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
In the previous Office Action issued March 11, 2026 (hereinafter “the previous Office Action”), claims 1-27 were pending.
This action is in response to the amendment and remarks filed July 13, 2026. In the amendment, claims 1-3, 6, 7,9, 11, 13, 14, 17, 22, and 24 were amended, no claims were canceled, and no claims were added. Thus, claims 1-27 are pending.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on July 20, 2026 is 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 1-27 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1: Claims 1-6 are directed to a processor [machine].
Claims 7-12 are directed to a system [machine].
Claims 13-19 are directed to a method [process].
Claims 20-27 are directed to a non-transitory, machine-readable medium [machine].
Regarding Claim 1:
Step 2A, Prong 1: The following limitations are directed to the abstract idea of a mental process [see MPEP 2106.04(a)(2) III. C.]. In particular, the claim recites mental processes that are concepts performed in the human mind or with pen and paper (including an observation, evaluation, judgement, or opinion).
generate an event prediction of a strength correlation between a first event and a second event based at least in part, on: a first time of publication of a first published text comprising content corresponding to a first event and a second time of publication of a second published text comprising content corresponding to a second event that is different from the first event
a determination of whether a length of time between the first time of publication of the first published text that the first event was extracted from and the second time of publication of the second published text that the second event was extracted from is greater than a predefined threshold length of time
As drafted, under their broadest reasonable interpretation (BRI), in view of the specification, the above limitations cover concepts performed in the human mind (observation, evaluation, judgement, or opinion). Given a sufficiently small set of data, nothing in the claim prohibits this process from being performed mentally or with pen and paper.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
A processor, comprising: one or more circuits to:
use one or more neural networks to:
provide output data that indicates the event prediction of the strength of correlation between the first event that was extracted from the first published text and the second event that was extracted from the second published text
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
A processor, comprising: one or more circuits to:
use one or more neural networks to:
provide output data that indicates the event prediction of the strength of correlation between the first event that was extracted from the first published text and the second event that was extracted from the second published text
Regarding Claim 2:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim. Additionally,
The following limitations are/remain directed to the abstract idea of a mental process [see MPEP 2106.04(a)(2) III. C.]. In particular, the claim recites mental processes that are concepts performed in the human mind (including an observation, evaluation, judgement, or opinion).
extract the first event and the second event from textual data
infer information about the textual data
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the one or more circuits are to
to train the one or more neural networks to
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the one or more circuits are to
to train the one or more neural networks to
Regarding Claim 3:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim. Additionally,
The following limitations are/remain directed to the abstract idea of a mental process [see MPEP 2106.04(a)(2) III. C.]. In particular, the claim recites mental processes that are concepts performed in the human mind (including an observation, evaluation, judgement, or opinion).
infer information based, at least in part, on a chronological order of the first event and the second event
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the one or more neural networks are to
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the one or more neural networks are to
Regarding Claim 4:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the one or more circuits are to train the one or more neural networks by pre-training the one or more neural networks to perform a plurality of different time-based tasks
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the one or more circuits are to train the one or more neural networks by pre-training the one or more neural networks to perform a plurality of different time-based tasks
Regarding Claim 5:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the one or more circuits use representations of uncut paragraphs of text to train the one or more neural networks
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the one or more circuits use representations of uncut paragraphs of text to train the one or more neural networks
Regarding Claim 6:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the event prediction is further based, at least in part, on where the first event and the second event appear in a publication
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the event prediction is further based, at least in part, on where the first event and the second event appear in a publication
Regarding Claim 7:
Claim 7 corresponds to claim 1.
Step 2A, Prong 1: This claim recites the same abstract ideas as in the corresponding claim.
Step 2A, Prong 2: This claim recites the same additional elements as in the corresponding claim. There are no additional elements in this claim that integrate the judicial exception into a practical application. The analysis of this claim at this step mirror that of corresponding claim, with the addition/exception the following limitations.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
A system, comprising: one or more processors to…
Step 2B: This claim recites the same additional elements as in the corresponding claim. There are no additional elements in this claim that amount to significantly more than the judicial exception. The analysis of this claim at this step mirror that of claim corresponding, with the addition/exception the following limitations.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
A system, comprising: one or more processors to…
Regarding Claim 8:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the one or more neural networks comprise a bidirectional encoder representations from transformers (BERT) learning model
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the one or more neural networks comprise a bidirectional encoder representations from transformers (BERT) learning model
Regarding Claim 9:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the one or more neural networks are trained to predict stock prices
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the one or more neural networks are trained to predict stock prices
Regarding Claim 10:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
The following limitation is directed to the abstract idea of a mental process [see MPEP 2106.04(a)(2) III. C.]. In particular, the claim recites mental processes that are concepts performed in the human mind (including an observation, evaluation, judgement, or opinion).
infer a causal relationship between events
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the one or more processors are further to train the one or more neural networks to
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the one or more processors are further to train the one or more neural networks to
Regarding Claim 11:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements do not meaningfully limit the judicial exception [see MPEP 2106.05(e)]. The claim simply recites additional information regarding the characteristics of the training of neural networks for the generation of an event prediction. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
wherein training the one or more neural networks is based, at least in part, on pre-training with multiple different tasks
wherein performance of the different tasks is based on when one or more events occurred
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements do not meaningfully limit the judicial exception [see MPEP 2106.05(e)]. The claim simply recites additional information regarding the characteristics of the training of neural networks for the generation of an event prediction. Therefore, the additional element does not amount to significantly more than the judicial exception.
wherein training the one or more neural networks is based, at least in part, on pre-training with multiple different tasks
wherein performance of the different tasks is based on when one or more events occurred
Regarding Claim 12:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
The following limitation is directed to the abstract idea of a mental process [see MPEP 2106.04(a)(2) III. C.]. In particular, the claim recites mental processes that are concepts performed in the human mind (including an observation, evaluation, judgement, or opinion).
calculate a distance between two or more events based, at least in part, whether the two or more events appeared in different publications
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the one or more processors train the one or more neural networks to
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the one or more processors train the one or more neural networks to
Regarding Claim 13:
Claim 13 corresponds to claim 1.
Step 2A, Prong 1: The claim recites the same abstract ideas as in claim 1.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application. The analysis of claim 13 at this step mirror that of claim 1.
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception. The analysis of claim 13 at this step mirror that of claim 1.
