claim s 1-3 and 5 are rejected under 35 U.S.C. 102( a)(1 ) as being anticipated by Haan et al. (“Causal Confusion in Imitation Learning”) . Regarding claim 1, Haan teaches: A learning system comprising: at least one memory storing instructions; and at least one processor configured to access the at least one memory and execute the instructions to: (Page 2, § 3 to page 3, line 15 and Algorithm 1 on page 6 teaches computing causal graphs. This computing environment inherently requires a memory storing instructions and processor for executing the instructions.) obtaining demonstrations of a target task from experts for training a model to generate a learned policy; (Page 2, § 3, line 1 to page 3, line 2; Page 5, § 4.1, lines 1-16 and § 4.2, from line 1 to page 6, line 10. A “target task” is driving a car, a “model” is a single graph-parameterized policy PNG media_image1.png 64 651 media_image1.png Greyscale , Equation (1) discloses training, and “demonstrations of a target task from experts” includes expert response A in expert query mode.) training the model by computing actions to be taken with respect to states; (Page 5, § 4.1, lines 1-16 and § 4.2, from line 1 to page 6, line 10. Training is described above.) generating dynamic causal graphs for each environment state, wherein each of [the] dynamic causal graphs is a Directed [Acyclic] Graph; (Page 2, § 3 to page 3, line 15 and page 5, § 4.1, lines 1-8 and equation (1) discloses generating a number of possible graphs. The system performs element-wise multiplication between state variable vector X i and a binary parameterization of graph G. The causal graphs are “dynamic” because the state variable vector X i changes them via element-wise multiplication.) encoding discovered causal relationships by updating state variable embeddings; and (Page 5, § 4.1, lines 2-8 and equation (1) discloses performing element-wise multiplication between X i and G. This encodes the causal graph and updates state variable vector X i into an embedding.) outputting the learned policy. (Page 5, § 4.1, lines 7-16, where training the policy network f ϕ in equation (1) generates the learned policy.) Regarding claim 2, Haan teaches: The learning system according to claim 1, wherein the one or more processors are configured to further execute the instructions to: conduct an imitation learning task and a state regression task by employing the updated state variable embeddings as evidence. (Page 2, § 3, line 1 to page 3, line 2; Page 5, § 4.1, lines 1-16 and § 4.2, from line 1 to page 6, line 10. An imitation learning task corresponds to training the policy network f ϕ in expert query mode, and a state regression task corresponds to equation (1). This training employs the product of element-wise multiplication between X i and G as the input to the neural network, which is a type of evidence.) Regarding claim 3, Haan teaches: The learning system according to claim 2, wherein the state regression task is used to provide auxiliary signals for learning causal edges among state variables. (On page 5, equation (1) corresponds to “a state regression task” as claimed. Training neural network parameters of the policy network results in learning causes between states, so the training provides information (“auxiliary signals”) for learning causal edges.) Regarding claim 5, Haan teaches: The learning system according to claim 1, wherein the state variable embeddings are updated with propagated messages from variables it depends on by employing an edge-aware update layer. (Page 5, § 4.1, lines 2-8 and equation (1) discloses performing element-wise multiplication between X i and G. State variable embeddings correspond to a product of this multiplication, and “variables” as claimed are X i and G. An edge-aware update layer corresponds to an input layer of the policy network because the multiplication is part of an input to the policy network. The policy network maps observations X to actions A, and the relationship between X and A is an edge.) Claim Rejections - 35 USC § 103 07-20-aia AIA 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. 07-23-aia AIA 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. 07-20-02-aia AIA This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention. 07-21-aia AIA Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over Haan et al. (“Causal Confusion in Imitation Learning”) and Wang et al. (“Adversarial Cooperative Imitation Learning for Dynamic Treatment Regimes”) . Regarding claim 4, Haan teaches: The learning system according to claim 2, wherein, for the imitation learning task, the learned policy is adversarially trained with a discrimination model to discriminate between predicted actions and demonstrations from expert by machine-learning algorithm. (Page 2, § 3, line 1 to page 3, line 2; Page 5, § 4.1, lines 1-16 and § 4.2, from line 1 to page 6, line 10, and Algorithm 1 where a “machine-learning algorithm” corresponds to the line “Fit w on D with linear regression” in Algorithm 1.) However, Haan does not explicitly teach: the learned policy is adversarially