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 .
Information Disclosure Statement
The information disclosure statements (IDS) submitted on 03/27/2023, are in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Rejections - 35 USC § 101
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 1-20 rejected under 35 U.S.C. 101 as being directed to a judicial exception (i.e., an abstract idea) without significantly more.
Regarding claim 1
Claim 1 – Step 1 – Is the claim to a process, machine, manufacture or composition of matter?
Yes, the claim is to a process.
Claim 1 – Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites an abstract idea.
processing, by the hardware accelerator, the input sequence of multi-modal feature vectors using a convolutional neural network comprising a stack of convolutional layers and a normalization laver following each convolutional laver except a final convolutional laver to generate a latent sequence that comprises a plurality of latent feature vectors; - recites mathematical concepts (e.g., convolution, normalization, vector transformations) and thus is an abstract idea under MPEP § 2106.04(a).
processing, by the hardware accelerator, the latent sequence of latent feature vectors using an aggregation neural network to generate an aggregated feature vector; - recites mathematical concepts (aggregation/weighting/combination of vectors) and thus is an abstract idea. See MPEP § 2106.04(a)(I).
processing, by the hardware accelerator, the aggregated feature vector using an output neural network to generate a prediction that characterizes the entity after the time window; - recites mathematical concepts and data evaluation/observation (producing a prediction from processed data) and thus is an abstract idea. See MPEP § 2106.04(a).
and reprocessing, by the hardware accelerator, the input sequence with the adjusted time window to generate an updated prediction. – recites iterative re-application of the same mathematical processing to produce an updated predicted result and is part of the abstract idea. See MPEP § 2106.04(a).
Claim 1 – Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application?
No. There are no additional elements that integrate the judicial exception into a practical application. The additional elements:
one or more computers – this recites generic computing tools. Merely invoking generic computing components to perform the abstract idea, constitutes mere instructions to apply the exception / use of a computer as a tool. See MPEP § 2106.05(f); § 2106.05(a).
and one or more storage devices storing instructions that when executed by the one or more computers cause the one or more computers to perform operations comprising – this recites generic storage and instructions. Merely storing and executing instructions on generic storage is a conventional implementation vehicle and does not provide a technological improvement or meaningful limitation on the abstract idea. This is a generic tool invocation / “apply it on a computer”. See MPEP § 2106.05(f) and MPEP § 2106.05(a).
receiving, by a hardware accelerator, an input sequence of multi-modal feature vectors characterizing an entity over a time window, wherein each multi-modal feature vector in the input sequence corresponds to a different time interval during the time window; - this recites mere data gathering / pre-solution activity, which is insignificant extra-solution activity. See MPEP § 2106.05(g). The phrase “hardware accelerator” as used here, is recited at a high level of generality and does not claim any specific accelerator mechanism that changes how the computer operates; it is therefore a generic tool/environment invocation. See MPEP § 2106.05(f); § 2106.05(a).
receiving, by the hardware accelerator, an additional multi-modal feature vector corresponding to a new time interval and in response:; - this limitation merely recites generic data gathering / insignificant extra-solution activity under MPEP § 2106.05(g), performed in a generic “hardware accelerator” environment that is not claimed as an improvement to computer functionality (MPEP § 2106.04(a)) and is merely a tool invocation (MPEP § 2106.05(f)).
adjusting the time window to incorporate the new time interval by removing a multi-model feature vector corresponding to the oldest time interval in the input sequence and adding the additional multi-modal feature vector corresponding to the new time interval; - this limitation recites routine data organization/manipulation around the abstract processing. The limitation recites a sliding-window data management operation (remove oldest / add newest) that prepares inputs for re-application of the abstract processing and is therefore insignificant extra-solution activity and/or routine data manipulation. See MPEP § 2106.05(g); § 2106.04(d).
“by a hardware accelerator” – the phrase as it appears is merely a generic tool invocation / field of use / result-oriented limitation, without any claimed accelerator mechanism that improves computer functionality. See MPEP § 2106.05(a); § 2106.05(f); § 2106.05(h).
Claim 1 – Step 2B – Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. There are no additional elements that amount to significantly more than the judicial exception. The additional elements are:
one or more computers – this limitation recites generic computing components that are well-understood, routine, and conventional (WURC) in the field. See MPEP § 2106.05(d).
and one or more storage devices storing instructions that when executed by the one or more computers cause the one or more computers to perform operations comprising – likewise, this limitation recites generic storage devices that store instructions for execution by one or more computers, which is well-understood, routine, and conventional (WURC) activity. See MPEP § 2106.05(d).
receiving, by a hardware accelerator, an input sequence of multi-modal feature vectors characterizing an entity over a time window, wherein each multi-modal feature vector in the input sequence corresponds to a different time interval during the time window; - “receiving” is routine data intake (MPEP § 2106.05(g)) and, at this level of generality, is well-understood, routine, and conventional (WURC). The recited “hardware accelerator” is recited generically, without nonconventional accelerator operation, and thus does not add significantly more. See MPEP § 2106.05(d); § 2106.05(f).
receiving, by the hardware accelerator, an additional multi-modal feature vector corresponding to a new time interval and in response:; - this recites routine data receiving (MPEP § 2106.05(g)) and generic accelerator environment (MPEP § 2106.05(f)). Both are well-understood, routine, and conventional (WURC). See MPEP § 2106.05(d).
adjusting the time window to incorporate the new time interval by removing a multi-model feature vector corresponding to the oldest time interval in the input sequence and adding the additional multi-modal feature vector corresponding to the new time interval; - this limitation recites routine data organization/manipulation used to set up re-application of the abstract processing. This amounts to insignificant extra-solution activity and well-understood, routine, and conventional (WURC) activity when claimed generically. See MPEP § 2106.05(g); § 2106.05(d).
“by a hardware accelerator” – this limitation is recited generically and is absent a nonconventional accelerator implementation. Rather it is a conventional tool usage that is well-understood, routine, and conventional (WURC) in the art. See MPEP § 2106.05(g) and § 2106.05(d).
Regarding claim 2
Claim 2 – Step 1 – Is the claim to a process, machine, manufacture or composition of matter?
Yes, the claim is to a process.
Claim 2 – Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites an abstract idea.
Claim 2 – Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application?
