DETAILED ACTION
This action is in response to the application filed 01/31/2024. Claims 1-20 are pending and have been examined.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
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.
Claim 1-20 rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding Claim 1:
Subject Matter Eligibility Analysis Step 1:
Claim 1 recites a method and is thus a process, one of the four statutory categories of patentable subject matter.
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 1 recites
transforming the time series data into a user-item graph; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
encoding, …the user-item graph into user embeddings and item embeddings (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
generating a plurality of losses according to a plurality of training tasks performed based on the user embeddings and, item embeddings; (i.e., the (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
generating, …one or more recommended items for a given user based on the updated user embeddings and the updated item embeddings; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
Therefore, claim 1 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 1 further recites additional elements of
the method comprising: receiving, … time series data indicating a user-item interaction history; (i.e., the broadest reasonable interpretation of receiving a data instance is mere data gathering, which is an insignificant extra solution activity (MPEP 2106.05(g))).
training, …by updating the user embeddings and the item embeddings via backpropagation based on a weighted sum of gradients of the plurality of losses (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…via a data interface… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…by a neural network encoder… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…the recommendation neural network model… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…by a neural network decoder… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
Therefore, claim 1 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
The additional elements of claim 1 do not provide significantly more than the abstract idea itself, taken alone and in combination because
the method comprising: receiving, … time series data indicating a user-item interaction history is the well understood, routine, and conventional activity of “transmitting or receiving data over a network” (see MPEP 2106.05(d)(II); OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network)).
training, …by updating the user embeddings and the item embeddings via backpropagation based on a weighted sum of gradients of the plurality of losses (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…via a data interface… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…by a neural network encoder… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…the recommendation neural network model… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…by a neural network decoder… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
Regarding Claim 2:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 2 recites
wherein the transforming the time series data into a user-item graph comprises: creating a first node representing at least one user and a second node representing at least one item; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend from the data into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
adding an edge between the first node and the second node representing that the at least one user has interacted with the at least one item; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like creating a link between two recently interacted elements, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
and associating the edge with a score representing a preference level of the at least one user towards the at least one item. (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like creating a score for the most interacted elements, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
Therefore, claim 2 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 2 does not further recite any additional elements. Therefore, claim 2 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
Since there are no additional elements, claim 2 does not provide significantly more than the abstract idea itself, taken alone and in combination. Therefore, claim 2 is subject-matter ineligible.
Regarding Claim 3:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 3 recites the same abstract ideas as claim 1. Therefore, claim 3 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 3 further recites additional elements of
plurality of training tasks comprise any one or more of: a task of predicting an item for a user; (This element does not integrate the abstract idea into a practical application because it recites a technological environment in which to apply a judicial exception (see MPEP 2106.05(h)).)
a task of predicting a social connection for a user; (This element does not integrate the abstract idea into a practical application because it recites a technological environment in which to apply a judicial exception (see MPEP 2106.05(h)).)
and a task of predicting an item category for a user; (This element does not integrate the abstract idea into a practical application because it recites a technological environment in which to apply a judicial exception (see MPEP 2106.05(h)).)
Therefore, claim 3 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
The additional elements of claim 3 do not provide significantly more than the abstract idea itself, taken alone and in combination because
plurality of training tasks comprise any one or more of: a task of predicting an item for a user specifies a particular technological environment to perform the abstract idea and cannot provide significantly more (see MPEP 2106.05(h)).
a task of predicting a social connection for a user specifies a particular technological environment to perform the abstract idea and cannot provide significantly more (see MPEP 2106.05(h)).
and a task of predicting an item category for a user specifies a particular technological environment to perform the abstract idea and cannot provide significantly more (see MPEP 2106.05(h)).
Therefore, claim 3 is subject-matter ineligible.
Regarding Claim 4:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 4 recites the same abstract ideas as claim 1. Therefore, claim 4 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 4 further recites additional elements of
training the recommendation neural network model comprises keeping the neural network encoder unchanged during the backpropagation. (This element does not integrate the abstract idea into a practical application because it amounts to mere “apply it on a computer” (see MPEP 2106.05(f)).)
Therefore, claim 4 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
The additional elements of claim 4 do not provide significantly more than the abstract idea itself, taken alone and in combination because
training the recommendation neural network model comprises keeping the neural network encoder unchanged during the backpropagation uses a computer as a tool to perform the abstract idea and cannot provide significantly more (see MPEP 2106.05(f)).
Therefore, claim 4 is subject-matter ineligible.