Regarding Claim 14:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
The following limitation is directed to the abstract idea of a mental process [see MPEP 2106.04(a)(2) III. C.]. In particular, the claim recites mental processes that are concepts performed in the human mind (including an observation, evaluation, judgement, or opinion).
masks a date on which at least one prior event indicated by the first published text or the second published text occurred
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein using the one or more neural networks
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein using the one or more neural networks
Regarding Claim 15:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
The following limitation is directed to the abstract idea of a mental process [see MPEP 2106.04(a)(2) III. C.]. In particular, the claim recites mental processes that are concepts performed in the human mind (including an observation, evaluation, judgement, or opinion).
infer a causal relationship between events and entities
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
further comprising training the one or more neural networks to
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
further comprising training the one or more neural networks to
Regarding Claim 16:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
The following limitation is directed to the abstract idea of a mental process [see MPEP 2106.04(a)(2) III. C.]. In particular, the claim recites mental processes that are concepts performed in the human mind (including an observation, evaluation, judgement, or opinion).
calculate a distance between prior events based, at least in part, on where one or more of the prior events appeared in a publication
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
further comprising training the one or more neural networks to
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
further comprising training the one or more neural networks to
Regarding Claim 17:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein training the one or more neural networks is based, at least in part, on a task of finding expressions in textual data that refer to an entity
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein training the one or more neural networks is based, at least in part, on a task of finding expressions in textual data that refer to an entity
Regarding Claim 18:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein training of the one or more neural networks is based, at least in part, on dwell times for one or more question-document pairs from a log
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein training of the one or more neural networks is based, at least in part, on dwell times for one or more question-document pairs from a log
Regarding Claim 19:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein training of the one or more neural networks is based, at least in part, on rewriting a question of a passage-question pair
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein training of the one or more neural networks is based, at least in part, on rewriting a question of a passage-question pair
Regarding Claim 20:
Claim 20 corresponds to claim 1.
Step 2A, Prong 1: The claim recites the same abstract ideas as in claim 1.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application. The analysis of claim 20 at this step mirror that of claim 1, with the exception of the following limitations.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
A non-transitory, machine-readable medium having stored thereon a set of instructions, which if performed by one or more processors, cause the one or more processors to at least:
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception. The analysis of claim 20 at this step mirror that of claim 1, with the exception of the following limitations.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
A non-transitory, machine-readable medium having stored thereon a set of instructions, which if performed by one or more processors, cause the one or more processors to at least:
Regarding Claim 21:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the one or more neural networks are to be trained based, at least in part, by masking entities
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the one or more neural networks are to be trained based, at least in part, by masking entities
Regarding Claim 22:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the one or more neural networks are to be trained based, at least in part, on masking capitalized phrases
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the one or more neural networks are to be trained based, at least in part, on masking capitalized phrases
Regarding Claim 23:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
The following limitation is directed to the abstract idea of a mental process [see MPEP 2106.04(a)(2) III. C.]. In particular, the claim recites mental processes that are concepts performed in the human mind (including an observation, evaluation, judgement, or opinion).
infer causal relationships between entities and events
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the set of instructions further include instructions, which if performed by the one or more processors, cause the one or more processors to:
train the one or more neural networks to
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the set of instructions further include instructions, which if performed by the one or more processors, cause the one or more processors to:
train the one or more neural networks to
Regarding Claim 24:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
The following limitation is directed to the abstract idea of a mental process [see MPEP 2106.04(a)(2) III. C.]. In particular, the claim recites mental processes that are concepts performed in the human mind (including an observation, evaluation, judgement, or opinion).
calculate one or more distances between at least two prior events extracted from one or more published texts based, at least in part, on a length of time between the occurrences of the at least two prior events
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the set of instructions further include instructions, which if performed by the one or more processors, cause the one or more processors to:
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the set of instructions further include instructions, which if performed by the one or more processors, cause the one or more processors to:
Regarding Claim 25:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements do not meaningfully limit the judicial exception [see MPEP 2106.05(e)]. The claim simply recites additional information regarding the characteristics of the training of neural networks for the generation of an event prediction. Therefore, the additional element does not amount to significantly more than the judicial exception.
wherein a first timestamp of metadata associated with the first published text indicates the first time of publication of the first published text and a second timestamp of metadata associated with the second published text indicates the second time of publication of the second published text
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements can be considered as generally linking the use of judicial exception to a particular technological environment or field of use [See MPEP § 2106.05(h)]. Therefore, the additional elements do not amount to significantly more than the judicial exception.
wherein a first timestamp of metadata associated with the first published text indicates the first time of publication of the first published text and a second timestamp of metadata associated with the second published text indicates the second time of publication of the second published text
Regarding Claim 26:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the set of instructions further include instructions, which if performed by the one or more processors, cause the one or more processors to:
train the one or more neural networks using market movement prediction tasks
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the set of instructions further include instructions, which if performed by the one or more processors, cause the one or more processors to:
train the one or more neural networks using market movement prediction tasks
Regarding Claim 27:
Step 2A, Prong 1: This claim recites the same abstract ideas as in the parent claim.
Step 2A, Prong 2: There are no additional elements in this claim that integrate the judicial exception into a practical application.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to integrate the judicial exception into a practical application.
wherein the set of instructions further include instructions, which if performed by the one or more processors, cause the one or more processors to:
train the one or more neural networks to predict stock prices
Step 2B: There are no additional elements in this claim that amount to significantly more than the judicial exception.
The following additional elements are adding the words “apply it” (or an equivalent) with the judicial exception, mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea [see MPEP 2106.05(f)], and therefore fails to amount to significantly more than the judicial exception.
wherein the set of instructions further include instructions, which if performed by the one or more processors, cause the one or more processors to:
train the one or more neural networks to predict stock prices
Claim Rejections - 35 USC § 103
The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
Claims 1-3, 5-7, 13, 20, and 26-27 are rejected under 35 U.S.C. 103 as being unpatentable over Trim et al. (US 20200394273), hereinafter Trim, in view of Nguyen et al. (US 20210049700), hereinafter Nguyen.
Regarding Claim 1:
Trim discloses:
A processor, comprising: one or more circuits to:
Trim, [0048], “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.
Trim discloses a processor [A processor, comprising: one or more circuits].
…of a strength of correlation between a first event and a second event
Trim, [0077], “A diachronic linguistic analyzer (DLA) module 440 of NLG system 430 analyzes each legacy report 420 in order to infer correlations between linguistic features of a report's natural-language text and one or more temporal characteristics of the report or of the raw data from which the report was generated by legacy report generator 410. As described above, these temporal characteristics may comprise, or be associated with, a report's publication date or generation date or a creation date or capture date of data used to generate the report.
Trim discloses inferring correlations between analyzed legacy reports [a strength of correlation between a first event and a second event].
based, at least in part, on: a first time of publication of a first published text comprising content corresponding to the first event and a second time of publication of a second published text comprising content corresponding to the second event that is different than the first event
Trim, [0093], “In another example, DLA module 440 might select a temporal characteristic based on differences in a duration of time, rather than one based on a threshold date or range of dates. In the previous storm-prediction example, such a characteristic would be a duration of time between the publication of a weather forecast and the time of occurrence of a weather event predicted by the report. As described above, reports that predict an event occurrent several weeks in the future do so in a different linguistic style than the style used by reports that predict an imminent weather event. Here, if a first diachronic group contains reports of weather events occurring at least two days after the publication date of the report, a second diachronic group would contain reports of weather events predicted to occur within two days of the report's publication date.”