No. There are no additional elements that integrate the judicial exception into a practical application. The additional elements:
wherein the stack of convolutional layers that each have a respective one- dimensional kernel. – this limitation merely characterizes the type of convolution kernel (one-dimensional) used in the neural-network mathematical processing. It does not recite any specific improvement to computer functionality or the hardware accelerator (e.g., no accelerator architecture, memory hierarchy, scheduling, tiling, dataflow, or other concrete implementation that changes how the computer operates). As such, it is at most an implementation detail / field-of-use style limitation and does not meaningfully limit the judicial exception or integrate it into a practical application. See MPEP § 2106.04(d); § 2106.05(a); § 2106.05(f); § 2106.05(h).
Claim 2 – Step 2B – Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. There are no additional elements that amount to significantly more than the judicial exception. The additional elements are:
wherein the stack of convolutional layers that each have a respective one- dimensional kernel. – specifying a one-dimensional kernel for convolutional layers is a conventional modeling choice for time-series / sequential vector inputs and, as claimed, merely describes how the abstract mathematical processing is carried out. It does not reflect a nonconventional arrangement of components or a specific improvement to the operation of the computer/accelerator. Accordingly, it does not amount to an inventive concept when considered individually or with claim 1’s generic computing/accelerator execution. See MPEP § 2106.05(d) (WURC / significantly more).
Regarding claim 3
Claim 3 – Step 1 – Is the claim to a process, machine, manufacture or composition of matter?
Yes, the claim is to a process.
Claim 3 – Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites an abstract idea.
Claim 3 – Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application?
No. There are no additional elements that integrate the judicial exception into a practical application. The additional elements:
wherein the aggregation neural network is a recurrent neural network. – reciting that the aggregation neural network “is a recurrent neural network” merely specifies a particular mathematical/modeling technique for carrying out the abstract processing (sequence aggregation). It does not recite a specific improvement to computer functionality (MPEP § 2106.05(a)), nor does it impose a meaningful limitation on the judicial exception beyond generally stating ‘do the abstract processing using an RNN’. This is an implementation detail / field-of-use type limitation (MPEP § 2106.05(h)) and, to the extent it is simply performing the abstract idea on generic computing/accelerator architecture, it is use of a computer as a tool / mere instructions to apply the exception (MPEP § 2106.05(f)).
Claim 3 – Step 2B – Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. There are no additional elements that amount to significantly more than the judicial exception. The additional elements are:
wherein the aggregation neural network is a recurrent neural network. – a recurrent neural network is a known, conventional neural-network architecture for processing sequential/temporal data. As claimed here, it is recited at a high level (without any nonconventional RNN structure, training regime, state update mechanism, memory gating detail, or accelerator-level implementation that improves computing operation), and therefore does not constitute an inventive concept when considered alone or in combination with claim 1’s generic computing/accelerator execution. See MPEP § 2106.05(d).
Regarding claim 4
Claim 4 – Step 1 – Is the claim to a process, machine, manufacture or composition of matter?
Yes, the claim is to a process.
Claim 4 – Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites an abstract idea.
Claim 4 – Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application?
No. There are no additional elements that integrate the judicial exception into a practical application. The additional elements:
wherein the output neural network comprises one or more fully-connected layers followed by an output layer. – this limitation merely recites a conventional neural-network architecture detail for implementing the abstract prediction computation. It does not recite a specific improvement to computer functionality (MPEP § 2106.05(a)), nor does it meaningfully limit the judicial exception beyond specifying a generic implementation choice for the output network (implementation detail / field-of-use, MPEP § 2106.05(h)). To the extent it is executed on generic “one or more computers” / “hardware accelerator”, it is use of a computer as a tool / mere instructions to apply the exception (MPEP § 2106.05(f)).
Claim 4 – Step 2B – Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. There are no additional elements that amount to significantly more than the judicial exception. The additional elements are:
wherein the output neural network comprises one or more fully-connected layers followed by an output layer. – fully-connected layers followed by an output layer are well-known, conventional building blocks for neural-network classifiers/regressions. As claimed, this limitation is high-level and does not recite any nonconventional arrangement or a particular technical implementation that improves operation of the computer/accelerator. Thus, it does not add an inventive concept when considered individually or in combination with claim 1. See MPEP § 2106.05(d) (WURC / “significantly more”).
Regarding claim 5
Claim 5 – Step 1 – Is the claim to a process, machine, manufacture or composition of matter?
Yes, the claim is to a process.
Claim 5 – Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites an abstract idea.
Claim 5 – Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application?
No. There are no additional elements that integrate the judicial exception into a practical application. The additional elements:
obtaining data characterizing the entity from a plurality of different data streams; – obtaining data is data gathering and constitutes insignificant extra-solution activity. See MPEP § 2106.05(g).
generating the multi-modal feature vectors in the input sequence by converting the data characterizing the entity into a standardized format. – this is routine data manipulation/organization that prepares data for the abstract processing (CNN/RNN/output NN prediction) and likewise is insignificant extra-solution activity. See MPEP § 2106.05(g); § 2106.04(d).
Claim 5 – Step 2B – Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. There are no additional elements that amount to significantly more than the judicial exception. The additional elements are:
obtaining data characterizing the entity from a plurality of different data streams; - this is a routine input-collection operation that is well-understood, routine, and conventional (WURC) when recited at this level of generality, and therefore does not supply “significantly more”. See MPEP § 2106.05(g).
generating the multi-modal feature vectors in the input sequence by converting the data characterizing the entity into a standardized format. – this recites conventional preprocessing step used to feed downstream analytics/ML, and is recited at a functional level that does not amount to an inventive concept, either alone or in combination with the remaining generic-computing implementation. MPEP § 2106.05(d) (WURC / “significantly more”); § 2106.05(g).
Regarding claim 6
Claim 6 – Step 1 – Is the claim to a process, machine, manufacture or composition of matter?
Yes, the claim is to a process.
Claim 6 – Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites an abstract idea.
Claim 6 – Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application?
No. There are no additional elements that integrate the judicial exception into a practical application. The additional elements:
wherein generating each multi-modal feature vector comprises: identifying respective features of each of a plurality of feature types that characterize the entity during the corresponding time interval for the multi-modal feature vector; - this is a high-level data analysis/feature extraction step. It does not recite any specific improvement to computer functionality (e.g., a specific sensor/control mechanism, a specific computer-memory/dataflow improvement, or any unconventional implementation that changes how the computer operates). It is part of preparing/evaluating data for the abstract model pipeline and is at most insignificant extra-solution activity. MPEP § 2106.05(g); and it does not satisfy the “improvement to computer functionality” consideration under MPEP § 2106.05(a).
and for each feature type, adding the identified respective features of the feature type to one or more entries of the multi-modal feature vector that correspond to the feature type. – this limitation is data organization/manipulation, i.e., mapping/coding identified features into vector entries by type. This is a conventional pre-processing step for downstream analytics/ML and is also insignificant extra-solution activity (organizing data so it can be processed by the abstract idea). MPEP § 2106.05(g); and is not a claimed improvement to computing technology under MPEP § 2106.05(a).
