DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
This non-final office action is in response to the preliminary amendment filed 6 December 2023.
Claims 39-58 are pending. Claims 39, 53, 56, and 58 are independent claims.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 23 February 2024 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Drawings
The examiner accepts the drawings filed on 6 December 2023.
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 29-58 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1:
According to Step 1 of the two Step analysis, claims 39-52 are directed toward an apparatus (machine). Claims 53-55 are directed toward an apparatus (machine). Claims 56-57 are directed toward a method (process). Claim 58 is directed toward a method (process). Therefore, each of these claims falls within one of the four statutory categories.
Claim 39:
Step 2A, Prong 1:
The claim recites:
learn importance of one or more parameters by using a training dataset (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation of parameters of a training dataset to learn the importance of the parameters)
define one or more masks for indicating an importance of the one or more parameters for model finetuning (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation to define one or more masks indicated the observed importance of parameters)
finetune at least one parameter of the one or more parameters based at least on the at least one or more mask (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an opinion to finetune the at least one parameter of the one or more parameters based at least on the at least one or more masks)
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional elements:
an apparatus comprising at least one processor
at least one non-transitory memory comprising computer program code, wherein the at least one non-transitory memory and the computer program code are configured to, with the processor, cause the apparatus to at least perform
The additional elements are recited at a high-level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
The claim recites the additional elements:
share at least one mask of the one or more masks with at least one of an encoder or a decoder
send or signal one or more weight updates corresponding to the at least one parameter in a bitstream to the decoder
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do not integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional elements:
an apparatus comprising at least one processor
at least one non-transitory memory comprising computer program code, wherein the at least one non-transitory memory and the computer program code are configured to, with the processor, cause the apparatus to at least perform
The additional elements are recited at a high-level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
The claim recites the additional elements:
share at least one mask of the one or more masks with at least one of an encoder or a decoder
send or signal one or more weight updates corresponding to the at least one parameter in a bitstream to the decoder
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 40:
With respect to claim 40, the claim depends upon claim 39. The analysis of claim 39 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
learn an ordering of the importance of the one or more parameters… by using the training dataset (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation to learn an ordering of the importance of the one or more parameters)
define one or more ordering lists for indicating the ordering of the importance of the one or more parameters (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an opinion to define an ordered list for indicating the order of importance of the one or more parameters)
recovers the at least one weight update from the bitstream by using the received at least one parameter and the at least one ordering list (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation to recover the weight update by using the received parameter and the ordered list)
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional elements:
the decoder
share at least one ordering list of the one or more ordering lists with at least one of the encoder or the decoder
The claim further recites the additional elements at a high-level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
The claim recites the additional element:
send or signal at least one parameter of the one or more parameters that are updated by the model finetuning and at least one weight update corresponding to the at least one parameters to the decoder
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do not integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional elements:
the decoder
share at least one ordering list of the one or more ordering lists with at least one of the encoder or the decoder
The claim further recites the additional elements at a high-level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
The claim recites the additional element:
send or signal at least one parameter of the one or more parameters that are updated by the model finetuning and at least one weight update corresponding to the at least one parameters to the decoder
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 41:
With respect to claim 41, the claim depends upon claim 40. The analysis of claim 40 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
wherein the apparatus is further caused to determine at least one of: the at least one mask of the one or more masks to be used by the encoder during the model finetuning or at least one ordering list of the one or more ordering lists to be used by the encoder during a model finetuning stage (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation to determine either the mask or the ordered list to be used by the encoder during finetuning)
Step 2A, Prong 2:
The claim does not recite additional elements considered under Step 2A, Prong 2.
Step 2B:
The claim does not recite additional elements considered under Step 2B.
Step 42:
With respect to claim 42, the claim depends upon claim 40. The analysis of claim 40 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
minimize an objective function with a suitable loss term (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation to minimize an objective function)
determine one or more importance scores for each parameter based on a minimized objective function (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation to determine one or more importance score for each parameter based on a minimized objective function)
determine at least one of: the at least one mask by using the one or more importance scores and a threshold value or the at least one ordering list by sorting the one or more parameters based on the one or more importance scores of the each one or more parameters (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation to determine at least one of a mask or an ordered list by using the importance scores)
Step 2A, Prong 2:
The claim does not recite additional elements considered under Step 2A, Prong 2.
Step 2B:
The claim does not recite additional elements considered under Step 2B.
Claim 43:
With respect to claim 43, the claim depends upon claim 42. The analysis of claim 42 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
wherein the one or more importance scores for the each parameter comprises one of: a sum of an absolute value of the weight update value of the each parameter for samples in the training dataset; a sum of an absolute value of a ratio of the weight update value over an original weight value on the each parameter for samples in the training dataset; a combination of an importance metric and a learned importance, wherein the combination is using a summation or multiplication scores of the one ore more parameters; or the non-linear combination is determined using a learning based approach (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation to determine importance scores for each parameter via one of four possible evaluations)
Step 2A, Prong 2:
The claim does not recite additional elements considered under Step 2A, Prong 2.
Step 2B:
The claim does not recite additional elements considered under Step 2B.
