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 .
Response to Amendment
As per the instant Application having Application number 18/260,156 the examiner acknowledges the applicant's submission of the preliminary amendment dated 06/30/2023. At this point, the abstract, specification, drawings, and claims 4-5, 7-8, 14, 16, and 18-19 have been amended. Claim 20 has been cancelled. Claims 1-19 are pending.
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(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.
Claim(s) 1-3, 5-6, 8-12, 16, and 18-19 is/are rejected under 35 U.S.C. 102 (a)(1) as being anticipated by Vijil et al., (US 2021/0374551 A1, hereinafter Vijil).
Regarding claim 1:
Vijil shows:
“A method of training a classification model, comprising: processing first sample data by using an auto-encoding module, so as to obtain reconstructed sample data, wherein the auto-encoding module comprises at least one autoencoder, the autoencoder comprises an encoder and a decoder,” (Paragraph [0039]: “The encoder training component 112 can supply the molecular structures to the one or more autoencoders to encode one or more latent representations, which can then be decoded by the one or more decoders to reconstruct the molecular structures.” And in paragraph [0043]: “In various embodiments, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more regressors and/or classifiers using a first training dataset. For example, the first training dataset 122 can include example molecular structures associated with a first attribute. Additionally, in one or more embodiments the first training dataset 122 can have the most molecular structure examples amongst the available training datasets 122”)
“and the first sample data comprises medical sample data;” (Paragraph [0043]: “In various embodiments, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more regressors and/or classifiers using a first training dataset. For example, the first training dataset 122 can include example molecular structures associated with a first attribute. Additionally, in one or more embodiments the first training dataset 122 can have the most molecular structure examples amongst the available training datasets 122”, and in paragraph [0045]: “FIG. 4 illustrates a diagram of an example, non-limiting graph 400 regarding exemplary training datasets 122 that can be analyzed by the joint training component 302 in accordance with one or more embodiments described herein. Repetitive description of like elements employed in other embodiments described herein is omitted for sake of brevity. For example, the system 100 can be employed to generate molecular structures for one or more therapeutic compounds. Graph 400 can characterize the distribution of quantitative estimate of medication-likeness (“QED”) with respect to an exemplary use case scenario regarding a target attribute profile 120 that delineates the following three attributes: small molecules size (e.g., having a molecular weight of less than 500 molar mass (“M”); high bonding affinity (e.g., having a pIC50 value of greater than 6); and histone deacetylase 1 (“HDAC1”) targeting”)
“processing first sample feature data of the first sample data by using a classification module, so as to obtain a first sample classification result;” (Paragraph [0043]: “the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more regressors and/or classifiers using a first training dataset. For example, the first training dataset 122 can include example molecular structures associated with a first attribute. Additionally, in one or more embodiments the first training dataset 122 can have the most molecular structure examples amongst the available training datasets 122. Thereby, the joint training component 302 can train the one or more autoencoders 204, regressors, and/or classifiers to predict the first attribute. For example, the joint training component 302 can utilize the first training dataset 122 to assist the autoencoders 204 (e.g., variational autoencoders) to learn a disentangled latent space 205 that can accurately predict whether a molecular structure exhibits the first attribute.”)
“jointly training the auto-encoding module and the classification module according to the first sample data, the reconstructed sample data, the first sample classification result, and a first sample classification label value of the first sample data;” (Paragraph [0039]: “The encoder training component 112 can supply the molecular structures to the one or more autoencoders to encode one or more latent representations, which can then be decoded by the one or more decoders to reconstruct the molecular structures.” And in paragraph [0044]: “Subsequently, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more additional regressors and/or classifiers using one or more additional training datasets 122. For example, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more second regressors and/or classifiers using a second training dataset 122. For instance, the second training dataset 122 can include example molecular structures associated with a second attribute. Thereby, the joint training component 302 can train the one or more autoencoders 204, regressors, and/or classifiers to predict the second attribute. In various embodiments, the joint training component 302 can continue to train the one or more autoencoders 204 (e.g., variational autoencoders) with additional regressors and/or classifiers on each of the training datasets 122, and thereby each of the attributes associated with the training datasets 122. By analyzing each training dataset 122 the transfer learning component 108 can apply lessons learned from training directed to a first attribute to training directed to additional attributes.”)