Claim 6 – Step 2B – Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. There are no additional elements that amount to significantly more than the judicial exception. The additional elements are:
wherein generating each multi-modal feature vector comprises: identifying respective features of each of a plurality of feature types that characterize the entity during the corresponding time interval for the multi-modal feature vector; - identifying features by type from data is a conventional feature-engineering / data preprocessing concept when claimed at this level of generality and does not supply “significantly more”. See MPEP § 2106.05(d) (WURC / “significantly more”).
and for each feature type, adding the identified respective features of the feature type to one or more entries of the multi-modal feature vector that correspond to the feature type. – populating entries of a vector (standard encoding/representation) is routine data formatting/organization ancillary to the abstract model processing, and is well-understood, routine, and conventional (WURC) / insignificant extra-solution activity absent any nonconventional mechanism. MPEP § 2106.05(d) and § 2106.05(g).
Regarding claim 7
Claim 7 – Step 1 – Is the claim to a process, machine, manufacture or composition of matter?
Yes, the claim is to a process.
Claim 7 – Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites an abstract idea.
Claim 7 – Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application?
No. There are no additional elements that integrate the judicial exception into a practical application. The additional elements:
wherein the convolutional neural network, the aggregation neural network, and the output neural network have been jointly trained on training data that includes a plurality of training input sequences and, for each training input sequence, a corresponding ground truth outcome. – this limitation adds a training context (joint training using training input sequences and corresponding ground truth outcomes), which merely specifies that the model used to perform the abstract prediction processing is trained in a conventional supervised-learning manner. This is an implementation detail describing how the abstract mathematical model is obtained/parameterized, which does not meaningfully limit the judicial exception. This is analogous to reciting “apply it with trained parameters”. See MPEP § 2106.04(d) and § 2106.05(h); and to the extent the limitation is merely “apply the exception” in a generic environment, MPEP §2106.05(f).
Claim 7 – Step 2B – Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. There are no additional elements that amount to significantly more than the judicial exception. The additional elements are:
wherein the convolutional neural network, the aggregation neural network, and the output neural network have been jointly trained on training data that includes a plurality of training input sequences and, for each training input sequence, a corresponding ground truth outcome. – joint training using supervised learning data with “training input sequences” and “ground truth outcome” labels is a conventional ML training approach, and as claimed, is recited at a high level without any nonconventional training procedure, objective function, constraint, hyperparameter regime, or computer/accelerator-level improvement. As such, it is well-understood, routine, and conventional (WURC) and does not supply an inventive concept either alone or in combination with claim 1’s generic computing/accelerator execution. See MPEP § 2106.05(d).
Regarding claim 8
Claim 8 – Step 1 – Is the claim to a process, machine, manufacture or composition of matter?
Yes, the claim is to a process.
Claim 8 – Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites an abstract idea.
Claim 8 – Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application?
No. There are no additional elements that integrate the judicial exception into a practical application. The additional elements:
wherein the entity is a financial asset, and wherein each multi-modal feature vector comprises technical analysis features and sentiment analysis features. – this recites a field-of-use limitation (applying the abstract predictive processing to the financial domain). Limiting an abstract idea to a particular field of use does not integrate the exception into a practical application. MPEP § 2106.05(h); 2106.04(d). The limitation further specifies types of input information/features used in the abstract evaluation/prediction, i.e., it is at most data selection / data gathering / data preparation for the abstract processing and constitutes insignificant extra-solution activity. See MPEP § 2106.05(g).
Claim 8 – Step 2B – Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. There are no additional elements that amount to significantly more than the judicial exception. The additional elements are:
wherein the entity is a financial asset, and wherein each multi-modal feature vector comprises technical analysis features and sentiment analysis features. – applying the abstract predictive processing to a financial asset is a field-of-use limitation and does not add an inventive concept. MPEP § 2106.05(h) § 2106.05(d). Using technical analysis features and sentiment analysis features as inputs is a conventional choice of data/features for financial prediction and, as claimed at this high level, amounts to routine data selection/characterization ancillary to the abstract idea. This does not constitute “significantly more”, either alone or as an ordered combination with claim 1’s generic implementation. MPEP § 2106.05(d) and § 2106.05(g).
Regarding claim 9
Claim 9 – Step 1 – Is the claim to a process, machine, manufacture or composition of matter?
Yes, the claim is to a process.
Claim 9 – Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites an abstract idea.
Claim 9 – Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application?
No. There are no additional elements that integrate the judicial exception into a practical application. The additional elements:
wherein each multi-modal feature vector further comprises fundamental analysis features. – this limitation merely adds another category of input information (“fundamental analysis features”) used in the abstract predictive processing. That is, at most, data selection / data gathering / data preparation for the judicial exception and constitutes insignificant extra-solution activity. See MPEP § 2106.05(g). It does not recite any improvement to computer functionality or a particular technological implementation beyond specifying additional data content. See MPEP § 2106.05(a); § 2106.04(d).
Claim 9 – Step 2B – Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. There are no additional elements that amount to significantly more than the judicial exception. The additional elements are:
wherein each multi-modal feature vector further comprises fundamental analysis features. – adding “fundamental analysis features” is a conventional data-feature choice in the financial prediction domain and, as claimed at a high level, amounts to routine data selection/characterization ancillary to the abstract idea. This does not supply an inventive concept either alone or in ordered combination with claims 1 and 8. See MPEP § 2106.05(d) and § 2106.05(g).
Regarding claim 10
Claim 10 – Step 1 – Is the claim to a process, machine, manufacture or composition of matter?
Yes, the claim is to a process.
Claim 10 – Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites an abstract idea.
Claim 10 – Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application?
No. There are no additional elements that integrate the judicial exception into a practical application. The additional elements:
wherein the prediction characterizes a predicted trading behavior of the financial asset at an end of a next time interval after the end of the time window. – this limitation merely specifies the field-of-use/result of the prediction (trading behavior prediction at a particular feature time). It does not recite any additional technical implementation that improves computer functionality or that applies the abstract idea in a manner that effects a technological improvement. See MPEP § 2106.05(h), MPEP § 2106.05(a).