trained with a discrimination model to discriminate between predicted actions and demonstrations from expert by machine-learning algorithm. But Wang teaches: the learned policy is adversarially trained with a discrimination model to discriminate between predicted actions and demonstrations from expert by machine-learning algorithm. (Page 1785, all of Abstract; and page 1788, § 3.1, lines 1-13 and 17-21 discloses a dynamic treatment regime policy π θ is trained with an adversarial discriminator D a , which discriminates between treatments predicted by the policy and real-world treatments. The “model” is the policy π θ , “demonstrations from expert” are real-world treatments prescribed by doctors, and a “machine-learning algorithm” is the training/updating of the policy π θ .) It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have incorporated Wang’s adversarial discriminator into Haan, and to have updated the model from Haan’s policy based on feedback from Wang’s adversarial discriminator. A motivation for the combination is to train the model to mimic successful treatments. (Page 1788, col. 1, § 3.1) 07-21-aia AIA Claim 6 is rejected under 35 U.S.C. 103 as being unpatentable over Haan et al. (“Causal Confusion in Imitation Learning”), Pamfil et al. (“DYNOTEARS: Structure Learning from Time-Series Data”), and Singer et al. (US 10438129 B1) . Regarding claim 6, Haan teaches: The learning system according to claim 1, wherein a sparsity constraint and an acyclicity constraint are employed to optimize the dynamic causal graphs, (Page 5, equation (1) teaches an objective function for training the policy network which takes causal graphs as input.) However, Haan does not explicitly teach: wherein a sparsity constraint and an acyclicity constraint are employed to optimize the dynamic causal graphs, and a template selection regularization loss is employed to enable consistency in template selection across similar time steps. But Pamfil teaches: wherein a sparsity constraint and an acyclicity constraint are employed to optimize the dynamic causal graphs, and (Page 1, § 1, from line 5 to col. 2, line 11, and Page 3, § 2.2, from line 1 to the fourth line below equation (5) discloses optimizing a dynamic causal graph.) It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have incorporated a sparsity constraint and an acyclicity constraint into Haan. A motivation for the combination is that an acyclicity constraint ensures edges go only forward in time, and a sparsity constraint is especially useful in cases with much fewer samples than variables. (Pamfil, page 3, col. 2, § 2.2, from line 1 to the fourth line below equation (5)) However, Haan and Pamfil do not explicitly teach: a template selection regularization loss is employed to enable consistency in template selection across similar time steps. But Singer teaches: a template selection regularization loss is employed to enable consistency in template selection across similar time steps. (Examiner treats “similar time steps” as “time steps”. C. 1, L. 49-53; C. 4, L. 28-38; C. 5, L. 49-52 and L. 64 to C. 6, L. 4; and C. 8, L. 8-16 disclose a template selection regularization penalty. The penalty is maintained across a particular number of training iterations (time steps).) It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have incorporated Singer’s regularization penalty into the combination of Haan and Pamfil. A motivation for the combination is that a regularization penalty may control model complexity and improve model generalization on inference data. (Singer, C. 5, L. 64 to C. 6, L. 4) 07-21-aia AIA Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Haan et al. (“Causal Confusion in Imitation Learning”) in view of Zhang et al. (US 20170235626 A1) . Regarding claim 7, Haan teaches: The learning system according to claim 1, However, Haan does not explicitly teach: wherein the causal graph is generated by optimizing the causal graph based on constraints. But Zhang teaches: wherein the causal graph is generated by optimizing the causal graph based on constraints. (Zhang’s claim 4 on page 4, the title “Uncover Causal Relations via Non-Negative Sparse Regression” above paragraph [0021], and paragraphs [0021]-[0023], where an optimization is the regression in Equation (4) and a constraint is a sparsity constraint.) It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have optimized Haan’s causal graph based on a sparsity constraint. A motivation for the combination is to capture the dependence relationships among variables in a time series. (Zhang, [0002], lines 11-15) Double Patenting Claim 1 is provisionally rejected on the ground of nonstatutory double patenting as being unpatentable over claim 15 of copending Application No. 17/877,081 (reference application). Although the claims at issue are not identical, they are not patentably distinct from each other. In claim 1, the limitation “A learning system comprising: at least one memory storing instructions; and at least one processor configured to access the at least one memory and execute the instructions