Trim discloses temporal characteristics including the publication date of reports [based, at least in part, on: a first time of publication of a first published text comprising content corresponding to the first event and a second time of publication of a second published text comprising content corresponding to the second event that is different than the first event]. The reports are interpreted as different from each other because Trim discloses finding the differences between their dates.
a determination of whether a length of time between the first time of publication of the first published text that the first event was extracted from and the second time of publication of the second published text that the second event was extracted from was greater than a predefined threshold length of time
Trim, [0093], “In another example, DLA module 440 might select a temporal characteristic based on differences in a duration of time, rather than one based on a threshold date or range of dates. In the previous storm-prediction example, such a characteristic would be a duration of time between the publication of a weather forecast and the time of occurrence of a weather event predicted by the report. As described above, reports that predict an event occurrent several weeks in the future do so in a different linguistic style than the style used by reports that predict an imminent weather event. Here, if a first diachronic group contains reports of weather events occurring at least two days after the publication date of the report, a second diachronic group would contain reports of weather events predicted to occur within two days of the report's publication date.”
Trim discloses determining the duration of time between two different publication dates [a determination of whether a length of time between the first time of publication of the first published text that the first event was extracted from and the second time of publication of the second published text that the second event was extracted from]. Then, Trim discloses an example where they form two groups, one occurring at least two days after, and one occurring within two days of the publication date [was greater than a predefined threshold]. In this case, the predefined threshold is two days.
…of the strength of correlation between the first event that was extracted from the first published text and the second event that was extracted from the second published text
As cited above with para. 77, Trim discloses inferring correlations between analyzed legacy reports [of the strength of correlation between the first event that was extracted from the first published text and the second event that was extracted from the second published text].
Trim does not explicitly disclose:
use one or more neural networks to: generate an event prediction…
provide output data that indicates the event prediction…
However, in the same field, analogous art Nguyen teaches:
use one or more neural networks to: generate an event prediction…
Nguyen, [0083], “The disclosed system may extract relevant signals from various financial data sources (e.g., unstructured news data and transaction data) to predict if companies will need capital funding in the future. In order to make such predictions, a neural network (e.g., natural language processor), a first machine learning model, and a second machine learning model are applied on both structured numerical and unstructured textual news data (financial fundamentals, news, press releases, earning calls, etc.)”
In para. 83, Nguyen discloses using neural networks and machine learning models [use one or more neural networks] and various financial data sources to predict if companies will need capital funding [generate an event prediction].
provide output data that indicates the event prediction…
Nguyen, [0299], “FIG. 24 is a diagram 2400 of a sample dashboard rendered on a graphical user interface, according to some embodiments outputting the results of one or more of the models discussed herein.”
[0085], “The first machine learning models are able to make time series forecasting with Mean Square Error (MSE) of less than 0.001 on a test sample; the neural networks are able to extract and correctly classify text into 7 categories with accuracy of 91.134% and f1 score of 0.91280. The results generated by the techniques are input for Naïve Bayes Inference algorithms to output the probability of whether a company will need capital in the future.”
In para. 299, Nguyen discloses a GUI that outputs the results of the models [provide output data], and para. 85 specifies the outputs of the algorithms are the probability of whether a company will need capital in the future [indicates the event prediction].
Trim, Nguyen, and the instant application are analogous art because they are all directed to natural language processing.
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Trim with Nguyen to perform event predictions in order to aid businesses with making financial predictions. “The disclosed system may extract relevant signals from various financial data sources (e.g., unstructured news data and transaction data) to predict if companies will need capital funding in the future. In order to make such predictions, a neural network (e.g., natural language processor), a first machine learning model, and a second machine learning model are applied on both structured numerical and unstructured textual news data (financial fundamentals, news, press releases, earning calls, etc.)” (Nguyen, [0083]).
Regarding Claim 2:
As discussed above, Trim in view of Nguyen teach [the] process of claim 1, and Nguyen further discloses:
wherein the one or more circuits are to extract the first event and the second event from textual data to train the one or more neural networks to infer information about the textual data
Nguyen, [0083], “The disclosed system may extract relevant signals from various financial data sources (e.g., unstructured news data and transaction data) to predict if companies will need capital funding in the future. In order to make such predictions, a neural network (e.g., natural language processor), a first machine learning model, and a second machine learning model are applied on both structured numerical and unstructured textual news data (financial fundamentals, news, press releases, earning calls, etc.)”
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Trim with Nguyen to perform event predictions in order to aid businesses with making financial predictions. “The disclosed system may extract relevant signals from various financial data sources (e.g., unstructured news data and transaction data) to predict if companies will need capital funding in the future. In order to make such predictions, a neural network (e.g., natural language processor), a first machine learning model, and a second machine learning model are applied on both structured numerical and unstructured textual news data (financial fundamentals, news, press releases, earning calls, etc.)” (Nguyen, [0083]).
Regarding Claim 3:
As discussed above, Trim in view of Nguyen teach [the] process of claim 1, and Nguyen further discloses:
wherein the one or more neural networks are to infer information based, at least in part, on a chronological order of the first event and the second event
Nguyen, [0173], “Neural network 104 is configured to receive the one or more numerical time-series data related to a set of historical financial transactions and process the received one or more numerical time-series data to generate one or more future feature data structures having a future feature value and a future instance value”
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Trim with Nguyen to perform event predictions in order to aid businesses with making financial predictions. “The disclosed system may extract relevant signals from various financial data sources (e.g., unstructured news data and transaction data) to predict if companies will need capital funding in the future. In order to make such predictions, a neural network (e.g., natural language processor), a first machine learning model, and a second machine learning model are applied on both structured numerical and unstructured textual news data (financial fundamentals, news, press releases, earning calls, etc.)” (Nguyen, [0083]).
Regarding Claim 5:
As discussed above, Trim in view of Nguyen teach [the] process of claim 1, and Nguyen further discloses:
wherein the one or more circuits use representations of uncut paragraphs of text to train the one or more neural networks
Nguyen, [0231], “The preprocessing unit 118 may process the received unstructured textual news data set 1102 in a variety of manners. For example, the preprocessing unit 118 can process the received unstructured textual news data set 1102 to remove stop words (e.g., commonly recurring work at no meaning resentenced), remove punctuation, remove additional spacing, remove numbers, and remove special characters”
[0233] “During operation, the word vectorizer 1106 may be configured to represent each of the one or more words in each document within the unstructured textual news data set 1102…by structuring the vectorized representation with a term frequency and inverse document frequency (TF-IDF) of text words within each document.”
[0256], “FIGS. 19A and 19B show an example implementation 1900A and 1900B of a Doc2Vec word vectorizer 1106 in Gensim. In the shown embodiment, the Doc2Vec word vectorizer 1106 trained for 100 epochs on the example new data set, with the minimum word count set to two in order to discard words with very few occurrences.”