Claim 44:
With respect to claim 44, the claim depends upon claim 40. The analysis of claim 40 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
identify an optimal number of the one or more parameters and the corresponding weight updates, and wherein to identify the optimal number of the one or more parameters and the corresponding weight updates (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation to identify an optimal number of parameters and corresponding weights)
perform a first model finetuning by selecting a first set of important parameters based on the one or more ordering lists (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation to perform a first model finetuning by selecting a set of importance parameters)
calculate a first rate distortion (RD) loss value including weight update overheads (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation to calculate a first rate distortion loss value)
repeat following until a stopping criterion is met: remove a predetermined number of important parameters from the first set of important parameters; continue the model finetuning from the first saved finetuned model to achieve a second finetuned model; and calculate a second RD loss value including the weight update overheads (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation to remove parameters and an evaluation to calculate a second RD loss value)
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional elements:
save the first finetuned model to achieve a first saved finetuned model
save the second finetuned model
The additional elements are recited at a high level of generality and amounts to extra-solution activity of receiving data. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do not integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional elements:
save the first finetuned model to achieve a first saved finetuned model
save the second finetuned model
The additional elements are recited at a high level of generality and amounts to extra-solution activity of receiving data. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 45:
With respect to claim 45, the claim depends upon claim 39. The analysis of claim 39 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
determine an optimal mask of the one or more masks (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses a judgement to determine an optimal mask of the one or more masks)
determine optimal weight updates corresponding to the optimal mask at a model finetuning stage (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses a judgement to determine an optimal weight corresponding to the optimal mask)
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional element:
send or signal an index of the optimal mask and the optimal weight updates corresponding to the optimal mask to the decoder
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
The claim recites the additional element:
wherein the decoder reconstructs the index of the optimal mask and the optimal weight updates from the bitstream and updates the parameters corresponding to the optimal weight updates in a model according to the optimal mask
The decoder is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do not integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional element:
send or signal an index of the optimal mask and the optimal weight updates corresponding to the optimal mask to the decoder
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
The claim recites the additional element:
wherein the decoder reconstructs the index of the optimal mask and the optimal weight updates from the bitstream and updates the parameters corresponding to the optimal weight updates in a model according to the optimal mask
The decoder is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 46:
With respect to claim 46, the claim depends upon claim 40. The analysis of claim 40 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
determine an optimal ordering list from the one or more ordering lists (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses a judgement to determine an optimal ordering list from the one or more ordering lists)
determine an optimal number of parameters from the one or more parameters (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses a judgement to determine an optimal number of parameters from the one or more parameters)
determine an optimal number of weight updates corresponding to the optimal parameters at a model finetuning stage (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses a judgement to determine an optimal number of weight updates corresponding to optimal parameters)
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional element:
send or signal an index of the optimal ordering list, the optimal number of parameters, and the optimal weight updates to the decoder
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
The claim recites the additional element:
wherein the decoder reconstructs the index of the optimal ordering list, the optimal number of parameters, and the optimal weight updates from the bitstream, and updates the parameters corresponding to the optimal weight updates in a model according to the optimal ordering list and the optimal number of parameters
The decoder is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do no integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional element:
send or signal an index of the optimal ordering list, the optimal number of parameters, and the optimal weight updates to the decoder
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
The claim recites the additional element:
wherein the decoder reconstructs the index of the optimal ordering list, the optimal number of parameters, and the optimal weight updates from the bitstream, and updates the parameters corresponding to the optimal weight updates in a model according to the optimal ordering list and the optimal number of parameters
The decoder is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 47:
With respect to claim 47, the claim depends upon claim 40. The analysis of claim 40 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
learn the one or more masks or the one or more ordered lists from an input content (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses a judgement to learn the one or more masks or ordered lists from input content)
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional element:
signal the learned one or more masks or ordered lists to the decoder
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do no integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional element:
signal the learned one or more masks or ordered lists to the decoder
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 48:
With respect to claim 48, the claim depends upon claim 40. The analysis of claim 40 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
define at least one of the one or more masks or the one or more ordering lists (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses a judgement to define the one or more masks or ordered lists from input content)
define a unique identifier for each mask or each ordering list, and indicate unique identifiers of the each mask or the each ordering list in or along the bitstream from the encoder to the decoder (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses a judgement to define a unique identifier for each mask or ordering list)
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional element:
deliver the at least one of the at least one mask or the at least one ordering list in or along the bitstream
deliver the at least one of the at least one mask or the at least one ordering list
deliver the at least one of the at least one mask or the at least one ordering list from the encoder to the decoder prior to or during delivery of the bitstream
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do no integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional element:
deliver the at least one of the at least one mask or the at least one ordering list in or along the bitstream
deliver the at least one of the at least one mask or the at least one ordering list
deliver the at least one of the at least one mask or the at least one ordering list from the encoder to the decoder prior to or during delivery of the bitstream
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 49:
With respect to claim 49, the claim depends upon claim 40. The analysis of claim 40 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
select a first number of parameters to be updated for a first unit of the bitstream (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation to select a first number of parameters to be updated)
indicate the first number of parameters and the updated first number of parameters associated with the first unit of the bitstream (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation to indicate the first number of parameters and update the first number of parameters associated with the first unit of the bitstream)
select a second number of parameters to be updated for a second unit of the bitstream, wherein the second unit follows the first unit in a bitstream order, and the second number of parameters follow the first number of parameters in the at least one ordering list (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses a judgement to select a second number of parameters to be updated)
indicate the second number of parameters and the updated second number of parameters associated with the second unit of the bitstream (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation to indicate the second number of parameters and the updated second number of parameters)
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional element:
use an updated neural network for encoding the first unit of the bitstream
use the updated neural network for encoding the second unit of the bitstream
The decoder is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do no integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional element:
The claim recites the additional element:
use an updated neural network for encoding the first unit of the bitstream
use the updated neural network for encoding the second unit of the bitstream
The decoder is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 50:
With respect to claim 50, the claim depends upon claim 49. The analysis of claim 49 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
select at least one of the first number of parameters or the second number of parameters according to one or more of the following: a bit budget for the updated first number of parameters or the updated second number of parameters, or a rate-distortion decision performed for the first unit and the first number of parameters separately from the second unit and the second number of parameters, wherein the first unit and the second unit may are selected to be a first coded picture and a second coded picture respectively, or a first hierarchical prediction structure of pictures and a second hierarchical prediction structure of the pictures (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation of a bit budget or rate-distortion to select at least one of the first number of parameters or second number of parameters)
Step 2A, Prong 2:
The claims fail to recite any additional elements considered under Step 2A, Prong 2.