Regarding Claim 5:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 5 recites
computing the plurality of gradients corresponding to the plurality of losses; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper, by calculating the plurality of gradient based on the losses, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
and modifying a magnitude of at least one gradient in the plurality of task gradients based on a current training epoch number. (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper, by changing the magnitude of one gradient based on the current training iteration, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
Therefore, claim 5 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 5 does not further recite any additional elements. Therefore, claim 5 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
Since there are no additional elements, claim 5 does not provide significantly more than the abstract idea itself, taken alone and in combination. Therefore, claim 5 is subject-matter ineligible
Regarding Claim 6:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 6 recites
generating a plurality of gradient weights from magnitudes of the plurality of gradients, including the at least one gradient with the modified magnitude; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper, by calculating the plurality of gradient weight based on the plurality of gradient magnitudes, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
and computing the weighted sum of gradients of the plurality of losses based on the generated plurality of gradient weights. (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper, by calculating the weighted sum of gradients of the plurality of losses, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
Therefore, claim 6 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 6 does not further recite any additional elements. Therefore, claim 6 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
Since there are no additional elements, claim 6 does not provide significantly more than the abstract idea itself, taken alone and in combination. Therefore, claim 6 is subject-matter ineligible.
Regarding Claim 7:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 7 recites
wherein one or more losses from the plurality of losses are computed further based on information including any one or more of: user profile information; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like computing the plurality of losses from the user profile information, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
and user social connection information. (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like computing the plurality of losses from the user social connection information, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
Therefore, claim 7 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 7 does not further recite any additional elements. Therefore, claim 7 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
Since there are no additional elements, claim 7 does not provide significantly more than the abstract idea itself, taken alone and in combination. Therefore, claim 7 is subject-matter ineligible.
Regarding Claim 8:
Subject Matter Eligibility Analysis Step 1:
Claim 8 recites a system and is thus an apparatus, one of the four statutory categories of patentable subject matter.
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 8 recites
transforming the time series data into a user-item graph; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
encoding, …the user-item graph into user embeddings and item embeddings (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
generating a plurality of losses according to a plurality of training tasks performed based on the user embeddings and, item embeddings; (i.e., the (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
generating, …one or more recommended items for a given user based on the updated user embeddings and the updated item embeddings; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
Therefore, claim 8 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 8 further recites additional elements of
A system for making recommendations, the system comprising:: a memory that stores a recommendation neural network model and a plurality of processor executable instructions; (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
a communication interface that receives time series data indicating a user-item interaction history; (i.e., the broadest reasonable interpretation of receiving a data instance is mere data gathering, which is an insignificant extra solution activity (MPEP 2106.05(g))).
and one or more, hardware processors that read and execute the plurality of processor-executable instructions from the memory to perform operations comprising (i.e., the broadest reasonable interpretation of receiving a data instance is mere data gathering, which is an insignificant extra solution activity (MPEP 2106.05(g))).
training, via backpropagation by updating the user embeddings and the item embeddings, …based on a weighted sum of gradients of the plurality of losses (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…via a data interface… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…by a neural network encoder… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…the recommendation neural network model… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…by a neural network decoder… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
Therefore, claim 8 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
The additional elements of claim 8 do not provide significantly more than the abstract idea itself, taken alone and in combination because
A system for making recommendations, the system comprising:: a memory that stores a recommendation neural network model and a plurality of processor executable instructions; (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
a communication interface that receives time series data indicating a user-item interaction history is the well understood, routine, and conventional activity of “transmitting or receiving data over a network” (see MPEP 2106.05(d)(II); OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network)).
and one or more, hardware processors that read and execute the plurality of processor-executable instructions from the memory to perform operations comprising is the well understood, routine, and conventional activity of “transmitting or receiving data over a network” (see MPEP 2106.05(d)(II); OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network)).
training, via backpropagation by updating the user embeddings and the item embeddings, …based on a weighted sum of gradients of the plurality of losses (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…via a data interface… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…by a neural network encoder… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…the recommendation neural network model… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…via backpropagation… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…by a neural network decoder… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
Regarding Claim 9:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 9 recites
The operation of transforming the time series data into a user-item graph comprises: creating a first node representing at least one user and a second node representing at least one item; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend from the data into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
adding an edge between the first node and the second node representing that the at least one user has interacted with the at least one item; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like creating a link between two recently interacted elements, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
and associating the edge with a score representing a preference level of the at least one user towards the at least one item. (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like creating a score for the most interacted elements, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
Therefore, claim 9 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 9 does not further recite any additional elements. Therefore, claim 9 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
Since there are no additional elements, claim 9 does not provide significantly more than the abstract idea itself, taken alone and in combination. Therefore, claim 9 is subject-matter ineligible.