Claim 10 – Step 2B – Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. There are no additional elements that amount to significantly more than the judicial exception. The additional elements are:
wherein the prediction characterizes a predicted trading behavior of the financial asset at an end of a next time interval after the end of the time window. – specifying that the model output is a prediction of trading behavior at a particular feature time is a well-understood, routine, and conventional (WURC) statement of an intended use/output in the finance domain and does not add a nonconventional technical mechanism. As such, it does not supply an inventive concept either alone or in combination with the generic implementation of claim 1. See MPEP § 2106.05(d); MPEP § 2106.05(h).
Regarding claim 11
Claim 11 – Step 1 – Is the claim to a process, machine, manufacture or composition of matter?
Yes, the claim is to a process.
Claim 11 – Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites an abstract idea.
processing, by the hardware accelerator, the input sequence of multi-modal feature vectors using a convolutional neural network comprising a stack of convolutional layers and a normalization laver following each convolutional laver except a final convolutional laver to generate a latent sequence that comprises a plurality of latent feature vectors; - this limitation recites mathematical concepts (convolution/normalization/vector transformations) used to generate latent vectors, which is an abstract idea under MPEP § 2106.04(a).
processing, by the hardware accelerator, the latent sequence of latent feature vectors using an aggregation neural network to generate an aggregated feature vector; - this limitation recites mathematical processing (aggregation/combination of vectors) via a neural network) and is part of the abstract idea (MPEP § 2106.04(a)).
and processing, by the hardware accelerator, the aggregated feature vector using an output neural network to generate a prediction that characterizes the entity after the time window; - this limitation recites mathematical processing to generate a prediction (data evaluation to produce an outcome), which is an abstract idea under MPEP § 2106.04(a).
and reprocessing, by the hardware accelerator, the input sequence with the adjusted time window to generate an updated prediction. – this limitation recites re-application of the same mathematical processing to generate an updated predicted result, which is part of the abstract idea under MPEP § 2106.04(a).
Claim 11 – Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application?
No. There are no additional elements that integrate the judicial exception into a practical application. The additional elements:
A method performed by one or more computers, the method comprising: - this is use of a computer as a tool / generic implementation and does not integrate the abstract idea into a practical application under MPEP § 210.05(f). It does not recite a specific improvement to computer functionality (MPEP § 2106.05(a)).
receiving, by a hardware accelerator, an input sequence of multi-modal feature vectors characterizing an entity over a time window, wherein each multi-modal feature vector in the input sequence corresponds to a different time interval during the time window; - this recites mere data gathering / pre-solution activity, which is insignificant extra-solution activity. See MPEP § 2106.05(g). The phrase “hardware accelerator” as used here, is recited at a high level of generality and does not claim any specific accelerator mechanism that changes how the computer operates; it is therefore a generic tool/environment invocation. See MPEP § 2106.05(f); § 2106.05(a).
receiving, by the hardware accelerator, an additional multi-modal feature vector corresponding to a new time interval and in response:; - this limitation merely recites generic data gathering / insignificant extra-solution activity under MPEP § 2106.05(g), performed in a generic “hardware accelerator” environment that is not claimed as an improvement to computer functionality (MPEP § 2106.04(a)) and is merely a tool invocation (MPEP § 2106.05(f)).
adjusting the time window to incorporate the new time interval by removing a multi-model feature vector corresponding to the oldest time interval in the input sequence and adding the additional multi-modal feature vector corresponding to the new time interval; - this limitation recites routine data organization/manipulation around the abstract processing. The limitation recites a sliding-window data management operation (remove oldest / add newest) that prepares inputs for re-application of the abstract processing and is therefore insignificant extra-solution activity and/or routine data manipulation. See MPEP § 2106.05(g); § 2106.04(d).
Claim 11 – Step 2B – Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. There are no additional elements that amount to significantly more than the judicial exception. The additional elements are:
A method performed by one or more computers, the method comprising: - a method “performed” by one or more computers is a well-understood, routine, and conventional (WURC) activity and does not add “significantly more” than the abstract idea. See MPEP § 2106.05(d).
receiving, by a hardware accelerator, an input sequence of multi-modal feature vectors characterizing an entity over a time window, wherein each multi-modal feature vector in the input sequence corresponds to a different time interval during the time window; - “receiving” is routine data intake (MPEP § 2106.05(g)) and, at this level of generality, is well-understood, routine, and conventional (WURC). The recited “hardware accelerator” is recited generically, without nonconventional accelerator operation, and thus does not add significantly more. See MPEP § 2106.05(d); § 2106.05(f).
receiving, by the hardware accelerator, an additional multi-modal feature vector corresponding to a new time interval and in response:; - this limitation merely recites generic data gathering / insignificant extra-solution activity under MPEP § 2106.05(g), performed in a generic “hardware accelerator” environment that is not claimed as an improvement to computer functionality (MPEP § 2106.04(a)) and is merely a tool invocation (MPEP § 2106.05(f)).
adjusting the time window to incorporate the new time interval by removing a multi-model feature vector corresponding to the oldest time interval in the input sequence and adding the additional multi-modal feature vector corresponding to the new time interval; - this limitation recites routine data organization/manipulation around the abstract processing. The limitation recites a sliding-window data management operation (remove oldest / add newest) that prepares inputs for re-application of the abstract processing and is therefore insignificant extra-solution activity and/or routine data manipulation. See MPEP § 2106.05(g); § 2106.04(d).
Regarding claim 12
Claim 12 – Step 1 – Is the claim to a process, machine, manufacture or composition of matter?
Yes, the claim is to a process.
Claim 12 – Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites an abstract idea.
Claim 12 – Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application?
No. There are no additional elements that integrate the judicial exception into a practical application. The additional elements:
wherein the stack of convolutional layers that each have a respective one-dimensional kernel. – this limitation merely specifies the type of convolution kernel used in the mathematical processing. It does not recite a specific improvement to computer functionality (MPEP § 2106.05(a)), and it is at most an implementation detail / field-of-use style limitation describing how the abstract idea is performed (MPEP § 2106.05(h)). It therefore does not integrate the exception into a practical application. MPEP § 2106.04(d).