Step 2B:
The claims fail to recite any additional elements considered under Step 2B.
Claim 51:
With respect to claim 51, the claim depends upon claim 50. The analysis of claim 50 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
wherein the bit budget is determined based on one or more of the following: a leftover bitrate in a previous unit the bitstream, or a leftover size in video packets, wherein in a packet-oriented networks one or more packers up to a maximum transfer unit (MTU) size are treated equally, and wherein the bit budget is be determined to fill in a packet up to an MTU size in an instance a unit of the bitstream carried in the packet has size less than the MTU size (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation to determine a bit budget)
Step 2A, Prong 2:
The claims fail to recite any additional elements considered under Step 2A, Prong 2.
Step 2B:
The claims fail to recite any additional elements considered under Step 2B.
Claim 52:
With respect to claim 52, the claim depends upon claim 49. The analysis of claim 49 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
determine and indicate in or along the bitstream: whether the second number of parameters follow the first number of parameters in the at least one ordering list, or whether the second number of parameters are related to a start of the at least one ordering list, … decodes associated weight updates according to an indicated position in the at least one ordering list (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation to determine whether the second number of parameters follow the first number of parameters or whether the second number of parameters are related to a start of the at least one ordering list)
Step 2A, Prong 2:
The claims fail to recite any additional elements considered under Step 2A, Prong 2.
Step 2B:
The claims fail to recite any additional elements considered under Step 2B.
Claim 54:
Step 2A, Prong 1:
The claim recites:
reconstruct the at least one weight updated from the bitstream (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation to reconstruct the at least one weighted update from the bitstream)
update corresponding at least one parameter based on at least one mask (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an opinion to update the at least one parameter of the one or more parameters based at least on the at least one or more masks)
wherein importance of the at least one parameter is learnt by using a training dataset (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation of parameters of a training dataset to learn the importance of the parameters)
wherein one or more masks are defined for indicating importance of one or more parameters for a model finetuning (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation to define one or more masks indicated the observed importance of parameters)
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional elements:
an apparatus comprising at least one processor
at least one non-transitory memory comprising computer program code, wherein the at least one non-transitory memory and the computer program code are configured to, with the processor, cause the apparatus to at least perform
The additional elements are recited at a high-level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
The claim recites the additional elements:
wherein the at least one mask is shared by an encoder and a decoder
receive one or more weight updates corresponding to at least one parameter in a bitstream
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do no integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional elements:
an apparatus comprising at least one processor
at least one non-transitory memory comprising computer program code, wherein the at least one non-transitory memory and the computer program code are configured to, with the processor, cause the apparatus to at least perform
The additional elements are recited at a high-level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
The claim recites the additional elements:
wherein the at least one mask is shared by an encoder and a decoder
receive one or more weight updates corresponding to at least one parameter in a bitstream
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 54:
With respect to claim 54, the claim depends upon claim 53. The analysis of claim 53 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
recover the one or more weight updates from the bitstream by using the received number of the one or more parameters and one or more ordering lists (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation to recover the one or more weight updates from the bitstream by using the received number of the one or more parameters and one or more ordering lists)
wherein the one or more ordering lists indicate the ordering of the importance of the one or more parameters (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation to identify the importance of the one or more parameters)
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional elements:
receive a number of one or more parameters that are updated at the model finetuning and one or more corresponding weight updates
wherein the one or more ordering list is shared by the encoder and the decoder
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
The claim recites the additional element:
wherein an ordering of importance of the one or more parameters in the decoder is learnt using the training dataset
The decoder is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do no integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional elements:
receive a number of one or more parameters that are updated at the model finetuning and one or more corresponding weight updates
wherein the one or more ordering list is shared by the encoder and the decoder
The additional elements amount to data gathering which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
The claim recites the additional element:
wherein an ordering of importance of the one or more parameters in the decoder is learnt using the training dataset
The decoder is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 55:
With respect to claim 55, the claim depends upon claim 53. The analysis of claim 53 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
reconstruct a first number of parameters and a respective first number of weight updates from the bitstream (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation to reconstruct a first number of parameters and a respective first number of weight updates)
reconstruct a second number of parameters and a respective second number of weight updates from the bitstream, wherein a second unit follows the first unit in a bitstream order, and the second number of parameters follow the first number of parameters in an importance ordering list (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation to reconstruct a second number of parameters and a respective second number of weight updates from the bitstream, wherein a second unit follows the first unit in a bitstream order, and the second number of parameters follow the first number of parameters in an importance ordering list)
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional elements:
update a corresponding first number of parameters in a model according to the one or more ordering list
use an updated neural network for decoding a first unit of the bitstream
update corresponding second number parameters in the model according to the one or more ordering list
use the updated neural network for decoding the second unit of the bitstream
The decoder is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do no integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional elements:
update a corresponding first number of parameters in a model according to the one or more ordering list
use an updated neural network for decoding a first unit of the bitstream
update corresponding second number parameters in the model according to the one or more ordering list
use the updated neural network for decoding the second unit of the bitstream
The decoder is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claims 56-58:
With respect to claims 56-58, the claims recite the limitations substantially similar to those in claims 53-54 and 39, respectively. Claims 56-58 are rejected under similar rationale.