Regarding Claim 10:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 10 recites the same abstract ideas as claim 8. Therefore, claim 10 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 10 further recites additional elements of
wherein plurality of training tasks comprise any one or more of: a task of predicting an item for a user; (This element does not integrate the abstract idea into a practical application because it recites a technological environment in which to apply a judicial exception (see MPEP 2106.05(h)).)
a task of predicting a social connection for a user; (This element does not integrate the abstract idea into a practical application because it recites a technological environment in which to apply a judicial exception (see MPEP 2106.05(h)).)
and a task of predicting an item category for a user; (This element does not integrate the abstract idea into a practical application because it recites a technological environment in which to apply a judicial exception (see MPEP 2106.05(h)).)
Therefore, claim 10 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
The additional elements of claim 10 do not provide significantly more than the abstract idea itself, taken alone and in combination because
plurality of training tasks comprise any one or more of: a task of predicting an item for a user specifies a particular technological environment to perform the abstract idea and cannot provide significantly more (see MPEP 2106.05(h)).
a task of predicting a social connection for a user specifies a particular technological environment to perform the abstract idea and cannot provide significantly more (see MPEP 2106.05(h)).
and a task of predicting an item category for a user specifies a particular technological environment to perform the abstract idea and cannot provide significantly more (see MPEP 2106.05(h)).
Therefore, claim 10 is subject-matter ineligible.
Regarding Claim 11:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 11 recites the same abstract ideas as claim 8. Therefore, claim 11 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 11 further recites additional elements of
the operation of training the recommendation neural network model comprises keeping the neural network encoder unchanged during the backpropagation. (This element does not integrate the abstract idea into a practical application because it amounts to mere “apply it on a computer” (see MPEP 2106.05(f)).)
Therefore, claim 11 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
The additional elements of claim 11 do not provide significantly more than the abstract idea itself, taken alone and in combination because
training the recommendation neural network model comprises keeping the neural network encoder unchanged during the backpropagation uses a computer as a tool to perform the abstract idea and cannot provide significantly more (see MPEP 2106.05(f)).
Therefore, claim 11 is subject-matter ineligible.
Regarding Claim 12:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 12 recites
the operation of training the recommendation neural network comprising: computing the plurality of gradients corresponding to the plurality of losses; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper, by calculating the plurality of gradient based on the losses, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
and modifying a magnitude of at least one gradient in the plurality of task gradients based on a current training epoch number. (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper, by changing the magnitude of one gradient based on the current training iteration, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
Therefore, claim 12 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 12 does not further recite any additional elements. Therefore, claim 12 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
Since there are no additional elements, claim 12 does not provide significantly more than the abstract idea itself, taken alone and in combination. Therefore, claim 12 is subject-matter ineligible
Regarding Claim 13:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 13 recites
generating a plurality of gradient weights from magnitudes of the plurality of gradients, including the at least one gradient with the modified magnitude; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper, by calculating the plurality of gradient weight based on the plurality of gradient magnitudes, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
and computing the weighted sum of gradients of the plurality of losses based on the generated plurality of gradient weights. (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper, by calculating the weighted sum of gradients of the plurality of losses, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
Therefore, claim 13 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 13 does not further recite any additional elements. Therefore, claim 13 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
Since there are no additional elements, claim 13 does not provide significantly more than the abstract idea itself, taken alone and in combination. Therefore, claim 13 is subject-matter ineligible.
Regarding Claim 14:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 14 recites
wherein one or more losses from the plurality of losses are computed further based on information including any one or more of: user profile information; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like computing the plurality of losses from the user profile information, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
and user social connection information. (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like computing the plurality of losses from the user social connection information, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
Therefore, claim 14 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 14 does not further recite any additional elements. Therefore, claim 14 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
Since there are no additional elements, claim 14 does not provide significantly more than the abstract idea itself, taken alone and in combination. Therefore, claim 14 is subject-matter ineligible.
Regarding Claim 15:
Subject Matter Eligibility Analysis Step 1:
Claim 15 recites A non-transitory machine-readable medium. The claim recites a non-transitory computer readable medium which is interpreted as an article of manufacture. An article of manufacture is one of the four statutory categories of invention.