“and obtaining the classification model according to the trained encoder and the trained classification module.” (Paragraph [0043]: “the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more regressors and/or classifiers using a first training dataset. For example, the first training dataset 122 can include example molecular structures associated with a first attribute. Additionally, in one or more embodiments the first training dataset 122 can have the most molecular structure examples amongst the available training datasets 122. Thereby, the joint training component 302 can train the one or more autoencoders 204, regressors, and/or classifiers to predict the first attribute. For example, the joint training component 302 can utilize the first training dataset 122 to assist the autoencoders 204 (e.g., variational autoencoders) to learn a disentangled latent space 205 that can accurately predict whether a molecular structure exhibits the first attribute.” And in paragraph [0044]: “Subsequently, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more additional regressors and/or classifiers using one or more additional training datasets 122. For example, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more second regressors and/or classifiers using a second training dataset 122. For instance, the second training dataset 122 can include example molecular structures associated with a second attribute. Thereby, the joint training component 302 can train the one or more autoencoders 204, regressors, and/or classifiers to predict the second attribute. In various embodiments, the joint training component 302 can continue to train the one or more autoencoders 204 (e.g., variational autoencoders) with additional regressors and/or classifiers on each of the training datasets 122, and thereby each of the attributes associated with the training datasets 122. By analyzing each training dataset 122 the transfer learning component 108 can apply lessons learned from training directed to a first attribute to training directed to additional attributes.”)
Regarding claim 2:
Vijil shows the method of claim 1 as claimed and specified above.
And Vijil shows “wherein the processing first sample data by using an auto-encoding module, so as to obtain reconstructed sample data comprises: processing the first sample data by using at least one encoder, so as to obtain first sample feature data; and processing the first sample feature data by using at least one decoder, so as to obtain the reconstructed sample data.” (Paragraph [0039]: “the encoder training component 112 can train an unconditional machine learning model by employing one or more autoencoders to analyze unlabeled data from a defined data distribution set. For example, the one or more autoencoders can be one or more unconditional variational autoencoders comprised within a neural network model, such as a bidirectional regional neural network. Additionally, the encoder training component 112 can train one or more decoders on the unlabeled data. For example, the defined data distribution set can characterize the molecular structures for a defined group of chemical compounds. The encoder training component 112 can supply the molecular structures to the one or more autoencoders to encode one or more latent representations, which can then be decoded by the one or more decoders to reconstruct the molecular structures. The encoder training component 112 can analyze the latent representations to determine one or more Kullback-Leibler divergence values (“KL-loss”). Additionally, the encoder training component 112 can compare the reconstructed molecular structures with the supplied molecular structures to determine one or more reconstruction loss values. Moreover, in various embodiments the encoder training component 112 can analyze the latent representations to determine one or more classification and/or regression loss values regarding one or more attributes attributed to the molecular structures”)
Regarding claim 3:
Vijil shows the method of claim 2 as claimed and specified above.
And Vijil shows “wherein the classification module comprises at least one classifier, and the first sample feature data comprises at least one first sample feature dimension data; and wherein the processing first sample feature data of the first sample data by using a classification module, so as to obtain a first sample classification result comprises: processing the first sample feature dimension data corresponding to the at least one classifier by using the at least one classifier, so as to obtain the first sample classification result.” (Paragraph [0043]: “the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more regressors and/or classifiers using a first training dataset. For example, the first training dataset 122 can include example molecular structures associated with a first attribute. Additionally, in one or more embodiments the first training dataset 122 can have the most molecular structure examples amongst the available training datasets 122. Thereby, the joint training component 302 can train the one or more autoencoders 204, regressors, and/or classifiers to predict the first attribute. For example, the joint training component 302 can utilize the first training dataset 122 to assist the autoencoders 204 (e.g., variational autoencoders) to learn a disentangled latent space 205 that can accurately predict whether a molecular structure exhibits the first attribute.” And in paragraph [0044]: “Subsequently, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more additional regressors and/or classifiers using one or more additional training datasets 122. For example, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more second regressors and/or classifiers using a second training dataset 122. For instance, the second training dataset 122 can include example molecular structures associated with a second attribute. Thereby, the joint training component 302 can train the one or more autoencoders 204, regressors, and/or classifiers to predict the second attribute. In various embodiments, the joint training component 302 can continue to train the one or more autoencoders 204 (e.g., variational autoencoders) with additional regressors and/or classifiers on each of the training datasets 122, and thereby each of the attributes associated with the training datasets 122. By analyzing each training dataset 122 the transfer learning component 108 can apply lessons learned from training directed to a first attribute to training directed to additional attributes.”)
Regarding claim 5:
Vijil shows the method of claim 1 as claimed and specified above.
And Vijil shows “wherein the first sample feature data comprises a plurality of first sample feature dimension data; wherein the method further comprises: determining at least one first sample feature dimension data from the plurality of first sample feature dimension data based on a feature selection method; and obtaining second sample feature data based on the at least one first sample feature dimension data; and wherein the processing first sample feature data of the first sample data by using a classification module, so as to obtain a first sample classification result comprises: processing the second sample feature data by using the classification module, so as to obtain the first sample classification result.” (Paragraph [0043]: “the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more regressors and/or classifiers using a first training dataset. For example, the first training dataset 122 can include example molecular structures associated with a first attribute. Additionally, in one or more embodiments the first training dataset 122 can have the most molecular structure examples amongst the available training datasets 122. Thereby, the joint training component 302 can train the one or more autoencoders 204, regressors, and/or classifiers to predict the first attribute. For example, the joint training component 302 can utilize the first training dataset 122 to assist the autoencoders 204 (e.g., variational autoencoders) to learn a disentangled latent space 205 that can accurately predict whether a molecular structure exhibits the first attribute.” And in paragraph [0044]: “Subsequently, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more additional regressors and/or classifiers using one or more additional training datasets 122. For example, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more second regressors and/or classifiers using a second training dataset 122. For instance, the second training dataset 122 can include example molecular structures associated with a second attribute. Thereby, the joint training component 302 can train the one or more autoencoders 204, regressors, and/or classifiers to predict the second attribute. In various embodiments, the joint training component 302 can continue to train the one or more autoencoders 204 (e.g., variational autoencoders) with additional regressors and/or classifiers on each of the training datasets 122, and thereby each of the attributes associated with the training datasets 122. By analyzing each training dataset 122 the transfer learning component 108 can apply lessons learned from training directed to a first attribute to training directed to additional attributes.”)