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 56 and 58 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Afrabandpay et al. (Response to MPEG call for proposals on incremental weight update compression, April 2021, hereafter Afrabandpay).
As per independent claim 56, Afrabandpay discloses:
receive one or more weight updates corresponding to at least one parameter in a bitstream
reconstruct at least one weight update from the bitstream (Sections 2.4 and 2.5: Here, the mask data and weight data is shared with the server for encoding/decoding)
update corresponding to at least one parameter is learnt by using a training dataset (Section 2.4: Here, a plurality of iterations is used to update the global weights sent from the server to the client)
wherein importance of the at least one parameter is learnt by using a training dataset (Section 2.1: Here, an importance function F calculates the weight update importance based on a set up elements (training dataset))
wherein one or more masks are defined for indicating importance of one or more parameters for a model finetuning (Section 2.3: Here, a quantization algorithm defines a gradient change bitmask)
wherein the at least one mask is shared by an encoder and decoder (Sections 2.4 and 2.5: Here, the mask data and weight data is shared with the server for encoding/decoding)
As per independent claim 58, Afrabandpay discloses an apparatus (page 2, paragraph 1: Here, a server and client apparatus are disclosed) cause the apparatus to perform:
learn importance of one or more parameters by using a training dataset (Section 2.1: Here, an importance function F calculates the weight update importance based on a set up elements (training dataset))
define one or more masks for indicating an importance of the one or more parameters for a model finetuning (Section 2.3: Here, a quantization algorithm defines a gradient change bitmask)
share at least one mask of the one or more masks with at least one of an encoder or a decoder (Sections 2.4 and 2.5: Here, the mask data and weight data is shared with the server for encoding/decoding)
finetune at least one parameters of the one or more parameters based at least on the at least one mask (Section 2.4: Here, a plurality of iterations is used to update the global weights sent from the server to the client)
send or signal one or more weight updates corresponding to the at least one parameters in a bitstream to the decoder (Sections 2.4 and 2.5: Here, the mask data and weight data is shared with the server for encoding/decoding)
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 39, 45, and 53 are rejected under 35 U.S.C. 103 as being unpatentable over Afrabandpay and further in view of Abe et al. (US 2019/0273921, published 5 September 20219, hereafter Abe).
As per independent claim 39, Afrabandpay discloses an apparatus (page 2, paragraph 1: Here, a server and client apparatus are disclosed) cause the apparatus to perform:
learn importance of one or more parameters by using a training dataset (Section 2.1: Here, an importance function F calculates the weight update importance based on a set up elements (training dataset))
define one or more masks for indicating an importance of the one or more parameters for a model finetuning (Section 2.3: Here, a quantization algorithm defines a gradient change bitmask)
share at least one mask of the one or more masks with at least one of an encoder or a decoder (Sections 2.4 and 2.5: Here, the mask data and weight data is shared with the server for encoding/decoding)
finetune at least one parameters of the one or more parameters based at least on the at least one mask (Section 2.4: Here, a plurality of iterations is used to update the global weights sent from the server to the client)
send or signal one or more weight updates corresponding to the at least one parameters in a bitstream to the decoder (Sections 2.4 and 2.5: Here, the mask data and weight data is shared with the server for encoding/decoding)
Afrabandpay fails to specifically disclose:
at least one processor
at least one non-transitory memory comprising computer program code, wherein the at least one non-transitory memory and the computer program code are configured to, with the at least one processor
However, Abe, which is analogous to the claimed invention because it is directed toward encoding/decoding, discloses:
at least one processor (paragraph 0467)
at least one non-transitory memory comprising computer program code, wherein the at least one non-transitory memory and the computer program code are configured to, with the at least one processor (paragraph 0008)
It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Abe with Afrabandpay, with a reasonable expectation of success, as it would have allowed for implementing an encoder/decoder in a system including memory and processing hardware (Abe: paragraph 0007).