to” is anticipated by reference claim 15 limitation “A system for learning a self-explainable imitator by discovering causal relationships between states and actions, the system comprising: a memory; and one or more processors in communication with the memory configured to”. Communication between the processor and memory and the processor being configured to perform steps indicates the processor executes instructions stored in the memory. The limitation “obtaining demonstrations of a target task from experts for training a model to generate a learned policy” is anticipated by reference claim 15 limitation “obtain demonstrations of a target task from experts for training a model to generate a learned policy”. The limitation “training the model by computing actions to be taken with respect to states” is anticipated by reference claim 15 limitation “train the model the learning component computing actions to be taken with respect to states” The limitation “generating dynamic causal graphs for each environment state, wherein each of dynamic causal graphs is a Directed Acrylic Graph” is anticipated by reference claim 15 limitation “generate dynamic causal graphs for each environment state in the learned policy, wherein each dynamic causal graph is a continuous directed acyclic graph”. The limitation “encoding discovered causal relationships by updating state variable embeddings” is anticipated by reference claim 15 limitation “encode discovered causal relationships by updating the state variable embeddings”. The limitation “outputting the learned policy” is anticipated by reference claim 15 limitation “output the learned policy”. This is an anticipatory-type, provisional nonstatutory double patenting rejection because the patentably indistinct claims have not in fact been patented. Claim 2 is provisionally rejected on the ground of nonstatutory double patenting as being unpatentable over claim 16 of copending Application No. 17/877,081 (reference application). Although the claims at issue are not identical, they are not patentably distinct from each other. In claim 2, the limitation “conduct an imitation learning task and a state regression task by employing the updated state variable embeddings as evidence” is anticipated by reference claim 16 limitation “an imitation learning task and a state regression task are conducted, via an action prediction component, by employing the updated state variable embeddings as input evidence.” This is an anticipatory-type, provisional nonstatutory double patenting rejection because the patentably indistinct claims have not in fact been patented. Claim 3 is provisionally rejected on the ground of nonstatutory double patenting as being unpatentable over claim 17 of copending Application No. 17/877,081 (reference application). Although the claims at issue are not identical, they are not patentably distinct from each other. In claim 3, the limitation “the state regression task is used to provide auxiliary signals for learning causal edges among state variables” is anticipated by reference claim 16 limitation “the state regression task is used to provide auxiliary signals for learning causal edges among state variables.” This is an anticipatory-type, provisional nonstatutory double patenting rejection because the patentably indistinct claims have not in fact been patented. Claim 5 is provisionally rejected on the ground of nonstatutory double patenting as being unpatentable over claim 19 of copending Application No. 17/877,081 (reference application). Although the claims at issue are not identical, they are not patentably distinct from each other. In claim 5, the limitation “the state variable embeddings are updated with propagated messages from variables it depends on by employing an edge-aware update layer” is anticipated by reference claim 19 limitation “the state variable embeddings are updated with propagated messages from variables it depends on by employing an edge-aware update layer.” This is an anticipatory-type, provisional nonstatutory double patenting rejection because the patentably indistinct claims have not in fact been patented. Claim 1 is provisionally rejected on the ground of nonstatutory double patenting as being unpatentable over claim 2 of copending Application No. 18/471,558 (reference application) in view of Pamfil et al. (“DYNOTEARS: Structure Learning from Time-Series Data”). Reference claim 2 incorporates the features of reference claim 1. In claim 1, the limitation “A learning system comprising: at least one memory storing instructions; and at least one processor configured to access the at least one memory and execute the instructions to” is taught by reference claim 1 limitation “An action prediction system comprising: at least one memory storing instructions; and at least one processor configured to access the at least one memory and execute the instructions to”. The limitation “ obtaining demonstrations of a target task from experts for training a model to generate a learned policy” is taught by reference claim 2 limitation “the action is