Regarding claim 6:
Vijil shows the method of claim 5 as claimed and specified above.
And Vijil shows “wherein the determining at least one first sample feature dimension data from the plurality of first sample feature dimension data based on a feature selection method comprises: determining, based on an importance evaluation strategy, an importance evaluation value corresponding to the plurality of first sample feature dimension data, so as to obtain a plurality of importance evaluation values, wherein the importance evaluation value indicates an importance of the first sample feature dimension data; and determining the at least one sample feature dimension data from the plurality of first sample feature dimension data according to the plurality of importance evaluation values.” (Paragraph [0043]: “the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more regressors and/or classifiers using a first training dataset. For example, the first training dataset 122 can include example molecular structures associated with a first attribute. Additionally, in one or more embodiments the first training dataset 122 can have the most molecular structure examples amongst the available training datasets 122. Thereby, the joint training component 302 can train the one or more autoencoders 204, regressors, and/or classifiers to predict the first attribute. For example, the joint training component 302 can utilize the first training dataset 122 to assist the autoencoders 204 (e.g., variational autoencoders) to learn a disentangled latent space 205 that can accurately predict whether a molecular structure exhibits the first attribute.” And in paragraph [0044]: “Subsequently, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more additional regressors and/or classifiers using one or more additional training datasets 122. For example, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more second regressors and/or classifiers using a second training dataset 122. For instance, the second training dataset 122 can include example molecular structures associated with a second attribute. Thereby, the joint training component 302 can train the one or more autoencoders 204, regressors, and/or classifiers to predict the second attribute. In various embodiments, the joint training component 302 can continue to train the one or more autoencoders 204 (e.g., variational autoencoders) with additional regressors and/or classifiers on each of the training datasets 122, and thereby each of the attributes associated with the training datasets 122. By analyzing each training dataset 122 the transfer learning component 108 can apply lessons learned from training directed to a first attribute to training directed to additional attributes.” And in paragraph [0039]: “Based on the KL-loss, reconstruction loss, and/or classification/regression loss values, the encoder training component 112 can determine whether the one or more autoencoders and/or decoders are pre-trained for subsequent transfer learning procedures described herein. For example, the encoder training component 112 can compare the one or more KL-loss, reconstruction loss, and/or classification/regression loss values to one or more defined thresholds to facilitate the determinations.”)
Regarding claim 8:
Vijil shows the method of claim 1 as claimed and specified above.
And Vijil shows “wherein the jointly training the auto-encoding module and the classification module according to the first sample data, the reconstructed sample data, the first sample classification result, and a first sample classification label value of the first sample data comprises: obtaining a first output value according to the first sample data and the reconstructed sample data based on a first loss function; obtaining a second output value according to the first sample classification result and the first sample classification label value based on a second loss function; and adjusting a model parameter of the auto-encoding module and a model parameter of the classification module according to the first output value and the second output value.” (Paragraph [0039]: “the encoder training component 112 can train an unconditional machine learning model by employing one or more autoencoders to analyze unlabeled data from a defined data distribution set. For example, the one or more autoencoders can be one or more unconditional variational autoencoders comprised within a neural network model, such as a bidirectional regional neural network. Additionally, the encoder training component 112 can train one or more decoders on the unlabeled data. For example, the defined data distribution set can characterize the molecular structures for a defined group of chemical compounds. The encoder training component 112 can supply the molecular structures to the one or more autoencoders to encode one or more latent representations, which can then be decoded by the one or more decoders to reconstruct the molecular structures. The encoder training component 112 can analyze the latent representations to determine one or more Kullback-Leibler divergence values (“KL-loss”). Additionally, the encoder training component 112 can compare the reconstructed molecular structures with the supplied molecular structures to determine one or more reconstruction loss values. Moreover, in various embodiments the encoder training component 112 can analyze the latent representations to determine one or more classification and/or regression loss values regarding one or more attributes attributed to the molecular structures”)
Regarding claim 9:
Vijil shows the method of claim 8 as claimed and specified above.