As per dependent claim 45, Afrabandpay and Abe disclose the limitations similar to those in claim 39, and the same rejection is incorporated herein. Afrabandpay discloses wherein the apparatus is further caused to:
determine an optimal mask of the one or more masks (Section 2.3: Here, a quantization algorithm defines a gradient change bitmask)
determine optimal weight updates corresponding to the optimal mask at a model finetuning stage (Sections 2.4 and 2.5: Here, the mask data and weight data is shared with the server for encoding/decoding)
send or signal an index of the optimal mask and the optimal weight updates corresponding to the optimal mask to the decoder, wherein the decoder reconstructs the index of the optimal mask and the optimal weight updates from the bitstream and updates the parameters corresponding to the optimal weight updates in a model according to the optimal mask (Section 2.4: Here, using predictive residual estimation, the amount of data transferred between clients/servers is limited. This is done based upon providing a global weight/mask and predicting the client weights at a time t)
As per independent claim 53, Afrabandpay discloses:
receive one or more weight updates corresponding to at least one parameter in a bitstream
reconstruct at least one weight update from the bitstream (Sections 2.4 and 2.5: Here, the mask data and weight data is shared with the server for encoding/decoding)
update corresponding to at least one parameter is learnt by using a training dataset (Section 2.4: Here, a plurality of iterations is used to update the global weights sent from the server to the client)
wherein importance of the at least one parameter is learnt by using a training dataset (Section 2.1: Here, an importance function F calculates the weight update importance based on a set up elements (training dataset))
wherein one or more masks are defined for indicating importance of one or more parameters for a model finetuning (Section 2.3: Here, a quantization algorithm defines a gradient change bitmask)
wherein the at least one mask is shared by an encoder and decoder (Sections 2.4 and 2.5: Here, the mask data and weight data is shared with the server for encoding/decoding)
Afrabandpay fails to specifically disclose:
at least one processor
at least one non-transitory memory comprising computer program code, wherein the at least one non-transitory memory and the computer program code are configured to, with the at least one processor
However, Abe, which is analogous to the claimed invention because it is directed toward encoding/decoding, discloses:
at least one processor (paragraph 0467)
at least one non-transitory memory comprising computer program code, wherein the at least one non-transitory memory and the computer program code are configured to, with the at least one processor (paragraph 0008)
It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Abe with Afrabandpay, with a reasonable expectation of success, as it would have allowed for implementing an encoder/decoder in a system including memory and processing hardware (Abe: paragraph 0007).
Claims 40-43 and 46-48 are rejected under 35 U.S.C. 103 as being unpatentable over Afrabandpay and Abe and further in view of Micaelian et al. (US 7836057, patented 16 November 2010, hereafter Micaelian).
As per dependent claim 40, Afrabandpay and Abe disclose the limitations similar to those in claim 39, and the same rejection is incorporated herein. Afrabandpay discloses:
learn an ordering of the importance of the one or more parameters in the decoder by using the training dataset (Section 2.1: Here, an importance function F calculates the weight update importance based on a set up elements (training dataset))
send or signal at least one parameters of the one or more parameters that are updated by the model finetuning and at least one weight update corresponding to the at least one parameters to the decoder, wherein the decoder recovers the at least one weight update from the bitstream by using the received at least one parameter (Sections 2.4 and 2.5: Here, the mask data and weight data is shared with the server for encoding/decoding)
Afrabandpay fails to specifically disclose:
define one or more ordering lists for indicating the ordering of the importance of the one or more parameters
share data with the at least one of the encoder or the decoder (paragraph 0181-0183: Here, the decoder receives the data for processing)
Abe discloses share data with the at least one of the encoder or the decoder (paragraph 0181-0183: Here, the decoder receives the data for processing). It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Abe with Afrabandpay, with a reasonable expectation of success, as it would have allowed for decoding input by a decoder (Abe: paragraph 0181).
Additionally, Micaelian, which is analogous to the claimed invention because it is directed toward sorting a list based upon weights, discloses define one or more ordering lists for indicating the ordering of the importance of the one or more parameters (claim 9: Here, an ordered list is generated based upon parameters and their associated importance weighting). It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Micaelian with Afrabandpay-Abe, with a reasonable expectation of success, as it would have allowed for providing an ordered list based upon importance weighting (Micaelian: claim 9).
As per dependent claim 41, Afrabandpay, Abe, and Micaelian disclose the limitations similar to those in claim 40, and the same rejection is incorporated herein. Afrabandpay discloses wherein the apparatus is further caused to determine at least one of:
the at least one mask of the one or more masks to be used by the encoder during the model finetuning (Sections 2.4 and 2.5: Here, the mask data and weight data is shared with the server for encoding/decoding) or
at least one ordering list of the one or more ordering lists to be used by the encoder during a model finetuning stage
As per dependent claim 42, Afrabandpay, Abe, and Micaelian disclose the limitations similar to those in claim 40, and the same rejection is incorporated herein. Afrabandpay discloses wherein to learn the importance of one or more parameters or to learn an ordering of the importance of the one or more parameters, the apparatus is further caused to:
minimize an objective function with a suitable loss term (Section 2.1(b): Here, elements that are determined to be less important have their weights set to zero and are not considered (dropped))
determine one or more importance scores for each parameter based on a minimized objective function (Section 2.2: Here, weights (importance scores) for each parameter are updated based upon the sparsification)
determine at least one of:
the at least one mask by using one or more importance scores and a threshold value (Sections 2.1(b) and 2.2: Here, in order to identify an element as “less important,” some implicit threshold must be met/not met. These “less important” elements are updated to zero. The scaling values are then updated based upon the sparsification)
or
the at least one ordering list by sorting the one or more parameters based on the one or more importance scores of the each one or more parameters
As per dependent claim 43, Afrabandpay, Abe, and Micaelian disclose the limitations similar to those in claim 40, and the same rejection is incorporated herein. Afrabandpay discloses wherein the one or more importance scores for each parameters comprises one of:
a sum of absolute value of the weight updated value of each parameter for samples in the training dataset
a sum of an absolute value of a ratio of the weight update value over an original weight value on the each parameter for samples in the training dataset
a combination of an importance metric and a learned importance, wherein the combination is using a summation or multiplication of scores of the one or more parameters (Sections 2.1 and 2.2: Here, a mean value is determined for non-zero values when determining the importance. In order to calculate a mean value, all non-zero elements must be summed (summation) and divided by the number of non-zero elements)
a non-linear combination of the importance metric and the learned importance, wherein the non-linear combination is determined using a linear based approach
As per dependent claim 46, Afrabandpay, Abe, and Micaelian disclose the limitations similar to those in claim 40, and the same rejection is incorporated herein. Afrabandpay discloses:
determine an optimal number of parameters from the one or more parameters (Section 2.2: Here, parameters are sparsified based upon their weights (importance scores). The examiner interprets the number of parameters that are not sparsified as being the “optimal number of parameters”)
determine an optimal number of weight updates corresponding to the optimal parameters at a model finetuning stage (Sections 2.4 and 2.5: Here, the mask data and weight data is shared with the server for encoding/decoding)
send or signal an index of the optimal ordering list, the optimal number of parameters, and the optimal weight updates to the decoder, wherein the decoder reconstructs the index of the optimal ordering list, the optimal number of parameters, and the optimal weight updates from the bitstream, and updates the parameters corresponding to the optimal weight updates in a model according to the optimal ordering list and the optimal number of parameters (Section 2.4: Here, using predictive residual estimation, the amount of data transferred between clients/servers is limited. This is done based upon providing a global weight/mask and predicting the client weights at a time t)
Afrabandpay fails to specifically disclose determine an optimal ordering list from the one or more ordering lists. Additionally, Micaelian, which is analogous to the claimed invention because it is directed toward sorting a list based upon weights, discloses determining an optimal ordering list from the one or more ordering lists (claim 9: Here, an ordered list is generated based upon parameters and their associated importance weighting. The examiner interprets this ordered list as being the “optimal ordering list from the one or more ordering list.” In this instance, it is the optimal list among a group of one possible ordering list). It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Micaelian with Afrabandpay-Abe, with a reasonable expectation of success, as it would have allowed for providing an ordered list based upon importance weighting (Micaelian: claim 9).