[i.e., the Doc2Vec word vectorizer is a neural network]
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Trim with Nguyen to perform event predictions in order to aid businesses with making financial predictions. “The disclosed system may extract relevant signals from various financial data sources (e.g., unstructured news data and transaction data) to predict if companies will need capital funding in the future. In order to make such predictions, a neural network (e.g., natural language processor), a first machine learning model, and a second machine learning model are applied on both structured numerical and unstructured textual news data (financial fundamentals, news, press releases, earning calls, etc.)” (Nguyen, [0083]).
Regarding Claim 7:
Claim 7 is a system claim corresponding to processor claim 1 and is rejected for at least the same reasons as given in the rejection of claim 1, with the exception of the following limitations.
Trim discloses:
A system, comprising: one or more processors
Trim, [0048], “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.
Trim discloses system with a processor [A system, comprising: one or more processors].
Regarding Claim 13:
Claim 13 is a method claim corresponding to processor claim 1 and is rejected for at least the same reasons as given in the rejection of claim 1.
Regarding Claim 20:
Claim 20 is a non-transitory, machine-readable medium claim corresponding to processor claim 1 and is rejected for at least the same reasons as given in the rejection of claim 1, with the exception of the following limitations.
Nguyen discloses:
A non-transitory, machine-readable medium having stored thereon a set of instructions, which if performed by one or more processors, cause the one or more processors to at least:
Trim, [0048], “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.
Trim discloses computer readable storage medium with a processor [A non-transitory, machine-readable medium having stored thereon a set of instructions, which if performed by one or more processors, cause the one or more processors to at least].
Regarding Claim 26:
As discussed above, Trim in view of Nguyen teach [the] non-transitory, machine-readable medium of claim 20, and Nguyen further discloses:
wherein the set of instructions further include instructions, which if performed by the one or more processors, cause the one or more processors to: train the one or more neural networks using market movement prediction tasks1
Nguyen, [0012], “In operation, the neural network receives raw numerical data inputs, such as market data relating to an entity including stock price data, volume data, and moving averages of the same, and can generate an output data structure representing a predicted price moving average for a future timestate.”
[0183], “The neural network includes one or more Recurrent Neural Network (RNN) layers (e.g., a first LSTM layer 704 and a second LSTM layer 706, one Kth-order Hidden Markov Models), for analysing, forecasting and detecting anomalies in numerical time-series data [i.e., market movement prediction tasks]…the neural network is trained to generate a future feature data structure having a future feature value and a future instance value for numerical time-series data. For example, during training, the neural network may ingest numerical time-series data and attempt to determine a future feature value for each instance in the numerical time-series data”
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Trim with Nguyen to perform event predictions in order to aid businesses with making financial predictions. “The disclosed system may extract relevant signals from various financial data sources (e.g., unstructured news data and transaction data) to predict if companies will need capital funding in the future. In order to make such predictions, a neural network (e.g., natural language processor), a first machine learning model, and a second machine learning model are applied on both structured numerical and unstructured textual news data (financial fundamentals, news, press releases, earning calls, etc.)” (Nguyen, [0083]).
Regarding Claim 27:
As discussed above, Trim in view of Nguyen teach [the] non-transitory, machine-readable medium of claim 20, and Nguyen further discloses:
wherein the set of instructions further include instructions, which if performed by the one or more processors, cause the one or more processors to: train the one or more neural networks to predict stock prices
Nguyen, [0225] “FIG. 9A is a diagram 900A generated where an example numerical time-series data for Gibson Energy, as listed on the TSX, was processed by the neural network 104 to predict a 30 day moving average of the stock price.”
It would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Trim with Nguyen to perform event predictions in order to aid businesses with making financial predictions. “The disclosed system may extract relevant signals from various financial data sources (e.g., unstructured news data and transaction data) to predict if companies will need capital funding in the future. In order to make such predictions, a neural network (e.g., natural language processor), a first machine learning model, and a second machine learning model are applied on both structured numerical and unstructured textual news data (financial fundamentals, news, press releases, earning calls, etc.)” (Nguyen, [0083]).
Claims 4 and 9 are rejected under 35 U.S.C. 103 as being unpatentable over Trim in view of Nguyen, and further in view of Gage et al. (US 20100235310), hereinafter Gage.
Regarding Claim 4:
As discussed above, Trim in view of Nguyen teach [the] processor of claim 1, but do not explicitly disclose:
wherein the one or more circuits are to train the one or more neural networks by pre-training the one or more neural networks to perform a plurality of different time-based tasks
However, in the same field, analogous art Gage teaches:
wherein the one or more circuits are to train the one or more neural networks by pre-training the one or more neural networks to perform a plurality of different time-based tasks
Gage, [0010], “The artificial neural network can be included in various methods, apparatus, and articles for use in predicting or profiling events.”
[0014], “a computer system for predicting a future event is provided…a trained artificial neural network [i.e., already trained / pre-trained]…and configured to produce new trainable nodes…”
[0051], “to predict or profile events in systems that have substantial dynamics over long time scales…used to predict or profile events in various finance systems. i.e., in stock markets, commodities markets, options markets…”
[0052], “Various embodiments may be useful in forecasting, for example in the field of sports and sporting events…”
Trim, Nguyen, Gage, and the instant application are analogous art because they are all directed to neural networks.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen to incorporate the teachings of Gage to pre-train neural networks to perform a variety of time-based tasks. Doing so would have allowed Trim in view of Nguyen to use Gage’s methods “ for use in predicting or profiling events.”, as suggested by Gage (see, e.g., Gage, [0010]).
Regarding Claim 9:
As discussed above, Trim in view of Nguyen teach [the] processor of claim 1, but do not explicitly disclose:
wherein the one or more neural networks are trained to predict stock prices
However, in the same field, analogous art Gage teaches:
wherein the one or more neural networks are trained to predict stock prices
Gage, [0010], “The artificial neural network can be included in various methods, apparatus, and articles for use in predicting or profiling events.”
[0014], “a computer system for predicting a future event is provided…a trained artificial neural network [i.e., already trained / pre-trained]…and configured to produce new trainable nodes…”
[0051], “to predict or profile events in systems that have substantial dynamics over long time scales…used to predict or profile events in various finance systems. i.e., in stock markets, commodities markets, options markets…”
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen to incorporate the teachings of Gage to pre-train neural networks to perform a variety of time-based tasks. Doing so would have allowed Trim in view of Nguyen to use Gage’s methods “ for use in predicting or profiling events.”, as suggested by Gage (see, e.g., Gage, [0010]).
Claims 6 and 25 are rejected under 35 U.S.C. 103 as being unpatentable over Trim in view of Nguyen, and further in view of Petroni et al. (US 20190012374), hereinafter Petroni.