And Vijil shows “wherein the obtaining a second output value according to the first sample classification result and the first sample classification label value based on a second loss function comprises: obtaining the second output value according to the first sample classification result and the first sample classification label value based on a third loss function, wherein the third loss function is determined according to the second loss function and a first penalty term.” (Paragraph [0039]: “the encoder training component 112 can train an unconditional machine learning model by employing one or more autoencoders to analyze unlabeled data from a defined data distribution set. For example, the one or more autoencoders can be one or more unconditional variational autoencoders comprised within a neural network model, such as a bidirectional regional neural network. Additionally, the encoder training component 112 can train one or more decoders on the unlabeled data. For example, the defined data distribution set can characterize the molecular structures for a defined group of chemical compounds. The encoder training component 112 can supply the molecular structures to the one or more autoencoders to encode one or more latent representations, which can then be decoded by the one or more decoders to reconstruct the molecular structures. The encoder training component 112 can analyze the latent representations to determine one or more Kullback-Leibler divergence values (“KL-loss”). Additionally, the encoder training component 112 can compare the reconstructed molecular structures with the supplied molecular structures to determine one or more reconstruction loss values. Moreover, in various embodiments the encoder training component 112 can analyze the latent representations to determine one or more classification and/or regression loss values regarding one or more attributes attributed to the molecular structures” And in paragraph [0060]: “At 904, the computer-implemented method 900 can comprise training (e.g., via encoder training component 112), by the system 100, one or more unconditional machine learning models by employing one or more autoencoders 204 (e.g., variational autoencoders) to analyze unlabeled data from a defined data distribution set. For example, the training at 904 can include computing KL-loss 210, classification/regression loss 212, and/or reconstruction loss 214 in accordance with various embodiments described herein.”)
Regarding claim 10:
Vijil shows the method of claim 1 as claimed and specified above.
And Vijil shows “processing second sample data by using an encoder of the classification model, so as to obtain third sample feature data; and optimizing the classification module of the classification model by using the third sample feature data.” (Paragraph [0043]: “the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more regressors and/or classifiers using a first training dataset. For example, the first training dataset 122 can include example molecular structures associated with a first attribute. Additionally, in one or more embodiments the first training dataset 122 can have the most molecular structure examples amongst the available training datasets 122. Thereby, the joint training component 302 can train the one or more autoencoders 204, regressors, and/or classifiers to predict the first attribute. For example, the joint training component 302 can utilize the first training dataset 122 to assist the autoencoders 204 (e.g., variational autoencoders) to learn a disentangled latent space 205 that can accurately predict whether a molecular structure exhibits the first attribute.” And in paragraph [0044]: “Subsequently, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more additional regressors and/or classifiers using one or more additional training datasets 122. For example, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more second regressors and/or classifiers using a second training dataset 122. For instance, the second training dataset 122 can include example molecular structures associated with a second attribute. Thereby, the joint training component 302 can train the one or more autoencoders 204, regressors, and/or classifiers to predict the second attribute. In various embodiments, the joint training component 302 can continue to train the one or more autoencoders 204 (e.g., variational autoencoders) with additional regressors and/or classifiers on each of the training datasets 122, and thereby each of the attributes associated with the training datasets 122. By analyzing each training dataset 122 the transfer learning component 108 can apply lessons learned from training directed to a first attribute to training directed to additional attributes.”)
Regarding claim 11:
Vijil shows the method of claim 10 as claimed and specified above.
And Vijil shows “wherein the third sample feature data comprises a plurality of second sample feature dimension data; wherein the optimizing the classification module of the classification model by using the third sample feature data comprises repeatedly performing operations until a performance test result of the classification module meets a predetermined performance condition, and the operations comprise: testing a model performance of the classification module by using candidate sample feature data, so as to obtain the performance test result; and determining, in response to determining that the performance test result does not meet the predetermined performance condition, at least one second sample feature dimension data from the plurality of second sample feature dimension data, so as to obtain new candidate sample feature data.” (Paragraph [0043]: “the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more regressors and/or classifiers using a first training dataset. For example, the first training dataset 122 can include example molecular structures associated with a first attribute. Additionally, in one or more embodiments the first training dataset 122 can have the most molecular structure examples amongst the available training datasets 122. Thereby, the joint training component 302 can train the one or more autoencoders 204, regressors, and/or classifiers to predict the first attribute. For example, the joint training component 302 can utilize the first training dataset 122 to assist the autoencoders 204 (e.g., variational autoencoders) to learn a disentangled latent space 205 that can accurately predict whether a molecular structure exhibits the first attribute.” And in paragraph [0044]: “Subsequently, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more additional regressors and/or classifiers using one or more additional training datasets 122. For example, the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more second regressors and/or classifiers using a second training dataset 122. For instance, the second training dataset 122 can include example molecular structures associated with a second attribute. Thereby, the joint training component 302 can train the one or more autoencoders 204, regressors, and/or classifiers to predict the second attribute. In various embodiments, the joint training component 302 can continue to train the one or more autoencoders 204 (e.g., variational autoencoders) with additional regressors and/or classifiers on each of the training datasets 122, and thereby each of the attributes associated with the training datasets 122. By analyzing each training dataset 122 the transfer learning component 108 can apply lessons learned from training directed to a first attribute to training directed to additional attributes.” And in paragraph [0039]: “Based on the KL-loss, reconstruction loss, and/or classification/regression loss values, the encoder training component 112 can determine whether the one or more autoencoders and/or decoders are pre-trained for subsequent transfer learning procedures described herein. For example, the encoder training component 112 can compare the one or more KL-loss, reconstruction loss, and/or classification/regression loss values to one or more defined thresholds to facilitate the determinations.”)