predicted by using a model and the updated state variable embeddings wherein the model is adversarially trained with a discrimination model to discriminate between predicted actions and demonstrations from expert by machine-learning algorithm.” Demonstrations from experts would have to be obtained before starting adversarial training, and predicting an action using a trained model indicates it has learned a policy. The limitation “training the model by computing actions to be taken with respect to states” is taught by reference claim 2 limitation “the action is predicted by using a model and the updated state variable embeddings wherein the model is adversarially trained with a discrimination model to discriminate between predicted actions and demonstrations from expert by machine-learning algorithm”. The limitation “generating dynamic causal graphs for each environment state” is taught by reference claim 1 limitation “generate a causal graph indicating relationships between the current states based on the current states” and “encode the causal graph by updating state variable embeddings”. Updating state variable embeddings of a causal graph would cause it to change and be dynamic. The limitation “encoding discovered causal relationships by updating state variable embeddings” is taught by reference claim 1 limitation “encode the causal graph by updating state variable embeddings”. The limitation “outputting the learned policy” is taught by reference claim 2 limitation “the action is predicted by using a model and the updated state variable embeddings”. Using a model to predict actions indicates the system has output or generated the learned policy. However, reference claim 2 does not explicitly teach (underlines indicate limitations not taught): wherein each of dynamic causal graphs is a Directed Acrylic Graph But Pamfil teaches: generating dynamic causal graphs for each environment state, wherein each of dynamic causal graphs is a Directed Acrylic Graph; (Page 1, col. 2, lines 3-15; Page 2, Fig. 1 and its caption; and Page 2 § 2.1, lines 1-8 discloses generating a dynamic causal graph such as the one depicted in Fig. 1.) It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have incorporated the dynamic feature of Pamfil’s causal graphs into reference claim 2. A motivation for the combination is to understand a time series whose variables influence each other in both a contemporaneous and a time-lagged manner. (Pamfil, Page 2 § 2.1, lines 1-8) This is an obviousness-type, provisional nonstatutory double patenting rejection. Claim 7 is provisionally rejected on the ground of nonstatutory double patenting as being unpatentable over claim 2 of copending Application No. 18/471,558 (reference application) in view of Pamfil et al. (“DYNOTEARS: Structure Learning from Time-Series Data”) and Zhang et al. (US 20170235626 A1). Reference claim 2 inherits the features of parent reference claim 1. Regarding claim 7, the limitation “the causal graph is generated” is taught by reference claim 1 limitation “generate a causal graph” in line 6. However, reference claim 2 and Pamfil do not explicitly teach: the causal graph is generated by optimizing the causal graph based on constraints. But Zhang teaches: the causal graph is generated by optimizing the causal graph based on constraints. (Zhang’s claim 4 on page 4, the title “Uncover Causal Relations via Non-Negative Sparse Regression” above paragraph [0021], and paragraphs [0021]-[0023], where an optimization is the regression in Equation (4) and a constraint is a sparsity constraint.) It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have optimized the dynamic causal graph in the combination of reference claim 2 and Pamfil based on a sparsity constraint. A motivation for the combination is to capture the dependence relationships among variables in a time series. (Zhang, [0002], lines 11-15) This is an obviousness-type, provisional nonstatutory double patenting rejection. Claim 1 is provisionally rejected on the ground of nonstatutory double patenting as being unpatentable over claim 3 of copending Application No. 18/471,564 (reference application). Although the claims at issue are not identical, they are not patentably distinct from each other. Reference claim 3 inherits the features of reference claim 1. In claim 1, the limitation “A learning system comprising: at least one memory storing instructions; and at least one processor configured to access the at least one memory and execute the instructions to” is anticipated by reference claim 1 limitation “A treatment prediction system comprising: at least one memory storing instructions; and at least one processor configured to access the at least one memory and execute the instructions to”. The limitations “obtaining demonstrations of a target task from experts for training a model to generate a learned policy” and “training the model by computing actions to be taken with respect to states” are both anticipated by reference claim 1 limitation “predict an interpretable treatment for the patient to be taken with respect to