As per dependent claim 47, Afrabandpay, Abe, and Micaelian disclose the limitations similar to those in claim 40, and the same rejection is incorporated herein. Afrabandpay discloses:
learn the one or more masks or one or more ordering lists from an input content (Section 2.3: Here, a quantization algorithm defines a gradient change bitmask)
signal the learned one or more masks or ordering lists to the decoder (Section 2.4: Here, using predictive residual estimation, the amount of data transferred between clients/servers is limited. This is done based upon providing a global weight/mask and predicting the client weights at a time t)
As per dependent claim 48, Afrabandpay, Abe, and Micaelian disclose the limitations similar to those in claim 40, and the same rejection is incorporated herein. Afrabandpay discloses wherein to share the at least one mask and the at least one ordered list, the apparatus is further configured to perform one or more of:
define at least one of the one or more masks or the one or more ordered lists (Section 2.3: Here, a quantization algorithm defines a gradient change bitmask)
define a unique identifier for each mask or each ordering list, and indicate unique identifiers of the each mask or the each ordering list in or along the bitstream from the encoder to the decoder
deliver the at least one of the at least one mask or the at least one ordering list in or along the bitstream (Figure 1: Here, the bitstream is used to provide weight updates/masks to the encoder)
deliver the at least one of the at least one mask or the at least one ordering list (Figure 1) or
deliver the at least one of the at least one mask or the at least one ordering list from the encoder to the decoder prior to or during delivery of the bitstream (Figure 1)
Claims 44, 49-50, and 52 are rejected under 35 U.S.C. 103 as being unpatentable over Afrabandpay, Abe, and Micaelian and further in view of Schwartz et al. (US 2022/0012022, published 13 January 2022, hereafter Schwartz).
As per dependent claim 44, Afrabandpay, Abe, and Micaelian disclose the limitations similar to those in claim 40, and the same rejection is incorporated herein. Afrabandpay discloses wherein the apparatus is further configured to identify an optimal number of the one or more parameters and the corresponding weight updates (Sections 2.1 and 2.2: Here, a sparsification is performed to identify parameters based upon importance and a sparsification percentage. This sparsification percentage defines the optimal number of parameters)
Afrabandpay fails to specifically disclose wherein to identify the optimal number of the one or more parameters and the corresponding weight updates, the apparatus if further caused to:
perform a first model finetuning by selecting a first set of important parameters based on the one or more ordering lists
save the first finetuning model to achieve a first saved finetuning model
calculate a first rate distortion (RD) loss value including weight update overheads
repeat following until a stopping criterion is met:
remove a predetermined number of important parameters from the first set of important parameters
continue the model finetuning from the first saved finetuned model to achieve a second finetuned model
save the second finetuned model
calculate a second RD loss value including the weight update overheads
Micaelian, which is analogous to the claimed invention because it is directed toward sorting a list based upon weights, discloses perform a first model finetuning by selecting a first set of important parameters based on the one or more ordering lists (claim 9: Here, an ordered list is generated based upon parameters and their associated importance weighting). It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Micaelian with Afrabandpay-Abe, with a reasonable expectation of success, as it would have allowed for providing an ordered list based upon importance weighting (Micaelian: claim 9).
Further, Schwartz, which is analogous to the claimed invention because it is directed toward deep learning using sparsification, discloses:
save the first finetuning model to achieve a first saved finetuning model (Figure 9: Here, upon each iteration of the input and testing, a finetuning model is generated)
calculate a first rate distortion (RD) loss value including weight update overheads (Figure 9, item 908; paragraphs 0135-0144: Here, a testing set is decomposed into a test and validation set. A difference value is calculated between these two sets (first rate distortion loss value))
repeat following until a stopping criterion is met:
remove a predetermined number of important parameters from the first set of important parameters (Figure 9; paragraph 0027: Here, the model is tuned based upon the sparsified matrices)
continue the model finetuning from the first saved finetuned model to achieve a second finetuned model (Figure 9: Here, after each iteration a new finetuned model is created and saved for further tuning)
save the second finetuned model (Figure 9)
calculate a second RD loss value including the weight update overheads (Figure 9, item 908; paragraphs 0135-0144: Here, each iteration of the model causes a new RD loss value to be calculated)
It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Schwartz with Afrabandpay-Abe-Micaelian, with a reasonable expectation of success, as it would have allowed for optimizing a model based upon user preferences (Schwartz: Abstract; paragraph 0001).