Regarding claim 12:
Vijil shows the method of claim 10 as claimed and specified above.
And Vijil shows “wherein the optimizing the classification module of the classification model by using the third sample feature data comprises: processing the third sample feature data by using the classification module, so as to obtain a second sample classification result; obtaining a third output value according to the second sample classification result and a second sample classification label value of the second sample data based on a fourth loss function, wherein the fourth loss function is determined according to the second loss function and a second penalty term; and adjusting a model parameter of the classification module according to the third output value.” (Paragraph [0039]: “the encoder training component 112 can train an unconditional machine learning model by employing one or more autoencoders to analyze unlabeled data from a defined data distribution set. For example, the one or more autoencoders can be one or more unconditional variational autoencoders comprised within a neural network model, such as a bidirectional regional neural network. Additionally, the encoder training component 112 can train one or more decoders on the unlabeled data. For example, the defined data distribution set can characterize the molecular structures for a defined group of chemical compounds. The encoder training component 112 can supply the molecular structures to the one or more autoencoders to encode one or more latent representations, which can then be decoded by the one or more decoders to reconstruct the molecular structures. The encoder training component 112 can analyze the latent representations to determine one or more Kullback-Leibler divergence values (“KL-loss”). Additionally, the encoder training component 112 can compare the reconstructed molecular structures with the supplied molecular structures to determine one or more reconstruction loss values. Moreover, in various embodiments the encoder training component 112 can analyze the latent representations to determine one or more classification and/or regression loss values regarding one or more attributes attributed to the molecular structures” And in paragraph [0060]: “At 904, the computer-implemented method 900 can comprise training (e.g., via encoder training component 112), by the system 100, one or more unconditional machine learning models by employing one or more autoencoders 204 (e.g., variational autoencoders) to analyze unlabeled data from a defined data distribution set. For example, the training at 904 can include computing KL-loss 210, classification/regression loss 212, and/or reconstruction loss 214 in accordance with various embodiments described herein.”)
Regarding claim 16:
Vijil shows the method of claim 11 as claimed and specified above.
And Vijil shows “comprising: acquiring target data, wherein the target data comprises medical target data; and inputting the target data into a classification model to obtain a classification result, wherein the classification model is trained using the method of claim 1.” (Paragraph [0039]: “the encoder training component 112 can train an unconditional machine learning model by employing one or more autoencoders to analyze unlabeled data from a defined data distribution set. For example, the one or more autoencoders can be one or more unconditional variational autoencoders comprised within a neural network model, such as a bidirectional regional neural network. Additionally, the encoder training component 112 can train one or more decoders on the unlabeled data. For example, the defined data distribution set can characterize the molecular structures for a defined group of chemical compounds. The encoder training component 112 can supply the molecular structures to the one or more autoencoders to encode one or more latent representations, which can then be decoded by the one or more decoders to reconstruct the molecular structures. The encoder training component 112 can analyze the latent representations to determine one or more Kullback-Leibler divergence values (“KL-loss”). Additionally, the encoder training component 112 can compare the reconstructed molecular structures with the supplied molecular structures to determine one or more reconstruction loss values. Moreover, in various embodiments the encoder training component 112 can analyze the latent representations to determine one or more classification and/or regression loss values regarding one or more attributes attributed to the molecular structures. Based on the KL-loss, reconstruction loss, and/or classification/regression loss values, the encoder training component 112 can determine whether the one or more autoencoders and/or decoders are pre-trained for subsequent transfer learning procedures described herein. For example, the encoder training component 112 can compare the one or more KL-loss, reconstruction loss, and/or classification/regression loss values to one or more defined thresholds to facilitate the determinations.” In paragraph [0043]: “the joint training component 302 can further train the one or more autoencoders 204 (e.g., variational autoencoders) with one or more regressors and/or classifiers using a first training dataset. For example, the first training dataset 122 can include example molecular structures associated with a first attribute. Additionally, in one or more embodiments the first training dataset 122 can have the most molecular structure examples amongst the available training datasets 122” And in paragraph [0045]: “FIG. 4 illustrates a diagram of an example, non-limiting graph 400 regarding exemplary training datasets 122 that can be analyzed by the joint training component 302 in accordance with one or more embodiments described herein. Repetitive description of like elements employed in other embodiments described herein is omitted for sake of brevity. For example, the system 100 can be employed to generate molecular structures for one or more therapeutic compounds. Graph 400 can characterize the distribution of quantitative estimate of medication-likeness (“QED”) with respect to an exemplary use case scenario regarding a target attribute profile 120 that delineates the following three attributes: small molecules size (e.g., having a molecular weight of less than 500 molar mass (“M”); high bonding affinity (e.g., having a pIC50 value of greater than 6); and histone deacetylase 1 (“HDAC1”) targeting”
Regarding claim 18:
Vijil shows the method of claim 1 as claimed and specified above.