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 15 recites
transforming time series data indicating a user-item interaction history into a user item graph; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
encoding, …the user-item graph into user embeddings and item embeddings (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
generating a plurality of losses according to a plurality of training tasks performed based on the user embeddings and, item embeddings; (i.e., the (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
generating, …one or more recommended items for a given user based on the updated user embeddings and the updated item embeddings; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
Therefore, claim 15 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 15 further recites additional elements of
A non-transitory machine-readable medium comprising a plurality of machine-executable instructions which, when executed by one or more processors, are adapted to cause the one or more processors to perform operations comprising: (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
training, …by updating the user embeddings and the item embeddings, via backpropagation based on a weighted sum of gradients of the plurality of losses (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…by a neural network encoder… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…the recommendation neural network model… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…by a neural network decoder… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
Therefore, claim 15 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
The additional elements of claim 15 do not provide significantly more than the abstract idea itself, taken alone and in combination because
A system for making recommendations, the system comprising:: a memory that stores a recommendation neural network model and a plurality of processor executable instructions; (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
a communication interface that receives time series data indicating a user-item interaction history is the well understood, routine, and conventional activity of “transmitting or receiving data over a network” (see MPEP 2106.05(d)(II); OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network)).
and one or more, …that read and execute the plurality of processor-executable instructions from the memory to perform operations comprising is the well understood, routine, and conventional activity of “transmitting or receiving data over a network” (see MPEP 2106.05(d)(II); OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network)).
training, …by updating the user embeddings and the item embeddings, via backpropagation based on a weighted sum of gradients of the plurality of losses (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…by a neural network encoder… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…the recommendation neural network model… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
…by a neural network decoder… (i.e., the generic computer components recited in this limitation merely add the words “apply it”, or an equivalent, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))).
Regarding Claim 16:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 16 recites
wherein the transforming the time series data indicating a user-item interaction history into a user-item graph comprises: creating a first node representing at least one user and a second node representing at least one item; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like converting a pattern or trend from the data into a graph, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
adding an edge between the first node and the second node representing that the at least one user has interacted with the at least one item; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like creating a link between two recently interacted elements, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
and associating the edge with a score representing a preference level of the at least one user towards the at least one item. (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper like creating a score for the most interacted elements, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
Therefore, claim 16 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 16 does not further recite any additional elements. Therefore, claim 16 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
Since there are no additional elements, claim 2 does not provide significantly more than the abstract idea itself, taken alone and in combination. Therefore, claim 16 is subject-matter ineligible.
Regarding Claim 17:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 17 recites the same abstract ideas as claim 15. Therefore, claim 17 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 17 further recites additional elements of
wherein plurality of training tasks comprise any one or more of: a task of predicting an item for a user; (This element does not integrate the abstract idea into a practical application because it recites a technological environment in which to apply a judicial exception (see MPEP 2106.05(h)).)
a task of predicting a social connection for a user; (This element does not integrate the abstract idea into a practical application because it recites a technological environment in which to apply a judicial exception (see MPEP 2106.05(h)).)
and a task of predicting an item category for a user; (This element does not integrate the abstract idea into a practical application because it recites a technological environment in which to apply a judicial exception (see MPEP 2106.05(h)).)
Therefore, claim 17 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
The additional elements of claim 17 do not provide significantly more than the abstract idea itself, taken alone and in combination because
wherein the plurality of training tasks comprise any one or more of: a task of predicting an item for a user specifies a particular technological environment to perform the abstract idea and cannot provide significantly more (see MPEP 2106.05(h)).
a task of predicting a social connection for a user specifies a particular technological environment to perform the abstract idea and cannot provide significantly more (see MPEP 2106.05(h)).
and a task of predicting an item category for a user specifies a particular technological environment to perform the abstract idea and cannot provide significantly more (see MPEP 2106.05(h)).
Therefore, claim 17 is subject-matter ineligible.
Regarding Claim 18:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 18 recites the same abstract ideas as claim 1. Therefore, claim 18 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 18 further recites additional elements of
training the recommendation neural network model comprises keeping the neural network encoder unchanged during the backpropagation. (This element does not integrate the abstract idea into a practical application because it amounts to mere “apply it on a computer” (see MPEP 2106.05(f)).)
Therefore, claim 18 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
The additional elements of claim 18 do not provide significantly more than the abstract idea itself, taken alone and in combination because
training the recommendation neural network model comprises keeping the neural network encoder unchanged during the backpropagation uses a computer as a tool to perform the abstract idea and cannot provide significantly more (see MPEP 2106.05(f)).