Regarding Claim 6:
As discussed above, Trim in view of Nguyen teach [the] process of claim 1, but do not explicitly disclose:
wherein the event prediction is further based, at least in part, on where the first event and the second event appear in a publication
However, in the same field, analogous art Petroni teaches:
wherein the event prediction is further based, at least in part, on where the first event and the second event appear in a publication
Petroni, [0171], “The location attribute extraction module 926 generates a location attribute for the event using the candidate attributes”
[i.e., the candidate attributes are directly associated with indications of events]
[0181], “The feature vector for each candidate location may be composed based on the news article and candidate attributes. For example, feature vector for each candidate location may be composed as a concatenation of the following…(4) a position offset of the candidate location in the news article”
[i.e., the position offset of the candidate location (event indication) in the news article (publication) represents an encoding of where the candidates appear in the publication for classification]).
Trim, Nguyen, Petroni, and the instant application are analogous art because they are all directed to natural language processing.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the neural network event prediction processor of Trim and Nguyen to incorporate the teachings of Petroni to use indications based on where two or more of the prior events appear in a publication. Doing so would have allowed Nguyen to use Petroni’s method for “classifying numeric references of the candidate attributes and adjacent word sequences as either representing an impact of the event or not”, as suggested by Petroni (Petroni, [0184]).
Regarding Claim 25:
As discussed above, Trim in view of Nguyen teach [the] non-transitory, machine-readable medium of claim 20, but do not explicitly disclose:
wherein a first timestamp of metadata associated with the first published text indicates the first time of publication of the first published text and a second timestamp of metadata associated with the second published text indicates the second time of publication of the second published text
However, in the same field, analogous art Petroni teaches:
wherein a first timestamp of metadata associated with the first published text indicates the first time of publication of the first published text and a second timestamp of metadata associated with the second published text indicates the second time of publication of the second published text
Petroni, [0020], “FIGS. 5l-5n is an exemplary metadata of an event detected cluster with ingested data of FIG. 5e as one of the related unit data”
FIG. 5g:
PNG
media_image1.png
48
177
media_image1.png
Greyscale
See, e.g., the timestamp on metadata associated with published texts, in FIG. 5g.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen to incorporate the teachings of Petroni to utilize timestamps of metadata associated with published texts of publications indicating their respective times of publication. Doing so would have allowed Nguyen to use Petroni’s method in order to “process and store this event information in an event representation form as used for the news article, social media and coreferenced events”, as suggested by Petroni (Petroni, [0240]).
Claims 8, 14, and 21-22 are rejected under 35 U.S.C. 103 as being unpatentable over Trim in view of Nguyen, and further in view of Hajarnis et al. (US 12056592), hereinafter Hajarnis.
Regarding Claim 8:
As discussed above, Trim in view of Nguyen teach [the] system of claim 7, but do not explicitly disclose:
wherein the one or more neural networks comprise a bidirectional encoder representations from transformers (BERT) learning model
However, in the same field, analogous art Hajarnis teaches:
wherein the one or more neural networks comprise a bidirectional encoder representations from transformers (BERT) learning model
Hajarnis, col 12, lines 30-36, “Implementations may then process different segments in parallel using a BERT neural network layer containing multiple BERT nodes, and then using forward and reverse direction sequence-to-sequence layers to further fine-tune the results to a specific task. Implementations of the disclosure have been tested to show superior performance for NLP tasks.”
Trim, Nguyen, Hajarnis, and the instant application are analogous art because they are all directed to natural language processing.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen in view of Hajarnis to incorporate the teachings of Hajarnis to implement bidirectional encoder representations from transformers (BERT) learning models in neural networks. Doing so would have allowed Trim in view of Nguyen to use Hajarnis' method in order to “To achieve deeper understanding of the underlying text”, as suggested by Hajarnis (Hajarnis, col. 2, lines 60-61).
Regarding Claim 14:
As discussed above, Trim in view of Nguyen teach [the] method of claim 13, but do not explicitly disclose:
wherein using the one or more neural networks masks a date on which at least one prior event indicated by the first published text or the second published text occurred
However, in the same field, analogous art Hajarnis teaches:
wherein using the one or more neural networks masks a date on which at least one prior event indicated by the first published text or the second published text occurred
Hajarnis, col 2, lines 63-67, “BERT may use a technique called Masked Language Modeling (MLM) that may randomly mask words in a sentence and then try to predict the masked words from other words in the sentence surrounding the masked words from both left and right of the masked words.”
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen in view of Hajarnis to incorporate the teachings of Hajarnis to use BERT models that employ masking techniques to make predictions from context in sentences for certain tokens such as dates (e.g., time, day, month, year etc.) for optimal training neural networks. Doing so would have allowed Trim in view of Nguyen to use Hajarnis' method in order to “To achieve deeper understanding of the underlying text”, as suggested by Hajarnis (Hajarnis, col. 2, lines 60-61).
Regarding Claim 21:
As discussed above, Trim in view of Nguyen teach [the] non-transitory machine-readable medium of claim 20, but do not explicitly disclose:
wherein the one or more neural networks are to be trained based, at least in part, by masking entities
However, in the same field, analogous art Hajarnis teaches:
wherein the one or more neural networks are to be trained based, at least in part, by masking entities
Hajarnis, col 2, lines 63-67, “BERT may use a technique called Masked Language Modeling (MLM) that may randomly mask words in a sentence and then try to predict the masked words from other words in the sentence surrounding the masked words from both left and right of the masked words.”
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen in view of Hajarnis to incorporate the teachings of Hajarnis to use BERT models that employ masking techniques to make predictions from context in sentences for certain tokens such as entities (e.g., objects, places, etc.) for optimal training neural networks. Doing so would have allowed Trim in view of Nguyen to use Hajarnis' method in order to “To achieve deeper understanding of the underlying text”, as suggested by Hajarnis (Hajarnis, col. 2, lines 60-61).
Regarding Claim 22:
As discussed above, Trim and Nguyen teach [the] non-transitory machine-readable medium of claim 20, but do not explicitly disclose:
wherein the one or more neural networks are to be trained based, at least in part, on masking capitalized phrases
However, in the same field, analogous art Hajarnis teaches:
wherein the one or more neural networks are to be trained based, at least in part, on masking capitalized phrases
Hajarnis, col 2, lines 63-67, “BERT may use a technique called Masked Language Modeling (MLM) that may randomly mask words in a sentence and then try to predict the masked words from other words in the sentence surrounding the masked words from both left and right of the masked words.”
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen in view of Hajarnis to incorporate the teachings of Hajarnis to use BERT models that employ masking techniques to make predictions from context in sentences for certain tokens such as capitalized phrases (e.g., people, places, acronyms, etc.) for optimal training neural networks. Doing so would have allowed Trim in view of Nguyen to use Hajarnis' method in order to “To achieve deeper understanding of the underlying text”, as suggested by Hajarnis (Hajarnis, col. 2, lines 60-61).
Claims 10-12, 15-17, and 23-24 are rejected under 35 U.S.C. 103 as being unpatentable over Trim in view of Nguyen, and further in view of Blair et al. (US 10878505), hereinafter Blair.