And Vijil shows “comprising: one or more processors; and a memory for storing one or more programs, wherein the one or more programs are configured to, when executed by the one or more processors, cause the one or more processors to implement the method of claim 1.” (Paragraph [0009]: “the program instructions can further cause the processor to train, by the processor, the autoencoder with a regressor using a first training dataset that includes molecular structures of a first group of compounds that exhibit a first attribute. Also, the program instructions can cause the processor to train, by the processor, autoencoder and regressor using a second training dataset that includes molecular structures of a second group of compounds that exhibit a second attribute. The target attribute profile comprises the first attribute and the second attribute. Additionally, the program instructions can cause the processor to train, by the processor, a classifier on a latent space learned by the autoencoder. Moreover, the program instructions can cause the processor to perform, by the processor, a conditional generation by executing the conditional machine learning model with rejection sampling in the latent space using a density model and the classifier. An advantage of such a computer program product method can be the generation of molecular structures that embody a defined combination of attributes of interest.”)
Regarding claim 19:
Vijil shows the method of claim 1 as claimed and specified above.
And Vijil shows “A computer readable storage medium having executable instructions therein, wherein the instructions are configured to, when executed by a processor, cause the processor to implement the method of claim 1.” (Paragraph [0009]: “the program instructions can further cause the processor to train, by the processor, the autoencoder with a regressor using a first training dataset that includes molecular structures of a first group of compounds that exhibit a first attribute. Also, the program instructions can cause the processor to train, by the processor, autoencoder and regressor using a second training dataset that includes molecular structures of a second group of compounds that exhibit a second attribute. The target attribute profile comprises the first attribute and the second attribute. Additionally, the program instructions can cause the processor to train, by the processor, a classifier on a latent space learned by the autoencoder. Moreover, the program instructions can cause the processor to perform, by the processor, a conditional generation by executing the conditional machine learning model with rejection sampling in the latent space using a density model and the classifier. An advantage of such a computer program product method can be the generation of molecular structures that embody a defined combination of attributes of interest.” And in paragraph [0087]: “The present invention may be a system, a method, and/or a computer program product at any possible technical detail level of integration. The computer program product may include a computer readable storage medium (or media) having computer readable program instructions thereon for causing a processor to carry out aspects of the present invention. The computer readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device.”)
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
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.
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.
Claim(s) 4, 7, and 13 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vijil in view of SHRIVASTAVA et al., (US 2021/0365965 A1, hereinafter Shrivastava).
Regarding claim 4:
Vijil shows the method of claim 1 as claimed and specified above.
But Vijil does not appear to explicitly recite “wherein the classification model comprises a first attention module; wherein the method further comprises: processing the first sample feature data of the first sample data by using the first attention module, so as to obtain first weighted sample feature data; and wherein the processing first sample feature data of the first sample data by using a classification module, so as to obtain a first sample classification result comprises: processing the first weighted sample feature data by using the classification module, so as to obtain the first sample classification result.”
However, Shrivastava teaches “wherein the classification model comprises a first attention module; wherein the method further comprises: processing the first sample feature data of the first sample data by using the first attention module, so as to obtain first weighted sample feature data; and wherein the processing first sample feature data of the first sample data by using a classification module, so as to obtain a first sample classification result comprises: processing the first weighted sample feature data by using the classification module, so as to obtain the first sample classification result.” (Paragraph [0098]: “the system can use data from health-care related sources. The system converts the data into sequenced activity data converted into numbers. The sequenced activities can include symptoms experienced on different days, meals eaten on each day, exercise on each day, etc. The resulting embeddings can be used to predict possible diagnosis, recommend treatments, predict treatment successes or side-effects, etc.” And in paragraph [0099]: “The system uses ML and weights to improve embeddings and predictions generated using the embeddings over time. In some examples, the system uses autoencoder frameworks during training using self-supervision to regenerate parts of the input and perform multi-task optimizations on different parts of the input without obtaining any input from customers or users. The encoder-decoder frameworks use sequential models with attention, which are combined with non-sequential data such as profile data, in linear layers in the encoder. This enables utilization of the self-supervised activity sequencing model 102 without customization (OOB), because users do not have to provide training labels for their data.”)
Vijil and Shrivastava are analogous in the arts because both Vijil and Shrivastava describe model training using medical data.