Therefore, claim 18 is subject-matter ineligible.
Regarding Claim 19:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 19 recites
computing the plurality of gradients corresponding to the plurality of losses; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper, by calculating the plurality of gradient based on the losses, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
and modifying a magnitude of at least one gradient in the plurality of task gradients based on a current training epoch number. (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper, by changing the magnitude of one gradient based on the current training iteration, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
Therefore, claim 19 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 19 does not further recite any additional elements. Therefore, claim 19 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
Since there are no additional elements, claim 19 does not provide significantly more than the abstract idea itself, taken alone and in combination. Therefore, claim 19 is subject-matter ineligible.
Regarding Claim 20:
Subject Matter Eligibility Analysis Step 2A Prong 1:
Claim 20 recites
generating a plurality of gradient weights from magnitudes of the plurality of gradients, including the at least one gradient with the modified magnitude; (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper, by calculating the plurality of gradient weight based on the plurality of gradient magnitudes, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a))).
and computing the weighted sum of gradients of the plurality of losses based on the generated plurality of gradient weights. (i.e., the broadest reasonable interpretation includes a step of observation, evaluation, and judgement and could be performed mentally or with pen and paper, by calculating the weighted sum of gradients of the plurality of losses, which is either a mental process of observation/evaluation/judgement (MPEP 2106.04(a)))
Therefore, claim 20 recites an abstract idea.
Subject Matter Eligibility Analysis Step 2A Prong 2:
Claim 20 does not further recite any additional elements. Therefore, claim 20 is not integrated into a practical application.
Subject Matter Eligibility Analysis Step 2B:
Since there are no additional elements, claim 20 does not provide significantly more than the abstract idea itself, taken alone and in combination. Therefore, claim 20 is subject-matter ineligible.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1, 3, 4, 7, 8, 10, 11, 14, 15, 17 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Li et al. (USPGPUB No. US 2023/0237093) (hereafter referred to as Li) in view of Gao et al (“A Survey of Graph Neural Networks for Recommender Systems: Challenges, Methods, and Directions”) (hereafter referred to as Gao), in further view of Hildebrandt et al (USPGPUB No. 2025/0165753) (hereafter referred to as Hildebrandt).
Regarding claim 1, Li teaches
A method for training a recommendation neural network model using multiple data sources, the method comprising: (Li, claim 11, “A method for training a neural network, comprising: receiving training data including relationships between a plurality of users and a plurality of content items [A method for training a recommendation neural network model using multiple data sources, the method comprising,].”).
receiving, via a data interface, … indicating a user-item interaction history; (Li, paragraph 7, “receiving training data including relationships between a plurality of users and a plurality of content items [receiving, via a data interface, … indicating a user-item interaction history;].”).
transforming the … into a user-item graph; (Li, paragraph 7, “One or more embodiments of the method, apparatus, and non-transitory computer readable medium include receiving training data including relationships between a plurality of users and a plurality of content items; generating a knowledge graph based on the training data, wherein the knowledge graph represents the relationships between the plurality of users and the plurality of content items; [transforming the … into a user-item graph; ]”)
encoding, by a neural network encoder, the user-item graph into user embeddings and item embeddings; (Li, paragraph 23, “In some embodiments, a multi-modal graph encoder of the item recommendation apparatus generates a first feature embedding representing a user and a second feature embedding representing a content item based on the knowledge graph [encoding, by a neural network encoder, the user-item graph into user embeddings and item embeddings;]”)
generating a plurality of losses according to a plurality of training tasks performed based on the user embeddings and, item embeddings; (Li, paragraph 53, “Some examples of the apparatus and method further include a training component configured to compute a loss function based on the first feature embedding and the second feature embedding and to update parameters of the multi-modal graph encoder based on the loss function [generating a plurality of losses according to a plurality of training tasks performed based on the user embeddings and, item embeddings;].”)
training the recommendation neural network model by updating the user embeddings and the item embeddings via, … based on a weighted sum of gradients of the plurality of losses; (Li, pg. 21 col. 63, “In some examples, training component 520 computes a loss function based on the first feature embedding and the second feature embedding. Training component 520 updates parameters of the multi-modal graph encoder 540 based on the loss function. In some examples, training component 520 identifies a first content item and a second content item. Training component 520 determines that a user prefers the first content item over the second content item using similarity scores for the first content item and the second content item. Training component 520 computes a ranking loss based on the determination, where the loss function includes the ranking loss [training the recommendation neural network model by updating the user embeddings and the item embeddings via, …based on a weighted sum of gradients of the plurality of losses;].”)
and generating, by a neural network decoder, one or more recommended items for a given user based on the updated user embeddings and the updated item embeddings. (Li, paragraph 8, “and a recommendation component configured to compare the first feature embedding to the second feature embedding to obtain similarity scores between the users and the content items and to identify recommended content items to the users based on the similarity scores. [and generating, by a neural network decoder, one or more recommended items for a given user based on the updated user embeddings and the updated item embeddings].”)