As per dependent claim 49, Afrabandpay, Abe, and Micaelian disclose the limitations similar to those in claim 40, and the same rejection is incorporated herein. Afrabandpay discloses wherein the apparatus is further caused to:
select a first number of parameters to be updated for a first unit of the bitstream updates (Sections 2.1 and 2.2: Here, a sparsification is performed to identify parameters based upon importance and a sparsification percentage. This sparsification is an update to a first number of parameters of the bitstream)
indicate the first number of parameters and the updated first number of parameters associated with the first unit of the bitstream (Sections 2.1 and 2.2: Here, a sparsification is performed to identify parameters based upon importance and a sparsification percentage)
Afrabandpay fails to specifically disclose:
use an updated neural network for encoding the first unit of the bitstream
select a second number of parameters to be updated for a second unit of the bitstream, wherein the second unit follows the first unit in the bitstream order, and the second number of parameters follow the first number of parameters in the at least one ordering list
indicate the second number of parameters and the updated second number of parameters associated with the second unit of the bitstream
use the updated neural network for encoding the second unit of the bitstream
Micaelian, which is analogous to the claimed invention because it is directed toward sorting a list based upon weights, parameters follow the first number of parameters in the at least one ordering list (claim 9: Here, an ordered list is generated based upon parameters and their associated importance weighting). It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Micaelian with Afrabandpay-Abe, with a reasonable expectation of success, as it would have allowed for providing an ordered list based upon importance weighting (Micaelian: claim 9).
Additionally, Schwartz, which is analogous to the claimed invention because it is directed toward deep learning using sparsification, discloses:
use an updated neural network for encoding the first unit (Figure 9; paragraph 0031: Here, a neural network is updated based upon the sparsification generated by a first set of testing and validation)
select a second number of parameters to be updated for a second unit of the bitstream, wherein the second unit follows the first unit in the order, and the second number of parameters follow the first number of parameters (Figure 9: Here, a next available (second number of parameters) set of testing/validation data is entered to adjust the latent feature values of the next layer)
indicate the second number of parameters and the updated second number of parameters associated with the second unit (Figure 9)
It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Schwartz with Afrabandpay-Abe-Micaelian, with a reasonable expectation of success, as it would have allowed for optimizing a model based upon user preferences (Schwartz: Abstract; paragraph 0001).
As per dependent claim 50, Afrabandpay, Abe, Micaelian, and Schwartz disclose the limitations similar to those in claim 49, and the same rejection is incorporated herein. Schwartz discloses wherein the apparatus is further caused to select at least one of the first number of parameters or the second number of parameters according to one or more of the following:
a bit budget for the updated first number of parameters or the updated second number of parameters or
a rate-distortion decision performed for the first unit and the first number of parameters separately from the second unit and the second number of parameters, wherein the first unit and the second unit may be selected to be first coded picture and a second code picture respectively (Figure 9; paragraph 0130: Here, a visual element is processed. This includes user/visual element interaction matrices to test and validate the sparse matrices and gauge a difference between the latent features sparse matrices (rate distortion) or a first hierarchical prediction structure of picture and a second hierarchical prediction structure of pictures
It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Schwartz with Afrabandpay-Abe-Micaelian, with a reasonable expectation of success, as it would have allowed for optimizing a model based upon user preferences (Schwartz: Abstract; paragraph 0001).
As per dependent claim 52, Afrabandpay, Abe, Micaelian, and Schwartz disclose the limitations similar to those in claim 49, and the same rejection is incorporated herein. Schwartz discloses wherein the apparatus is further caused to determine and indicate in or along the bitstream:
whether the second number of parameters follow the first number of parameters in the at least one ordering list (Figure 9: Here, a next available (second number of parameters) set of testing/validation data is entered to adjust the latent feature values of the next layer) or
whether the second number of parameters are related to a start of the at least one ordering list
wherein the decoder decodes associated weight updates according to an indicated position in the at least one ordering list
It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Schwartz with Afrabandpay-Abe-Micaelian, with a reasonable expectation of success, as it would have allowed for optimizing a model based upon user preferences (Schwartz: Abstract; paragraph 0001).
Claims 51 is rejected under 35 U.S.C. 103 as being unpatentable over Afrabandpay, Abe, Micaelian, and Schwartz and further in view of Thirumalai (US 2021/0312666, published 7 October 2021).