Regarding Claim 10:
As discussed above, Trim in view of Nguyen teach [the] system of claim 7, but do not explicitly disclose:
wherein the one or more processors are further to train the one or more neural networks to infer a causal relationship between events
However, in the same field, analogous art Blair teaches:
wherein the one or more processors are further to train the one or more neural networks
Blair, col. 16, lines 10-15, “commodity-specific neural networks 190…which inform and enable the development of comprehensive forecasts within the commodity forecasting algorithm 180.”
to infer a causal relationship between events
Blair, col. 15, lines 27-31, “The time-series modeling engine 170 converts information…that reflects the temporally-relevant explanation of the information therein to understand how events affect commodity values over time.”
[i.e., a causal relationship between events]
Col. 31, lines 25-43, “At step 350 the process the performs the commodity forecasting algorithm…with the sequence of discrete time data points 178 to create a set of classified, normalized content 186 for the selected commodity 102 that represents variables having an influence over the commodity state over the specified period of time, as well as”
Under the broadest reasonable interpretation (BRI), variables having influence over commodity states for a specified period of time suggests that the variables relate to supply and demand impact in commodity prices or market conditions which corresponds to a causal relationship between events. Additionally, under the BRI, a time-series modeling engine (described as a supervised instantiation of machine learning in Blair, col. 15 lines 5-6), that’s used for understanding how events affect commodity prices, would reasonably include a machine learning model such as a neural network, which are known to include supervised learning models capable of modeling time-series data.
Trim, Nguyen, Blair, and the instant application are analogous art because they are all directed to natural language processing.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen in view of Blair to incorporate the teachings of Blair to train neural networks to infer a causal relationship between events. Doing so would have allowed Trim and Nguyen to use Blair's method in order to “to process large amounts of data from both historical pricing data and other, non-traditional data sources, which can be analyzed in differing temporal contexts (i.e., over time and in real time).”, as suggested by Blair (Blair, col. 1, lines 55-57).
Regarding Claim 11:
As discussed above, Trim in view of Nguyen teach [the] system of claim 7, but do not explicitly disclose:
wherein training the one or more neural networks is based, at least in part, on pre-training with multiple different tasks
wherein performance of the different tasks is based on when on or more events occurred
However, in the same field, analogous art Blair teaches:
wherein training the one or more neural networks is based, at least in part, on pre-training with multiple different tasks
Blair, col. 28, lines 12-33, “In unsupervised meta learning, machine learning algorithms themselves propose their own task distributions within this environment, given un-curated and unlabeled data [i.e., pre-training on unlabeled data], by automatically and continuously curating new commodity-specific data as it is ingested. In the present invention, proposing (or learning) a task distribution includes optimizing parameters for the reward function that maps variables (sentiment) to different reward functions (price prediction) for each commodity… In this manner, unsupervised meta learning causes the system to discover its own optimized task distributions based on reward functions that it learns based on environmental similarities. Unsupervised meta learning is therefore utilized to learn best approaches for improving the production neural network(s) 191 and the training neural network(s) 192”
wherein performance of the different tasks is based on when on or more events occurred
Blair, col.27, lines 43-52, “at least to improve upon training data sets for training neural network(s) 192, and the outcomes of the production neural network(s) 191. The one or more deep learning meta networks 200 are therefore a deeper layer within the neural network modeling layer. Both of these functions are designed to produce outcomes in forecasts of future prices which are as close as possible to simulations of commodity prices using historical pricing data, and improve the overall accuracy [i.e., correct performance] of the multi-layer machine learning-based model.”
Col. 20, lines 22-46, “Neural networks having a recurrent architecture may also have stored, or controlled, internal states which permit storage under direct control of the neural network, making them more suitable for inputs having a temporal nature…In the present invention, where output data 220 is in the form of forecasted states, or prices, of commodities at some future, an understanding of the influence of various events and sentiment on a state over a period of time lead to more highly accurate and reliable forecasts.”
[i.e., the neural network performance is dependent on when events occurred]
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen in view of Blair to incorporate the teachings of Blair to pre-train neural networks with tasks whose correct performance is dependent on time-based events. Doing so would have allowed Trim in view of Nguyen to use Blair’s method in order to “develop(ing) and apply(ing) a deep learning aspect to the model, (so that) the algorithm(s) can determine on its own if a prediction is accurate or not through its own neural network.” as suggested by Blair (Blair, cols. 2, lines 12-15).
Regarding Claim 12:
As discussed above, Trim in view of Nguyen teach [the] system of claim 7, but do not explicitly disclose:
wherein the one or more processors train the one or more neural networks to calculate a distance between two or more events based
at least in part, whether the two or more events appeared in different publications
However, in the same field, analogous art Blair teaches:
wherein the one or more processors train the one or more neural networks to calculate a distance between two or more events based
Blair, col. 18, lines 57-64, “The neural network modeling layer therefore uses the commodity-specific indicators for each taxonomy (keywords, frequencies of their occurrence and distance relationships between them), to construct the nodes and connections of the production neural networks 191. Each textual vector corresponds to an assigned keyword, frequency, or distance relationship, and each node and connection is designed to model a particular aspect of the commodity 102.”
Col. 26-27, lines 65-9, “These activities are utilized to infer relevance in new documents that are ingested into the data analytics platform 100 that are un-curated, or unlabeled. In this manner, as the new information about each commodity is ingested, the neural network modeling layer is able to train itself both to identify relevant documents, and build on existing taxonomies of indicators for each commodity being analyzed. Still further, identifying such clusters 202 also enables development of a taxonomy corpus that is entity-specific as to the entity generating the corporate communications documents 136, and allows for training on entity-specific neural networks.”
at least in part, whether the two or more events appeared in different publications
Blair col. 9, lines 8-24, “News reports 131…Other textual information sources 133 include any other sources of unstructured data that must be processed to develop a sentiment 161 therefrom. These may include magazine articles, blog posts, forum comments (or comments to any news article, magazine article, or blog post) and any other digital sources of informational reports or comments.”
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen to incorporate the teachings of Blair to train neural networks to calculate distances between multiple events that are based on the locality of said events in different publications . Doing so would have allowed Trim in view of Nguyen to use Blair’s method in order to “to integrate a knowledge-based approach into such neural networks that applies rules developed from fundamental economic and commodity-specific indicators in forecasting commodity states”, as suggested by Blair (Blair, col. 3-4, lines 64-3).