Li does not explicitly disclose receiving and transforming time series data, but Gao does disclose
…time series data… (Gao, paragraph 2.1.3, line 18, “…uses the user’s historical behavior sequence to learn timestamp-aware sequential patterns to recommend the next item that the user may be interested in; […time series data…].”).
…time series data… (Gao, paragraph 2.1.3, line 18, “…uses the user’s historical behavior sequence to learn timestamp-aware sequential patterns to recommend the next item that the user may be interested in; […time series data…]”)
Li and Gao are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li and Gao to teach the above limitation(s). The motivation for doing so is advantageous because a recommendation model that “… uses the user’s historical behavior sequence to learn timestamp-aware sequential patterns to recommend the next item that the user may be interested in.” (Gao, paragraph 2.1.3, line 18).
Li and Gao do not explicitly disclose backpropagation, but Hildebrandt does disclose
…backpropagation… (Hildebrandt, paragraph 84, The result thereof, at reference sign (3) are final encodings h.sub.u.sup.L h.sub.i.sup.(L) which enter the decoder network DN at reference sign (4). This means that the trainable parameters are tuned by gradient descent of the initial item and user encodings via backpropagating the gradient induced by the loss function (see below) through the layers of the GNN without changing the weight matrices W.sup.(l). In other words, the initial encodings are adapted, so that the result of the prediction is better. Paragraph 85, “For a backpropagation it is determined which element, e.g., initial encoding has what influence on the value of the loss function. In contrast, for a usual forward pass a prediction would be calculated, a loss function would be determined, and the weights would be adapted to minimize the loss function, by using its gradient, and thus optimize the prediction. […backpropagation…].”)
Li, Gao, and Hildebrandt are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, and Hildebrandt to teach the above limitation(s). The motivation for doing so is advantageous because a recommendation model that has “… trainable parameters are tuned by gradient descent of the initial item and user encodings via backpropagating the gradient induced by the loss function (see below) through the layers of the GNN without changing the weight matrices W.sup.(l).” (Hildebrandt, paragraph 84).
Regarding claim 3, Li in view of Gao, in view of Hildebrandt teaches the method of claim 1. Li and Hildebrandt do not explicitly disclose, but Gao does disclose
wherein the plurality of training tasks comprise any one or more of: a task of predicting an item for a user; (Gao, paragraph 2.1.3, line 18, “The sequential recommendation method extracts information from these behavioral sequences and predicts the user’s next interaction item, as shown in Figure 4 [wherein the plurality of training tasks comprise any one or more of: a task of predicting an item for a user;].”)
a task of predicting a social connection for a user; (Gao, paragraph 3:17, line 14, “In some trivial tasks, the recommender system’s data input can be considered a kind of relational data, such as user-item interactions or user-user social relations. Paragraph 2.1.3, line 13, “Hence, social relations are often integrated into recommender systems to enhance the final performance, which is called social recommendation. Figure 3 illustrates the data input of social recommendation, of which user interactions are determined by both preferences and social factors (social influence and social homophily). [a task of predicting a social connection for a user;].”)
and a task of predicting an item category for a user. (Gao, paragraph 2.1.2, line 15, “After the matching stage, multiple sources of candidate items from different channels are merged into one list and then scored by a single ranking model. Specifically, the ranking model ranks these items according to the scores, and the top dozens of items are selected. [and a task of predicting an item category for a user.;].”)
Li, Hildebrandt, and Gao are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Hildebrandt, and Gao to teach the above limitation(s). The motivation for doing so is advantageous because “The sequential recommendation method extracts information from these behavioral sequences and predicts the user’s next interaction item, as shown in Figure 4” (Gao, paragraph 2.1.3, line 18,) and “Hence, social relations are often integrated into recommender systems to enhance the final performance, which is called social recommendation. Figure 3 illustrates the data input of social recommendation, of which user interactions are determined by both preferences and social factors (social influence and social homophily).” (Gao, paragraph 2.1.3, line 13).