Claim 12 – Step 2B – Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. There are no additional elements that amount to significantly more than the judicial exception. The additional elements are:
wherein the stack of convolutional layers that each have a respective one-dimensional kernel. – specifying one-dimensional kernels is a well-understood, routine, and conventional (WURC) CNN design choice for sequential/time-series inputs and, as claimed at a high level, does not supply an inventive concept. See MPEP § 2106.05(d).
Regarding claims 13-19 (dependent method claims)
Each of claims 13-19 is a dependent method claim that is analogous to a previously analyzed dependent system claim (i.e., each adds only a particular NN architecture choice, training context, data acquisition/feature-generation detail, or finance-specific field-of-use / feature-content limitation to the already analyzed abstract prediction pipeline of independent claim 11 / claim 1).
For each of claims 13-19, the additional functional language (e.g., specifying an RNN for aggregation, specifying fully-connected layers, specifying training data/ground truth, obtaining data streams and converting to standardized format, identifying feature types and inserting them into vector entries, specifying “financial asset” and “technical/sentiment/fundamental” feature content, and/or specifying a trading-behavior characterization of the prediction) does not materially change the § 101 analysis.
Accordingly, claims 13-19 are rejected under 35 U.S.C. § 101 for the same reasons as claim 11 (and the corresponding dependent claim analyses already provided for the analogous claim-set).
Regarding independent claim 20 (analogous to claim 11 / claim 1)
Independent claim 20 is the analog of independent claim 11 (and system claim 1) and recites the same core operations: receiving time-windowed multi-modal feature vectors, processing via CNN (including the normalization-layer stack detail), aggregating to an aggregated feature vector, generating a prediction, then receiving a new interval, sliding the window, and reprocessing to update the prediction, implemented using generic computing/accelerator execution.
This form does not materially change the eligibility analysis and accordingly, independent claim 20 is rejected under 35 U.S.C. § 101 for the same reasons as independent claim 11 / claim 1.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Patrick Cheung (US20210056413A1) in view of Xuan-Hong Dang (US11915123B2) and in further view of Chaowei Tan (EP3611699A1).
Regarding claim 1, Cheung in view of Dang and in further view of Tan, teach a system comprising:
one or more computers, and one or more storage devices storing instructions that when executed by the one or more computers cause the one or more computers to perform operations comprising:
receiving, by a hardware accelerator, an input sequence of multi-modal feature vectors characterizing an entity over a time window, wherein each multi-modal feature vector in the input sequence corresponds to a different time interval during the time window; - Cheung teaches this limitation. Cheung teaches execution on hardware including FPGA/ASIC (i.e., accelerator hardware):
“The processor 91 may be any hardware device capable of executing instructions stored in memory or storage or otherwise processing data. As such, the processor 91 may include a microprocessor, field programmable gate array (FPGA), application-specific integrated circuit (ASIC), or other similar devices.” (Cheung, p. 8, ¶[0108])
Cheung in addition, teaches time-series physiological data and bounded windows, for example:
“(e.g. a time series of vital sign measurements and laboratory test results).” (Cheung, p. 4, ¶[0046])
“truncated to keep the last 100 measurement points while others which have fewer
than 100 points were zero-padded to maintain 100 points in length.” (Cheung, p. 8, ¶[0099])
These teachings correspond to receiving a sequence over a time window with per-interval samples.
processing, by the hardware accelerator, the input sequence of multi-modal feature vectors using a convolutional neural network comprising a stack of convolutional layers ; - Cheung teaches this limitation in part. Cheung teaches convolution producing feature-map outputs and pooling into a predictive representation:
“A max pooling operation is applied to each filter to extract a scalar y(j)=MAX(Y(:, j)).” (Cheung, p. 7, ¶[0096])
This evidences per-filter feature maps prior to pooling, i.e., latent features over the input positions).
processing, by the hardware accelerator, the aggregated feature vector using an output neural network to generate a prediction that characterizes the entity after the time window; - Cheung teaches this limitation. Cheung teaches a final network head producing a prediction:
“fed to a final connected network … followed by a sigmoid function … to produce prediction” (Cheung, p. 7, ¶[0096])
Cheung’s prediction is explicitly about the future period after the observation window:
“predict if the HF patient would be readmitted within the next 30 days.” (Cheung, p. 7, ¶[0098])
adjusting the time window to incorporate the new time interval by removing a multi-model feature vector corresponding to the oldest time interval in the input sequence ; - Cheung teaches this limitation in part. Cheung teaches bounded/fixed-length time windows:
“Those resampled time series which have more than 100 points were truncated to keep the last 100 measurement points while others which have fewer than 100 points were zero-padded to maintain 100 points in length.” (Cheung, p. 8, ¶[0099])
Cheung does not teach:
processing, by the hardware accelerator, the latent sequence of latent feature vectors using an aggregation neural network to generate an aggregated feature vector;
receiving, by the hardware accelerator, an additional multi-modal feature vector corresponding to a new time interval and in response:;
“… and adding the additional multi-modal feature vector corresponding to the new time interval;”
Dang, however, teaches these limitations:
processing, by the hardware accelerator, the latent sequence of latent feature vectors using an aggregation neural network to generate an aggregated feature vector; - Dang discloses an:
“Asynchronous RNN (AsyncLSTM) that iteratively fuses encoded features from multiple different data modalities,” (Dang, col. 4, lines 5-6)
Dang further teaches aggregation into a single vector:
“Encodings from the text data (114) and encodings from the time series data (132) are combined in a concatenation layer (150) to aggregate the encoding into a single vectors” (Dang, col. 5, lines 24-27)
Thus, Dang teaches using an aggregation neural network to produce an aggregated feature vector.
receiving, by the hardware accelerator, an additional multi-modal feature vector corresponding to a new time interval and in response:; - Dang teaches per-timestep sequence ingestion/processing:
“At each timestep I in the series sequence (220) …” (Dang, col. 6, lines 6-7)
This supports receipt/handling of new time-interval data in a running sequence.
“… and adding the additional multi-modal feature vector corresponding to the new time interval;” - Dang teaches advancing through a time-series by timestep:
“As the AsyncLSTM advances in the time series sequence” (Dang, col. 5, lines 46-47)
Given Cheung’s fixed-length “last N points” window and Dang’s timestep-wise advancement, it would have been obvious and predictable implementation detail to maintain the fixed window length as new intervals arrive by dropping the oldest interval and adding the newest (standard sliding-window update to preserve bounded compute/latency).
and reprocessing, by the hardware accelerator, the input sequence with the adjusted time window to generate an updated prediction. – Dang teaches the concept of incorporating data from a new time interval (i.e., the “new vector comes in”):
“incorporates … to every step it performs reasoning … in the time series” (Dang, col. 6, lines 5-6)
Cheung teaches producing predictions from the processed representation via the final network head. Combining Cheung’s fixed-window inference with Dang’s timestep-wise sequential reasoning makes it obvious to re-run inference for the updated window to output an updated prediction (predictable results: “current” prediction as new interval data arrives).