Therefore, it would be obvious to one of ordinary skill in the art at the filing date of the instant application, having the teachings of Vijil and Shrivastava before him or her, to modify the teachings of Vijil to include the teachings of Shrivastava in order to support attention modules and the use of weights to increase the accuracy and flexibility of training of Vijil and thereby increase the marketability of Vijil.
Regarding claim 7:
Vijil shows the method of claim 5 as claimed and specified above.
But Vijil does not appear to explicitly recite “wherein the classification model comprises a second attention module; wherein the method further comprises: processing the second sample feature data by using the second attention module, so as to obtain second weighted sample feature data; and wherein the processing the second sample feature data by using the classification module, so as to obtain the first sample classification result comprises: processing the second weighted sample feature data by using the classification module, so as to obtain the first sample classification result.”
However, Shrivastava teaches “wherein the classification model comprises a first attention module; wherein the method further comprises: processing the first sample feature data of the first sample data by using the first attention module, so as to obtain first weighted sample feature data; and wherein the processing first sample feature data of the first sample data by using a classification module, so as to obtain a first sample classification result comprises: processing the first weighted sample feature data by using the classification module, so as to obtain the first sample classification result.” (Paragraph [0098]: “the system can use data from health-care related sources. The system converts the data into sequenced activity data converted into numbers. The sequenced activities can include symptoms experienced on different days, meals eaten on each day, exercise on each day, etc. The resulting embeddings can be used to predict possible diagnosis, recommend treatments, predict treatment successes or side-effects, etc.” And in paragraph [0099]: “The system uses ML and weights to improve embeddings and predictions generated using the embeddings over time. In some examples, the system uses autoencoder frameworks during training using self-supervision to regenerate parts of the input and perform multi-task optimizations on different parts of the input without obtaining any input from customers or users. The encoder-decoder frameworks use sequential models with attention, which are combined with non-sequential data such as profile data, in linear layers in the encoder. This enables utilization of the self-supervised activity sequencing model 102 without customization (OOB), because users do not have to provide training labels for their data.”)
Vijil and Shrivastava are analogous in the arts because both Vijil and Shrivastava describe model training using medical data.
Therefore, it would be obvious to one of ordinary skill in the art at the filing date of the instant application, having the teachings of Vijil and Shrivastava before him or her, to modify the teachings of Vijil to include the teachings of Shrivastava in order to support attention modules and the use of weights to increase the accuracy and flexibility of training of Vijil and thereby increase the marketability of Vijil.
Regarding claim 13:
Vijil shows the method of claim 10 as claimed and specified above.
But Vijil does not appear to explicitly recite “wherein the classification model comprises a third attention module; wherein the method further comprises: processing the third sample feature data by using the third attention module, so as to obtain third weighted sample feature data; and wherein the optimizing the classification module of the classification model by using the third sample feature data comprises: optimizing the classification module by using the third weighted sample feature data.”
However, Shrivastava teaches “wherein the classification model comprises a third attention module; wherein the method further comprises: processing the third sample feature data by using the third attention module, so as to obtain third weighted sample feature data; and wherein the optimizing the classification module of the classification model by using the third sample feature data comprises: optimizing the classification module by using the third weighted sample feature data.” (Paragraph [0098]: “the system can use data from health-care related sources. The system converts the data into sequenced activity data converted into numbers. The sequenced activities can include symptoms experienced on different days, meals eaten on each day, exercise on each day, etc. The resulting embeddings can be used to predict possible diagnosis, recommend treatments, predict treatment successes or side-effects, etc.” And in paragraph [0099]: “The system uses ML and weights to improve embeddings and predictions generated using the embeddings over time. In some examples, the system uses autoencoder frameworks during training using self-supervision to regenerate parts of the input and perform multi-task optimizations on different parts of the input without obtaining any input from customers or users. The encoder-decoder frameworks use sequential models with attention, which are combined with non-sequential data such as profile data, in linear layers in the encoder. This enables utilization of the self-supervised activity sequencing model 102 without customization (OOB), because users do not have to provide training labels for their data.”)
Vijil and Shrivastava are analogous in the arts because both Vijil and Shrivastava describe model training using medical data.
Therefore, it would be obvious to one of ordinary skill in the art at the filing date of the instant application, having the teachings of Vijil and Shrivastava before him or her, to modify the teachings of Vijil to include the teachings of Shrivastava in order to support attention modules and the use of weights to increase the accuracy and flexibility of training of Vijil and thereby increase the marketability of Vijil.
Claim(s) 14-15 and 17 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vijil in view of Jordan et al., (US 2022/0028551A1, hereinafter Jordan).
Regarding claim 14:
Vijil shows the method of claim 1 as claimed and specified above.
But Vijil does not appear to explicitly recite “wherein the medical sample data comprises at least one of multi-omics sample data or medical sample image data.”