Regarding claim 4, Li in view of Gao, in view of Hildebrandt teaches the method of claim 1. Li and Gao do not explicitly disclose, but Hildebrandt does disclose
wherein the training the recommendation neural network model comprises keeping the neural network encoder unchanged during the backpropagation. (Hildebrandt, paragraph 28, “The initial encodings are used as input to the encoder network. The weights of the encoder network are randomly initialized. According to an embodiment they are at least partly frozen or/and set to zero. This further reduces computational complexity. Claim 4, “The method according to claim 1, wherein at least a part of the weights of the encoder network is kept constant after the initialization. Paragraph 83, “According to an embodiment, the weight matrix W.sup.(1), where I denotes an individual layer from 1 to L is randomly initialized according to some distribution (e.g., using Xavier initializations). The weight matrix is not trained further, i.e., no adaptations to W.sub.i.sup.(1), where i=0,1 and 1=1, L are made due to the training, the weights W.sup.(1) are kept constant or are “frozen”, as indicated by the asterisks. Paragraph 84, “The result thereof, at reference sign (3) are final encodings h.sub.u.sup.L h.sub.i.sup.(L) which enter the decoder network DN at reference sign (4). This means that the trainable parameters are tuned by gradient descent of the initial item and user encodings via backpropagating the gradient induced by the loss function (see below) through the layers of the GNN without changing the weight matrices W.sup.(l). In other words, the initial encodings are adapted, so that the result of the prediction is better. Paragraph 85, “For a backpropagation it is determined which element, e.g., initial encoding has what influence on the value of the loss function. In contrast, for a usual forward pass a prediction would be calculated, a loss function would be determined, and the weights would be adapted to minimize the loss function, by using its gradient, and thus optimize the prediction. In other words, the initial encodings are adapted, so that the result of the prediction is better. [wherein the training the recommendation neural network model comprises keeping the neural network encoder unchanged during the backpropagation;].”)
Li, Gao, and Hildebrandt are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, and Hildebrandt to teach the above limitation(s). The motivation for doing so is advantageous because “For a backpropagation it is determined which element, e.g., initial encoding has what influence on the value of the loss function. In contrast, for a usual forward pass a prediction would be calculated, a loss function would be determined, and the weights would be adapted to minimize the loss function, by using its gradient, and thus optimize the prediction. In other words, the initial encodings are adapted, so that the result of the prediction is better.” (Hildebrandt, paragraph 85).
Regarding claim 7, Li in view of Gao, in view of Hildebrandt teaches the method of claim 1. Li and Gao do not explicitly disclose, but Hildebrandt does disclose
wherein one or more losses from the plurality of losses are computed further based on information including any one or more of: user profile information; (Hildebrandt paragraph 145, “For the prediction extracted by the decoder network a loss function is calculated which depends on the training objective. According to an embodiment, as training objective a “personalized ranking may be employed, i.e., a ranking of recommended items for a specific user [wherein one or more losses from the plurality of losses are computed further based on information including any one or more of: user profile information;].”)
and user social connection information (Hildebrandt, paragraph 32, “The loss function is defined such that higher scores are achieved if there is an observed interaction between a user and an item than if there is none. [a task of predicting a social connection for a user;].”)
Li, Gao, and Hildebrandt are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, and Hildebrandt to teach the above limitation(s). The motivation for doing so is advantageous because “The loss function is defined such that higher scores are achieved if there is an observed interaction between a user and an item than if there is none.” (Hildebrandt, paragraph 32).
Regarding claim 8, Li teaches
A system for making recommendations, the system comprising: (Li, abstract, “Systems and methods for item recommendation are described. [A method for training a recommendation neural network model using multiple data sources, the method comprising,].”).
a memory that stores a recommendation neural network model and a plurality of processor executable instructions; (Li, paragraph 36, “Item recommendation apparatus 110 may also include a processor unit, a memory unit, an I/O module, a training component, and a recommendation component. The training component is used to train a machine learning model (or an item recommendation network [a memory that stores a recommendation neural network model and a plurality of processor executable instructions;].”) (Li, paragraph 56, “In some cases, processor unit 505 is configured to execute computer-readable instructions stored in a memory to perform various functions [a memory that stores a recommendation neural network model and a plurality of processor executable instructions
a communication interface that receives … indicating a user-item interaction history; (Li, paragraph 7, “receiving training data including relationships between a plurality of users and a plurality of content items [a communication interface that receives … indicating a user-item interaction history;].”).