As per dependent claim 51, Afrabandpay, Abe, Micaelian, and Schwartz disclose the limitations similar to those in claim 50, and the same rejection is incorporated herein. Afrabandpay fails to specifically disclose wherein the bit budget is determined based on one or more of the following:
a leftover bitrate in a previous unit the bitstream or
a leftover size in video packets; wherein in a packet-oriented network one or more packets up to a maximum transfer unit (MTU) size are treated equally, and wherein the bit budget is determined to fill in a packet up to an MTU size in an instance a unit of the bitstream carried in the packet has size less than the MTU size
However, Thiraumalia, which is analogous to the claimed invention because it is directed toward a bit budget, discloses wherein the bit budget is determined based on one or more of the following:
a leftover bitrate in a previous unit the bitstream (paragraph 0006: Here, a remaining bit budget (leftover bitrate) is determined by subtracting a length of the encoded subunit from a bit budget)
a leftover size in video packets; wherein in a packet-oriented network one or more packets up to a maximum transfer unit (MTU) size are treated equally, and wherein the bit budget is determined to fill in a packet up to an MTU size in an instance a unit of the bitstream carried in the packet has size less than the MTU size
It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Thirumalai with Afrabandpay-Abe-Micaelian-Schwartz, with a reasonable expectation of success, as it would have allowed for determining a remaining bit budget and whether it is sufficient to perform an operation (Thirumalai: Abstract).
Claims 54 and 57 are rejected under 35 U.S.C. 103 as being unpatentable over Afrabandpay and further in view of Micaelian.
As per dependent claim 54, Afrabandpay discloses the limitations similar to those in claim 53, and the same rejection is incorporated herein. Afrabandpay discloses:
receive a number of one or more parameters that are updated at the model finetuning and one or more corresponding weight updates (Section 2.1: Here, an importance function F calculates the weight update importance based on a set up elements (training dataset))
recover the one or more weight updates from the bitstream by using the received number of the one or more parameters (Section 2.1: Here, an importance function F calculates the weight update importance based on a set up elements (training dataset))
wherein an ordering of importance of the one or more parameters in the decoder is learnt by using the training dataset (Section 2.1: Here, an importance function F calculates the weight update importance based on a set up elements (training dataset))
wherein the one or more ordering lists indicate the ordering of the importance of the one or more parameters (Sections 2.1 and 2.2: Here, a sparsification is performed to identify parameters based upon importance and a sparsification percentage)
wherein the one or more ordering list is shared by the encoder and the decoder (Sections 2.4 and 2.5: Here, the mask data and weight data is shared with the server for encoding/decoding)
Afrabandpay fails to specifically disclose one or more ordering lists.
However, Micaelian, which is analogous to the claimed invention because it is directed toward sorting a list based upon weights, discloses define one or more ordering lists for indicating the ordering of the importance of the one or more parameters (claim 9: Here, an ordered list is generated based upon parameters and their associated importance weighting). It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Micaelian with Afrabandpay-Abe, with a reasonable expectation of success, as it would have allowed for providing an ordered list based upon importance weighting (Micaelian: claim 9).
With respect to claim 57, the claim recites the limitations substantially similar to those in claim 54. Claim 57 is rejected under similar rationale.
Claim 55 is rejected under 35 U.S.C. 103 as being unpatentable over Afrabandpay and Micaelian, and further in view of Schwartz.
As per dependent claim 55, Afrabandpay and Abe disclose the limitations similar to those in claim 53 and the same rejection is incorporated herein. Afrabandpay discloses:
reconstruct a first number of parameters and a respective first number of weight updates from the bitstream (Section 2.4: Here, using predictive residual estimation, the amount of data transferred between clients/servers is limited. This is done based upon providing a global weight/mask and predicting the client weights at a time t)
update a corresponding first number of parameters in a model (Sections 2.4 and 2.5: Here, the mask data and weight data is shared with the server for encoding/decoding)
Afrabandpay fails to specifically disclose:
one or more ordering list
use an updated neural network for decoding a first unit of the bitstream
reconstruct a second number of parameters and a respective second number of weight updates from the bitstream, wherein a second unit follows the first unit in a bitstream order, and the second number of parameters follow the first number of parameters in an importance ordering list
update corresponding second number parameters in the model according to the one or more ordering lists
use the updated neural network for decoding the second unit of the bitstream
However, Micaelian, which is analogous to the claimed invention because it is directed toward sorting a list based upon weights, discloses define one or more ordering lists for indicating the ordering of the importance of the one or more parameters (claim 9: Here, an ordered list is generated based upon parameters and their associated importance weighting). It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Micaelian with Afrabandpay-Abe, with a reasonable expectation of success, as it would have allowed for providing an ordered list based upon importance weighting (Micaelian: claim 9).
Additionally, Schwartz discloses:
use an updated neural network for decoding a first unit (Figure 9: Here, a second iteration of test/validation user/visual element interaction matrices is performed)
reconstruct a second number of parameters and a respective second number of weight updates from the bitstream, wherein a second unit follows the first unit in a bitstream order, and the second number of parameters follow the first number of parameters (Figure 9: Here, a next available (second number of parameters) set of testing/validation data is entered to adjust the latent feature values of the next layer)
update corresponding second number parameters in the model (Figure 9: Here, based upon the second iteration, parameters are updated to finetune weight matrices)
use the updated neural network for decoding the second unit (Figure 9)
It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Schwartz with Afrabandpay-Abe-Micaelian, with a reasonable expectation of success, as it would have allowed for optimizing a model based upon user preferences (Schwartz: Abstract; paragraph 0001).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Georgescu et al. (US 9668699): Discloses sparsifying to tune a neural network (column 24, lines 23-43)
Sasturkar et al. (US 2017/0147418): Discloses generating a list using an importance parameter (paragraph 0074)
Bobley et al. (US 11087409): Discloses generating a result list based upon an importance parameter (column 11, lines 39-51)
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KYLE R STORK whose telephone number is (571)272-4130. The examiner can normally be reached 8am - 2pm; 4pm - 6pm.
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, Omar Fernandez Rivas can be reached at 571/272-2589. 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.
/KYLE R STORK/Primary Examiner, Art Unit 2128
Prosecution Insights