Cheung nor Dang teach:
“and a normalization laver following each convolutional laver except a final convolutional laver to generate a latent sequence that comprises a plurality of latent feature vectors”
Tan, however, teaches these limitations:
“and a normalization laver following each convolutional laver except a final convolutional laver to generate a latent sequence that comprises a plurality of latent feature vectors” – Tan teaches convolution layers each followed by batch normalization:
“each of the volumetric convolutional layers precedes a batch normalization layer.” (Tan, p. 2, col. 2, lines 51-52)
It would have been obvious to a POSITA to incorporate Tan’s post-convolution normalization (e.g., batch normalization) into Cheung’s convolutional stack to improve stability/training/performance of CNN feature extraction while keeping the final convolution un-normalized as a routine design choice for the terminal conv output (predictable result).
Cheung provides a multimodal neural prediction system implemented on processor hardware including FPGA/ASIC and teaches bounded time-series windows and CNN-based feature extraction producing a prediction. Dang teaches an RNN/AsyncLSTM aggregation/fusion network that iteratively fuses multimodal encodings over time and reasons per timestep, producing aggregated vectors and outputs. Tan teaches placing normalization (batch normalization) after convolution layers. It would have been obvious to combine these teachings to obtain the claimed CNN + (normalized conv stack) + recurrent aggregation + prediction pipeline operating on bounded time windows with timestep updates, with a reasonable expectation of success.
Regarding claim 2, Cheung in view of Dang, and in further view of Tan, teach the system of claim 1,
wherein the stack convolutional layers that each have a respective one- dimensional kernel. – Cheung teaches this limitation. Cheung teaches a plurality/stack of convolutional layers and explicitly identifies kernel size for those layers, in the context of applying convolution to time series / sequences:
“a two-stage convolutional neural network 150b employs a convolutional module including two stacked convolutional neural network stages S151 and S152 (FIG. SA) applying convolution and max pooling” (Cheung, p. 6, ¶[0087])
“each time series is considered as a channel input … where Cl and C2 are the numbers of convolutional filters in stages S151 and S152, Size is the kernel size …” (Cheung, p. 6, ¶[0086])
Cheung also teaches using specific filter widths (i.e., kernel widths) for the CNN:
“The three filter widths were set to 3, 4 and 5 with 20 filters each for CNN.” (Cheung, p. 8, ¶[0101])
And Cheung explains that its convolution is performed on a sequence of inputs (i.e., 1-D along the sequence dimension):
“Convolution stage S143 is performed on the sequence of inputs …” (Cheung, p. 6, ¶[0081])
Thus, Cheung expressly teaches (i) multiple/stacked convolutional stages, and (ii) each such convolution uses a kernel size / filter width, applied to a sequence/time-series input.
Regarding claim 3, Cheung in view of Dang and in further view of Tan, teach the system of claim 1,
wherein the aggregation neural network is a recurrent neural network. – Cheung does not teach this limitation. Cheung teaches an “aggregation” stage combining feature vectors from multiple networks, but does not specify that the aggregation network is an RNN. Dang, however, expressly teaches using a recurrent neural network (RNN) to analyze encoded multi-modal sets:
“leverage an artificial recurrent neural network (RNN) to analyze the encoded first and second data sets, including iteratively and asynchronously fuse the first and second encodings” (Dang, col. 18, lines 42-45)
This maps directly to the claim limitation because Dang’s RNN is used to fuse / combine (i.e., aggregate) the encodings/vectors from multiple modalities (first and second data sets).
It would have been obvious to a person having ordinary skill in the art to implement Cheung’s aggregation neural network as the RNN taught by Dang, because Dang expressly teaches RNN-based fusion of multimodal encodings/vectors and RNNs are a known design choice for modeling temporal dependencies and aggregating sequential/multimodal representations, yielding predictable results.
Regarding claim 4, Cheung in view of Dang and in further view of Tan, teach the system of claim 1,
wherein the output neural network comprises one or more fully-connected layers followed by an output layer. – Cheung teaches this limitation. Cheung discloses an output/prediction network having a fully connected stage followed by a sigmoid function stage (i.e., an output layer):
“cross-modal convolutional neural network 160a has a neural architecture employing a module including … a fully connected stage S162 and a sigmoid function stage S163 … to produce a prediction 117a” (Cheung, p. 7, ¶[0094])
“Scalars from
Ν
x
c
o
n
v
filters are then concatenated to form a compact predictive feature vector which is then fed to a final connected network at stage S162 followed by a sigmoid function at stage S163 to produce prediction 117a.” (Cheung, p. 7, ¶[0096])
These passages teach that the output neural network includes at least one fully-connected layer followed by an output layer that produces the prediction.
Cheung expressly teaches the limitation of claim 4. Therefore, claim 4 is unpatentable for the reasons set forth for claim 1.
Regarding claim 5, Cheung in view of Dang and in further view of Tan, teach the system of claim 1, the operations further comprising:
obtaining data characterizing the entity from a plurality of different data streams; - Cheung does not teach this limitation. Dang, however, teaches obtaining input data from multiple different streams/modalities. For example, Dang teaches different modalities are provided as different feeds/streams, including a textual stream and a time-series stream:
“the first data feed,
f
e
e
d
A
(452A), represents a time-stamped textual modality, … and the second data feed,
f
e
e
d
B
(452B), represents time series data modality.” (Dang, col. 8, lines 58-62)
“Input … includes time series data from the second input … and textual data from the first input” (Dang, col. 5, 51-53)
generating the multi-modal feature vectors in the input sequence by converting the data characterizing the entity into a standardized format. – Cheung does not teach this limitation. Dang, however, teaches converting heterogeneous, different-format multimodal inputs into vectors (i.e., encoding into a common vector format):
“the multi-modal data set comprising data in different formats from two or more modalities … encode the first data set into first encodings comprising one or more first vectors and encode the second data set into second encodings comprising one or more second vectors;” (Dang, col. 19, lines 16-18, 22-25)
“The first data feed manager functions to encode the first data set into a first set of vectors.” (Dang, col. 1, lines 34-35)
Thus, Dang expressly teaches the limitations of claim 5. Therefore, claim 5 is unpatentable for the reasons set forth for claim 1.