the health states using the updated state variable embeddings as input to the machine learning model, wherein the machine learning model is adversarially trained with a discrimination model to discriminate between predicted actions and demonstrations from an expert by machine-learning algorithm”. Demonstrations from experts would have to be obtained before starting adversarial training, and predicting an interpretable treatment using a trained machine learning model indicates it has learned a policy. The limitation “generating dynamic causal graphs for each environment state, wherein each of dynamic causal graphs is a Directed Acrylic Graph” is anticipated by reference claim 1 limitations “generate a causal graph with a machine learning model that indicates relationships between the health states” and “encode the causal graph by updating state variable embeddings”, and by reference claim 3 limitation “wherein the causal graph is a Directed Acyclic Graph indicating relationships between the health states.” Updating state variable embeddings of the causal graph implies it to change and be dynamic. The limitation “encoding discovered causal relationships by updating state variable embeddings” is anticipated by reference claim 1 limitation “encode the causal graph by updating state variable embeddings that enforce consistency between causal edges that represent the causal relationship and behavior of the machine learning model”. The limitation “outputting the learned policy” is anticipated by reference claim 1 limitation “output the predicted interpretable treatment to a display used by a healthcare provider, a doctor, or a nurse.” This is an anticipatory-type, provisional nonstatutory double patenting rejection because the patentably indistinct claims have not in fact been patented. Claim 7 is provisionally rejected on the ground of nonstatutory double patenting as being unpatentable over claim 3 of copending Application No. 18/471,564 (reference application) in view of Zhang et al. (US 20170235626 A1). Reference claim 3 inherits the features of reference claim 1. Regarding claim 7, the limitation “the causal graph is generated” is taught by reference claim 1 limitation “generate a causal graph”. However, reference claim 3 does not explicitly teach: the causal graph is generated by optimizing the causal graph based on constraints. But Zhang teaches: the causal graph is generated by optimizing the causal graph based on constraints. (Zhang’s claim 4 on page 4, the title “Uncover Causal Relations via Non-Negative Sparse Regression” above paragraph [0021], and paragraphs [0021]-[0023], where an optimization is the regression in Equation (4) and a constraint is a sparsity constraint.) It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have optimized the dynamic causal graph in reference claim 3 based on a sparsity constraint. A motivation for the combination is to capture the dependence relationships among variables in a time series. (Zhang, [0002], lines 11-15) This is an obviousness-type, provisional nonstatutory double patenting rejection. Conclusion Any inquiry concerning this communication or earlier communications from the examiner should be directed to Asher H. Jablon whose telephone number is (571)270-7648. The examiner can normally be reached Monday - Friday, 9:00 am - 6:00 pm. 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, Abdullah Al Kawsar can be reached at (571)270-3169. 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. /A.H.J./Examiner, Art Unit 2127 /ABDULLAH AL KAWSAR/Supervisory Patent Examiner, Art Unit 2127 Application/Control Number: 18/471,570 Page 2 Art Unit: 2127 Application/Control Number: 18/471,570 Page 3 Art Unit: 2127 Application/Control Number: 18/471,570 Page 4 Art Unit: 2127 Application/Control Number: 18/471,570 Page 5 Art Unit: 2127 Application/Control Number: 18/471,570 Page 6 Art Unit: 2127 Application/Control Number: 18/471,570 Page 7 Art Unit: 2127 Application/Control Number: 18/471,570 Page 8 Art Unit: 2127 Application/Control Number: 18/471,570 Page 9 Art Unit: 2127 Application/Control Number: 18/471,570 Page 10 Art Unit: 2127 Application/Control Number: 18/471,570 Page 11 Art Unit: 2127 Application/Control Number: 18/471,570 Page 12 Art Unit: 2127 Application/Control Number: 18/471,570 Page 13 Art Unit: 2127 Application/Control Number: 18/471,570 Page 14 Art Unit: 2127 Application/Control Number: 18/471,570 Page 15 Art Unit: 2127 Application/Control Number: 18/471,570 Page 16 Art Unit: 2127 Application/Control Number: 18/471,570 Page 17 Art Unit: 2127 Application/Control Number: 18/471,570 Page 18 Art Unit: 2127 Application/Control Number: 18/471,570 Page 19 Art Unit: 2127 Application/Control Number: 18/471,570 Page 20 Art Unit: 2127 Application/Control Number: 18/471,570 Page 21 Art Unit: 2127 Application/Control Number: 18/471,570 Page 22 Art Unit: 2127 Application/Control Number: 18/471,570 Page 23 Art Unit: 2127 Application/Control Number: 18/471,570 Page 24 Art Unit: 2127 Application/Control Number: 18/471,570 Page 25 Art Unit: 2127
Prosecution Insights