Regarding Claim 15:
As discussed above, Trim in view of Nguyen teach [the] method of claim 13, but do not explicitly disclose:
further comprising training the one or more neural networks to infer a causal relationship between events and entities
However, in the same field, analogous art Blair teaches:
further comprising training the one or more neural networks to infer a causal relationship between events and entities
Blair, col. 29, lines 27-43, “neural networks configured to learn how to improve the outcomes generated by the one or more neural networks 190, reinforced by a deeper understanding of how the input data 110 is influenced by, characterized by, or affected by its environment…in the present invention, a selected commodity 102—has a price that is necessarily influenced by characteristics which affect sentiment (its “environment”) through various interactions, such as economic policy and activity, weather, etc. [i.e., events] … building a further taxonomy of indicators 205 from corporate communications documents 136. ”
And with reference to the entities, see e.g., Blair, col. 7, lines 20-24, “may also include corporate communications documents 136…filings that corporate entities are required to file with governmental regulatory agencies”
Under the broadest reasonable interpretation (BRI), learning how input data is influenced, or affected by its environment including interactions such as economic policy, can be interpreted as inferring a causal relationship between events and entities, because corporate entities are impacted by economic policy events, such as federal interest rate changes, and said relationships are learned or inferred by the neural network.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen to incorporate the teachings of Blair to train neural networks to infer a causal relationship between events and entities. Doing so would have allowed Trim in view of Nguyen to use Blair's method in order to “to process large amounts of data from both historical pricing data and other, non-traditional data sources, which can be analyzed in differing temporal contexts (i.e., over time and in real time).”, as suggested by Blair (Blair, col. 1, lines 55-57).
Regarding Claim 16:
As discussed above, Trim in view of Nguyen teach [the] method of claim 13, but do not explicitly disclose:
further comprising training the one or more neural networks to calculate a distance between prior events based, at least in part, on where the one or more prior events appeared in a publication
However, in the same field, analogous art Blair teaches:
further comprising training the one or more neural networks to calculate a distance between prior events based, at least in part,
Blair, col. 18, lines 57-64, “The neural network modeling layer therefore uses the commodity-specific indicators for each taxonomy (keywords, frequencies of their occurrence and distance relationships between them), to construct the nodes and connections of the production neural networks 191. Each textual vector corresponds to an assigned keyword, frequency, or distance relationship, and each node and connection is designed to model a particular aspect of the commodity 102.”
Blair, col. 26-27, lines 65-9, “These activities are utilized to infer relevance in new documents that are ingested into the data analytics platform 100 that are un-curated, or unlabeled. In this manner, as the new information about each commodity is ingested, the neural network modeling layer is able to train itself both to identify relevant documents, and build on existing taxonomies of indicators for each commodity being analyzed. Still further, identifying such clusters 202 also enables development of a taxonomy corpus that is entity-specific as to the entity generating the corporate communications documents 136, and allows for training on entity-specific neural networks”
on where the one or more prior events appeared in a publication
Blair col. 9, lines 8-24, “News reports 131…Other textual information sources 133 include any other sources of unstructured data that must be processed to develop a sentiment 161 therefrom. These may include magazine articles, blog posts, forum comments (or comments to any news article, magazine article, or blog post) and any other digital sources of informational reports or comments”
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen to incorporate the teachings of Blair to train neural networks to calculate a distance between prior events based on their locality in a publication. Doing so would have allowed Trim in view of Nguyen to use Blair’s method in order to “to integrate a knowledge-based approach into such neural networks that applies rules developed from fundamental economic and commodity-specific indicators in forecasting commodity states”, as suggested by Blair (Blair, cols. 3-4, lines 64-3).
Regarding Claim 17:
As discussed above, Trim in view of Nguyen teach [the] method of claim 13, but do not explicitly disclose:
wherein training the one or more neural networks is based, at least in part, on a task of finding expressions in textual data that refer to an entity
However, in the same field, analogous art Blair teaches:
wherein training the one or more neural networks is based, at least in part, on a task of finding expressions in textual data that refer to an entity
Blair, col 27. lines 1-8, “In this manner, as the new information about each commodity is ingested, the neural network modeling layer is able to train itself both to identify relevant documents, and build on existing taxonomies of indicators for each commodity being analyzed. Still further, identifying such clusters 202 also enables development of a taxonomy corpus that is entity-specific as to the entity generating the corporate communications documents 136, and allows for training on entity-specific neural networks.”
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen to incorporate the teachings of Blair to train neural networks on a task of finding expressions in textual data that refer to entities. Doing so would have allowed Trim in view of Nguyen to use Blair method in order to “enables development of a taxonomy corpus that is entity-specific as to the entity generating the corporate communications documents 136, and allows for training on entity-specific neural networks.”, as suggested by Blair (Blair, col 27, lines 7-9).
Regarding Claim 23:
As discussed above, Trim in view of Nguyen teach [the] non-transitory machine-readable medium of claim 20, but do not explicitly disclose:
wherein the set of instructions further include instructions, which if performed by the one or more processors, cause the one or more processors to: train the one or more neural networks to infer causal relationships between entities and events
However, in the same field, analogous art Blair teaches:
wherein the set of instructions further include instructions, which if performed by the one or more processors, cause the one or more processors to: train the one or more neural networks
Blair, col. 16, lines 10-15, “These assessments are performed in conjunction with one or more commodity-specific neural networks 190 and one or more deep learning meta networks 200 as described further below, which inform and enable the development of comprehensive forecasts within the commodity forecasting algorithm 180”
to infer causal relationships between entities and events
Blair, col. 15, lines 27-31, “The time-series modeling engine 170 converts information…that reflects the temporally-relevant explanation of the information therein to understand how events affect commodity values over time.”
Col. 31, lines 25-43, “At step 350 the process the performs the commodity forecasting algorithm…with the sequence of discrete time data points 178 to create a set of classified, normalized content 186 for the selected commodity 102 that represents variables having an influence over the commodity state over the specified period of time, as well as”
With respect to the time-series modeling engine and the aspects of inferring causal relationships between events, the Broadest Reasonable Interpretation (BRI) analysis in claim 10, applies equally here.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen to incorporate the teachings of Blair to train neural networks to infer a causal relationship between events and entities. Doing so would have allowed Trim in view of Nguyen to use Blair's method in order to “to process large amounts of data from both historical pricing data and other, non-traditional data sources, which can be analyzed in differing temporal contexts (i.e., over time and in real time).”, as suggested by Blair (Blair, col. 1, lines 55-57).
Regarding Claim 24:
As discussed above, Trim in view of Nguyen teach [the] non-transitory machine-readable medium of claim 20, but do not explicitly disclose:
wherein the set of instructions further include instructions, which if performed by the one or more processors, cause the one or more processors to
calculate one or more distances between at least two prior events extracted from one or more published texts based, at least in part, on a length of time between occurrences of the at least two prior events
However, in the same field, analogous art Blair teaches:
wherein the set of instructions further include instructions, which if performed by the one or more processors, cause the one or more processors to:
calculate one or more distances between at least two prior events extracted from one or more published texts based, at least in part, on a length of time between occurrences of the at least two prior events
Blair col. 18, lines 57-64, “The neural network modeling layer therefore uses the commodity-specific indicators for each taxonomy (keywords, frequencies of their occurrence and distance relationships between them), to construct the nodes and connections of the production neural networks 191. Each textual vector corresponds to an assigned keyword, frequency, or distance relationship, and each node and connection is designed to model a particular aspect of the commodity 102.”