Regarding claim 6, Cheung in view of Dang and in further view of Tan, teach the system of claim 5, wherein generating each multi-modal feature vector comprises:
identifying respective features of each of a plurality of feature types that characterize the entity during the corresponding time interval for the multi-modal feature vector; - Cheung does not teach this limitation. Dang, however, teaches identifying/deriving modality-specific (feature-type-specific) vectors on a per-time-interval basis. For example, Dang discloses that the model receives multiple modalities (text and time series), and that the textual items are collected/processed at time period t, i.e., during a corresponding time interval:
“with n being the total number of news stories collected at the time stamp t. Each vector is represented for a textual article, such as news story, or a sequence of related words.” (Dang, col. 5, lines 4-7)
“With respect to the time series data, an input sample (220) at each time point t is a sequence of m values …” (Dang, col. 5, lines 54-56)
Dang also makes clear these are different modalities being encoded/handled as vectors (i.e., distinct feature types being identified/extracted per interval):
“the AsynchLSTM receives two modalities of data, including … text data, and … numerical time series data…” (Dang, col. 5, lines 33-37)
These disclosures teach that, for a given time period/time point t (corresponding time interval), the system processes distinct feature types/modalities (text/time-series) into corresponding vectors (i.e., identifying respective features of each of the plurality of feature types for that interval).
for each feature type, adding the identified respective features of the feature type to one or more entries of the multi-modal feature vector that correspond to the feature type. – Cheung teaches this limitation. Cheung teaches producing and retaining separate feature vectors corresponding to different data types (feature types), and then combining them at an upper layer:
“at a lower neural network layer … input the encoded data … to produce an encoded feature vector, (2) input the embedded data … to output an embedded feature vector, and (3) input the sampled data into a sampled neural network to output an sampled feature vector, and at an upper neural network layer, ( 4) input at least two of the encoded feature vector, the embedded feature vector and the sampled feature vector into a convolutional neural network…” (Cheung, p.2, ¶[0009])
This disclosure supports the claimed concept of, for each feature type, placing the identified features into corresponding “entries” (i.e., structured components/slots) of a multimodal vector representation - here, the encoded feature vector / embedded feature vector / sampled feature vector corresponding to different feature types – prior to combination/fusion.
Given Dang’s per-time-interval modality-specific vector formation (text vectors at time stamp t; time-series sample at time point t) and Cheung’s structured per-data feature vector organization (encoded/embedded/sampled feature vectors), it would have been obvious to combine Dang’s modality/time-interval identification with Cheung’s per-feature-type vector assembly so that, for each time interval, the respective feature-type features are placed into corresponding entries/components of the multimodal feature vector to preserve semantic alignment for downstream fusion/prediction.
Regarding claim 7, Cheung in view of Dang and in further view of Tan, teach the system of claim 1,
wherein . – Cheung teaches this limitation in part. Cheung expressly teaches training using many training inputs and corresponding ground truth labels. For example, Cheung’s explicit “training” disclosure:
“batch size of 256 for training the model. Binary cross-entropy was used as the loss function to adjust the weights. Training was stopped when no further improvement on the validation loss is found after 25 epochs.” (Cheung, p. 8, ¶[0102])
Cheung frames the prediction target as 30-day readmission:
“predict if the HF patient would be readmitted within the next 30 days.” (Cheung, p. 7, ¶[0098])
And gives readmissions occurring within 30 days after discharge as a counted outcome:
“The AXCNN was applied to a 30-day unplanned readmission data for heart failure (HF) collected from a large hospital system in Arizona, United States. The dataset consisted of patient encounter information for 6730 HF patients of age 18 or over (mean: 72.7, std:
14.4), 60% are males, between October 2015 and June 2017. Among them 853 patients have at least a readmission within 30 days after discharge” (Cheung, p. 7, ¶[0098])
Cheung does not teach this portion of the claim limitation:
“… the convolutional neural network, the aggregation neural network, and the output neural network have been jointly trained on … “
Dang, however, teaches this limitation:
“… the convolutional neural network, the aggregation neural network, and the output neural network have been jointly trained on … “ – Dang expressly discloses joint training of network parameters with the overall model:
“are the network parameters jointly trained with the entire model.” (Dang, col. 6, lines 2-3)
This provides direct support for the “jointly trained” concept (i.e., end-to-end training across model components).
Cheung teaches training a predictive neural network using a training set with corresponding ground truth labels/outcomes. Dang teaches that network parameters may be “jointly trained with the entire model”. Therefore, it would have been obvious to a POSITA to jointly train the convolutional neural network, aggregation neural network, and output neural network of the combined Cheung/Dang system using training input sequences and corresponding ground truth outcomes, as such end-to-end and supervised training of multi-component neural architectures was a known technique yielding predictable results (i.e., coordinated optimization of the feature extractor, fusion/aggregation, and prediction head).
Regarding claim 8, Cheung in view of Dang and in further view of Tan, teach the system of claim 1,
wherein the entity is a financial asset – Cheung does not teach this limitation. Dang, however, teaches this limitation. Dang expressly applies its multimodal system to a stock / financial market use case:
“In the venue of financial market use case … adjusted close price of a stock.” (Dang, col. 9, lines 43-44)
“… a third data feed (506) is in time series data in the form of stock exchange data.” (Dang, col. 10, lines 12-13)
These disclosures teach that the entity being characterized/predicted is a financial asset (e.g., a stock) and that stock-market data (stock exchange data) is an input modality.
and wherein each multi-modal feature vector comprises technical analysis features and sentiment analysis features. – Cheung does not teach this limitation. Dang, however, teaches technical/price-derived features used to compute a market prediction, including explicit price-differential computations:
“A market predication is computed via a second order differential of an adjusted close price of a stock.” (Dang, col. 9, lines 45-46)
“… time series data directed to the price of the stock” (Dang, col. 9, lines 48-49)
“… stock exchange data.” (Dang, col. 10, line 13)
These teachings correspond to technical-analysis-type inputs/features derived from stock price time series (adjusted close price; price changes; stock exchange data).