However, Jordan teaches “wherein the medical sample data comprises at least one of multi-omics sample data or medical sample image data.” (Paragraph [0015]: “Advantageously, the embodiments provided herein overcome the above and other problems by describing a multi-omic classifier to predict responses to PD-1/PD-L1 and CTLA-4 checkpoint blockade in various clinical indications, including but not limited to non-small cell lung cancer (NSCLC), melanoma, bladder cancer, and breast cancer. In one embodiment, the classifier is developed from training data that includes diagnostic imaging scans at baseline and follow-up intervals, along with existing biomarkers, relevant clinical, molecular, demographic, response and survival data. Examples of existing biomarkers used in clinical practice include: PD-L1 expression immunohistochemistry, tumor mutation burden (TMB), mutation mismatch repair (MMR), microsatellite instability (MSI), and neutrophil-to-lympocyte ratio (NLR). Furthermore, there is early evidence suggesting that laboratory tests, such as Lactate Dehydrogenase (LDH), S100 proteins and related blood serum proteins are predictive of immunotherapy response and pseudoprogression, specifically. In the near future, features and biomarkers extracted from the microbiome are expected to play a significant role as well.”)
Vijil and Jordan are analogous in the arts because both Vijil and Jordan describe model training using medical data.
Therefore, it would be obvious to one of ordinary skill in the art at the filing date of the instant application, having the teachings of Vijil and Jordan before him or her, to modify the teachings of Vijil to include the teachings of Jordan in order to support image medical data of Jordan to increase marketability of Vijil.
Regarding claim 15:
Vijil shows the method of claim 14 as claimed and specified above.
But Vijil does not appear to explicitly recite “wherein the multi-omics sample data comprise tumor multi-omics sample data, and the classification model is configured to determine a type of tumor.”
However, Jordan teaches “wherein the multi-omics sample data comprise tumor multi-omics sample data, and the classification model is configured to determine a type of tumor.” (Paragraph [0046]: “Model inputs using multiple resolutions and region-of-interest (ROI) sizes. A) The CNN model may prefer a subregion (ROI) of one or more CT scans as an input. ROIs of varying size and resolution may be used to create a redundant representation of the input CT image (or subregion) in the vicinity of the tumor location. By using multiple ROI sizes, the model can accommodate for tumors of different size and shape. For example, if only an ROI spanning 5×5×5 cm around the tumor was used, the model would likely not perform well on large tumors. Conversely, if a 50×50×50 cm ROI was used, the classifier would likely not perform well for smaller tumors that require high spatial resolution and fidelity. Combining ROI regions with small and large spatial dimensions in one model facilitates complementary learning of imaging features at the local context (e.g. tumor shape, texture, and intensity profile) and at the global context (e.g. location of the lesion within the body and with respect to other organs, lymph node involvement, patient's body mass composition and muscle reserve, overall health or vital organs, microcalcifications, etc.) and may ultimately results in more predictive and more robust treatment response and survival prediction models.”)
Regarding claim 17:
Vijil shows the method of claim 16 as claimed and specified above.
But Vijil does not appear to explicitly recite “wherein the medical target data comprises at least one of multi-omics target data or medical target image data.”
However, Jordan teaches “wherein the medical target data comprises at least one of multi-omics target data or medical target image data.” (Paragraph [0015]: “Advantageously, the embodiments provided herein overcome the above and other problems by describing a multi-omic classifier to predict responses to PD-1/PD-L1 and CTLA-4 checkpoint blockade in various clinical indications, including but not limited to non-small cell lung cancer (NSCLC), melanoma, bladder cancer, and breast cancer. In one embodiment, the classifier is developed from training data that includes diagnostic imaging scans at baseline and follow-up intervals, along with existing biomarkers, relevant clinical, molecular, demographic, response and survival data. Examples of existing biomarkers used in clinical practice include: PD-L1 expression immunohistochemistry, tumor mutation burden (TMB), mutation mismatch repair (MMR), microsatellite instability (MSI), and neutrophil-to-lympocyte ratio (NLR). Furthermore, there is early evidence suggesting that laboratory tests, such as Lactate Dehydrogenase (LDH), S100 proteins and related blood serum proteins are predictive of immunotherapy response and pseudoprogression, specifically. In the near future, features and biomarkers extracted from the microbiome are expected to play a significant role as well.”)
Vijil and Jordan are analogous in the arts because both Vijil and Jordan describe model training using medical data.
Therefore, it would be obvious to one of ordinary skill in the art at the filing date of the instant application, having the teachings of Vijil and Jordan before him or her, to modify the teachings of Vijil to include the teachings of Jordan in order to support image medical data of Jordan to increase marketability of Vijil.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Dong (US 2020/0210808 A1), part of the prior art made of record, describes the use of autoencoders with reconstruction of claim 1 through the use of an autoencoder used to reconstruct training data in paragraph [0020].
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SHANE D WOOLWINE whose telephone number is (571)272-4138. The examiner can normally be reached M-F 9:30-6:00 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, MIRANDA HUANG can be reached at (571) 270-7092. 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.
SHANE D. WOOLWINE
Primary Examiner
Art Unit 2124
/SHANE D WOOLWINE/Primary Examiner, Art Unit 2124
Prosecution Insights