Blair col. 26-27, lines 65-9, “These activities are utilized to infer relevance in new documents that are ingested into the data analytics platform 100 that are un-curated, or unlabeled. In this manner, as the new information about each commodity is ingested, the neural network modeling layer is able to train itself both to identify relevant documents, and build on existing taxonomies of indicators for each commodity being analyzed. Still further, identifying such clusters 202 also enables development of a taxonomy corpus that is entity-specific as to the entity generating the corporate communications documents 136, and allows for training on entity-specific neural networks.”
[i.e., distance relationships between textual vectors of multiple commodities (financial events) is calculated]
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen to incorporate the teachings of Blair to train neural networks to calculate distances between multiple past events that are based on the length of times between occurrences of said events . Doing so would have allowed Trim in view of Nguyen to use Blair’s method in order to “to integrate a knowledge-based approach into such neural networks that applies rules developed from fundamental economic and commodity-specific indicators in forecasting commodity states”, as suggested by Blair (Blair, cols. 3-4, lines 64-3).
Claims 18 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Trim in view of Nguyen, and further in view of Gong et al. (WO 2021146003), hereinafter Gong.
Regarding Claim 18:
As discussed above, Trim in view of Nguyen teach [the] method of claim 13, but do not explicitly disclose:
wherein training of one or more neural networks is based, at least in part, on dwell times for one or more question-document pairs from a log
However, in the same field, analogous art Gong teaches:
wherein training of one or more neural networks is based, at least in part, on dwell times for one or more question-document pairs from a log
Gong, [0020], “The collection cost of implicit relevance feedbacks for web documents is relatively low, the quantity is large, and the burden on users is not increased. Various features for mining implicit relevance feedbacks for Web documents from user behaviors have been proposed, e.g., click information, average dwell time, number of page visits, etc.”
[0003], “Embodiments of the present disclosure provide methods and apparatuses for providing QA training data and training a QA model based on implicit relevance feedbacks. A question-passage pair and corresponding user behaviors may be obtained from a search log. Behavior features may be extracted from the user behaviors. A relevance score between the question and the passage may be determined, through an implicit relevance feedback model, based on the behavior features.”
[0068], “The QA model 610 shown in FIG.6 may have an architecture based on various technologies. For example, the QA model 610 may be based on a deep neural network, e.g., bidirectional long short term memory (BiLSTM), bidirectional encoder representation from transformers (BERT), etc.” [i.e., neural network]
Nguyen, Nguyen, Gong, and the instant application are analogous art because they are all directed to neural networks.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen in view of Gong to incorporate the teachings of Gong to provide techniques for training neural networks based on dwell times for question-document pairs from a log. Doing so would have allowed Trim in view of Nguyen to incorporate Gong’s method for obtaining "…good indicator(s) for the relevance between the passage and the question.” [i.e., training neural networks on user behavior indicators for question-document pairs from a log], (Gong, [0052]).
Regarding Claim 19:
As discussed above, Trim in view of Nguyen teach [the] method of claim 13, but do not explicitly disclose:
wherein training of one or more neural networks is based, at least in part, on rewriting a question of a passage-question pairs
However, in the same field, analogous art Gong teaches:
wherein training of one or more neural networks is based, at least in part,
Gong, [0030], “The relevant question block 130 may include questions relevant to or similar to the user question in the search block 110. These relevant questions may include, e.g., questions frequently searched by other users. In FIG. 1, multiple questions relevant to the user question "summer flu treatment" are shown in the relevant question block 130, e.g., "What causes summer flu?", "Medicines for summer flu?", etc. When the user clicks on a relevant question…” [i.e., generated rewritten questions])
on rewriting a question of a passage-question pairs
Gong, [0083], “In an implementation, the method 700 may further comprise: adding the question-passage pair and the relevance label into a QA training data set as a QA training data example.” [i.e., passage-question pairs used for training data set]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Trim and Nguyen in view of Gong to incorporate the teachings of Gong to provide techniques for training neural networks to rewrite questions for question-passage pairs from a log. Doing so would have allowed Trim in view of Nguyen to incorporate Gong’s method for inferring relevant questions for question-document pairs as explained here: “These relevant questions may include, e.g., questions frequently searched by other users”, (Gong, [0030]).
Response to Arguments
Applicant's arguments filed July 13, 2026 (“Remarks”) have been fully considered but they are not persuasive.
35 U.S.C. § 103:
Remarks, pp. 9-13. Applicant’s arguments with respect to claims 1, 7, 13, and 20 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
35 U.S.C. § 101:
Remarks, pg. 13-15. Applicant argues amended claim 1 integrates into a practical application. In particular, Applicant cites para. 2 of the specification as disclosure for improvements in addition to the recent Appeals Review Panel decision (Ex parte Desjardins). Examiner respectfully disagrees.
MPEP 2106.05(a):
“If it is asserted that the invention improves upon conventional functioning of a computer, or upon conventional technology or technological processes, a technical explanation as to how to implement the invention should be present in the specification. That is, the disclosure must provide sufficient details such that one of ordinary skill in the art would recognize the claimed invention as providing an improvement…After the examiner has consulted the specification and determined that the disclosed invention improves technology, the claim must be evaluated to ensure the claim itself reflects the disclosed improvement in technology”
In view of the improvement cited in paragraph 2, Examiner consulted the specification for the technical details. Paragraph 2 discusses improvements to training neural networks to facilitate question-answering and casual inferencing. Examiner found that the disclosure (e.g. paragraphs [0145]-[0150]) details a question-answering application. However, the claim itself merely recites using neural networks to generate an event prediction and output data indicative of the event prediction. The claim does not recite limitations that reflect any question-answering facilitation. Therefore, the claim does not reflect the disclosed improvement in technology. For at least these reasons, amended claim 1 remains directed to the judicial exception and does not integrate into a practical application.
Remarks, pg. 15-16. Applicant further argues that amended claim 1 amounts to significantly more based on a combination of elements. Examiner respectfully disagrees. The combination of claim limitations are directed to generating an event prediction (a mental process), which uses a distance calculation (a mental process). The prediction and the distance calculation are then in combination applied to the field of publications involving published texts. For at least these reasons, amended claim 1 does not amount to significantly more than the judicial exception.
Claims 7, 13, and 20, corresponding to claim 1, are rejected for at least the same reasons.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to STEVEN PHUNG whose telephone number is (703) 756-1499. The examiner can normally be reached Monday-Thursday: 9:00AM-4:00PM ET.
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, KAMRAN AFSHAR can be reached at (571) 272-7796. 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.
/S.H.P./Examiner, Art Unit 2125
/KAMRAN AFSHAR/Supervisory Patent Examiner, Art Unit 2125
1 Examiner interprets the phrase “market movement prediction tasks” in view of Paragraphs [0090-0091]
Prosecution Insights