In addition, Dang expressly defines and uses market sentiment (optimistic/pessimistic) in the financial embodiment:
“Market sentiment for day t+1 is optimistic … Market sentiment for t+1 is pessimistic …” (Dang, col. 9, lines 52-54)
Dang expressly teaches the limitations of claim 8. Therefore, claim 8 is unpatentable for the reasons set forth for claim 1.
Regarding claim 9, Cheung in view of Dang and in further view of Tan, teach the system of claim 8,
wherein each multi-modal feature vector further comprises fundamental analysis features. – Cheung does not teach this limitation. Dang, however, teaches incorporating “quarterly reports” (a fundamental-information modality) into the multimodal vector representation and expressly discloses that quarterly reports are provided as an input modality in the financial-market embodiment:
“As shown, quarterly reports are represented in a first feed (502) and news articles are represented in a second feed (504). Both the quarterly reports and the news articles are textual data, and are received by textual module (512) …” (Dang, col. 9-10, lines 65-67, 1-2)
Dang further teaches that this textual input (including quarterly reports) is converted into a representation vector (i.e., a feature vector representation):
“The textual module (512) learns semantic dependencies among words present in the first feed (502) and the second feed (504) and aggregates them into a representative vector for each document …” (Dang, col. 10, lines 3-6)
Dang also teaches that outputs from the financial-market architecture fuse information associated with the financial time-series behaviors/performance with the news-related content:
“Output (574) … is directed at news filtering … in association with stock time-series behaviors and performance (572). More specifically, the news (570) and the performance behavior (572) are fused into the output (574).” (Dang, col. 10, lines 16-21)
It would have been obvious to a POSITA to include quarterly-report-derived features (fundamental information) in the multi-modal feature vectors of the Cheung/Dang/Tan combination in order to enrich the representation of the financial asset with fundamental information for improved prediction accuracy, which is a predictable use of known categories of financial data within multimodal prediction pipelines.
Regarding claim 10, Cheung in view of Dang and in further view of Tan, teach the system of claim 8,
wherein the prediction characterizes a predicted trading behavior of the financial asset at an end of a next time interval after the end of the time window. – Cheung does not teach this limitation. Dang, however, teaches predicting future (next-interval) behavior of a financial asset (e.g., stock market sentiment / behavior) and expressly teaches predicting next-day/next-interval market behavior (market sentiment) for a stock:
“Market sentiment for day t+1 is optimistic … Market sentiment for t+1 is pessimistic …” (Dang, col. 9, lines 52-54)
This passage teaches a prediction directed to a financial asset/market at a next time interval (t+1) following the current time t (i.e., after the time window ending at time t).
Dang also characterizes the output as a behavior label for the time series over a defined period of time:
“… produces output (170) in the form of a label of behavior of the time series data for a define period of time.” (Dang, col. 5, lines 28-30)
These teachings correspond to the “prediction characterizes a predicted training behavior … at the end of a next time interval” since Dang teaches predicting a labeled behavior outcome for time series data and explicitly provides a next-interval (t+1) sentiment prediction in a financial-market context.
Dang expressly teaches the limitations of claim 10. Therefore, claim 10 is unpatentable for the reasons set forth for claim 8 and subsequently, claim 1.
Regarding claims 11-19
Each of claims 11-19 is the method counterpart of a corresponding already-analyzed system claim (11 [Wingdings font/0xDF][Wingdings font/0xE0] 1; 12 [Wingdings font/0xDF][Wingdings font/0xE0] 2; 13 [Wingdings font/0xDF][Wingdings font/0xE0] 3; 14 [Wingdings font/0xDF][Wingdings font/0xE0] 4; 15 [Wingdings font/0xDF][Wingdings font/0xE0] 5; 16 [Wingdings font/0xDF][Wingdings font/0xE0] 6; 17 [Wingdings font/0xDF][Wingdings font/0xE0] 7; 18 [Wingdings font/0xDF][Wingdings font/0xE0] 8; 19 [Wingdings font/0xDF][Wingdings font/0xE0] 9). As such, claims 11-19 recite substantially the same subject matter as claims 1-9, respectively, but in method form.
Accordingly, the same teachings and combination rationales apply: Cheung teaches the core predictive pipeline, including CNN-based processing and a final connected network head to produce a prediction, implemented on computing hardware, with fixed-length windowing/truncation behavior and convolutional training constructs. Dang teaches recurrent-network based multimodal fusion/aggregation (RNN/AsyncLSTM) over temporal sequences and multimodal feeds, including time-stepped processing and multimodal encodings/fusion behavior. Tan teaches the CNN architectural technique of placing a normalization layer (batch normalization) following convolutional layers.
Therefore, because each of claims 11-19 is the method analog of a corresponding system claim already mapped to the combination, and merely recites the same underlying operations in method form (without adding any materially different structure or technological mechanism), claims 11-19 are rejected under 35 U.S.C. § 103 over Cheung in view of Dang and further in view of Tan for the same reasons discussed above with respect to claims 1-9, respectively.
Regarding claim 20
Independent claim 20 is the non-transitory computer-readable storage medium (CRM) analog of independent claim 11 (and substantively mirrors the independent pipeline of claim 1/11), i.e., it recites instructions stored on a non-transitory medium that, when executed, cause performance of the same time-windowed multimodal prediction pipeline (CNN with normalization; aggregation network; output network; update via new interval and reprocessing).
This CRM form does not materially change the §103 analysis, because Cheung expressly teaches the generic “stored instructions on non-transitory media” wrapper. For example, Cheung discloses: “storage 95 may store instructions for execution by the processor 91” and further describes implementation on “non-transitory machine-readable media”. The claimed CRM “instructions” therefore correspond to generic computer instructions stored on a generic non-transitory medium, which is conventional, and the underlying functional steps are taught/obvious for the same reasons as for independent claim 11 / claim 1 based on Cheung in view of Dang and in further view of Tan.
Accordingly, independent claim 20 is rejected under 35 U.S.C. § 103 as being unpatentable over Cheung in view of Dang and further in view of Tan for the same reasons as independent claim 11 (and substantively independent claim 1), with Cheung additionally teaching the claimed non-transitory storage / stored-instructions elements.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Paul Coleman whose telephone number is (571)272-4687. The examiner can normally be reached Mon-Fri.
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, David Yi can be reached at (571) 270-7519. 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.
/PAUL COLEMAN/Examiner, Art Unit 2126
/VAN C MANG/Primary Examiner, Art Unit 2126