one or more hardware processors that read and execute the plurality of processor-executable instructions from the memory to perform operations comprising: (Li, paragraph 56, “A processor unit 505 is an intelligent hardware device, (e.g., a general-purpose processing component, a digital signal processor (DSP), a central processing unit (CPU), a graphics processing unit (GPU), a microcontroller, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a programmable logic device, a discrete gate or transistor logic component, a discrete hardware component, or any combination thereof). In some cases, processor unit 505 is configured to operate a memory array using a memory controller. In other cases, a memory controller is integrated into the processor. In some cases, processor unit 505 is configured to execute computer-readable instructions stored in a memory to perform various functions. In some embodiments, processor unit 505 includes special-purpose components for modem processing, baseband processing, digital signal processing, or transmission processing. [one or more hardware processors that read and execute the plurality of processor-executable instructions from the memory to perform operations comprising:].”)
transforming the … into a user-item graph; (Li, paragraph 7, “One or more embodiments of the method, apparatus, and non-transitory computer readable medium include receiving training data including relationships between a plurality of users and a plurality of content items; generating a knowledge graph based on the training data, wherein the knowledge graph represents the relationships between the plurality of users and the plurality of content items; [transforming the … into a user-item graph; ]”)
encoding, by a neural network encoder, the user-item graph into user embeddings and item embeddings; (Li, paragraph 23, “In some embodiments, a multi-modal graph encoder of the item recommendation apparatus generates a first feature embedding representing a user and a second feature embedding representing a content item based on the knowledge graph [encoding, by a neural network encoder, the user-item graph into user embeddings and item embeddings;]”)
generating a plurality of losses according to a plurality of training tasks performed based on the user embeddings and, item embeddings; (Li, paragraph 53, “Some examples of the apparatus and method further include a training component configured to compute a loss function based on the first feature embedding and the second feature embedding and to update parameters of the multi-modal graph encoder based on the loss function [generating a plurality of losses according to a plurality of training tasks performed based on the user embeddings and, item embeddings;].”)
training the recommendation neural network model by updating the user embeddings and the item embeddings via, … based on a weighted sum of gradients of the plurality of losses; (Li, pg. 21 col. 63, “In some examples, training component 520 computes a loss function based on the first feature embedding and the second feature embedding. Training component 520 updates parameters of the multi-modal graph encoder 540 based on the loss function. In some examples, training component 520 identifies a first content item and a second content item. Training component 520 determines that a user prefers the first content item over the second content item using similarity scores for the first content item and the second content item. Training component 520 computes a ranking loss based on the determination, where the loss function includes the ranking loss [training the recommendation neural network model by updating the user embeddings and the item embeddings via, …based on a weighted sum of gradients of the plurality of losses;].”)
and generating, by a neural network decoder, one or more recommended items for a given user based on the updated user embeddings and the updated item embeddings. (Li, paragraph 8, “and a recommendation component configured to compare the first feature embedding to the second feature embedding to obtain similarity scores between the users and the content items and to identify recommended content items to the users based on the similarity scores. [and generating, by a neural network decoder, one or more recommended items for a given user based on the updated user embeddings and the updated item embeddings].”)
Li does not explicitly disclose receiving and transforming time series data, but Gao does disclose
…time series data… (Gao, paragraph 2.1.3, line 18, “…uses the user’s historical behavior sequence to learn timestamp-aware sequential patterns to recommend the next item that the user may be interested in; […time series data…].”).
…time series data… (Gao, paragraph 2.1.3, line 18, “…uses the user’s historical behavior sequence to learn timestamp-aware sequential patterns to recommend the next item that the user may be interested in; […time series data…]”)
Li and Gao are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li and Gao to teach the above limitation(s). The motivation for doing so is advantageous because a recommendation model that “… uses the user’s historical behavior sequence to learn timestamp-aware sequential patterns to recommend the next item that the user may be interested in.” (Gao, paragraph 2.1.3, line 18).
Li and Gao do not explicitly disclose backpropagation, but Hildebrandt does disclose
backpropagation in “training the recommendation neural network model by updating the user embeddings and the item embeddings via backpropagation based on a weighted sum of gradients of the plurality of losses;” (Hildebrandt, paragraph 84,The result thereof, at reference sign (3) are final encodings h.sub.u.sup.L h.sub.i.sup.(L) which enter the decoder network DN at reference sign (4). This means that the trainable parameters are tuned by gradient descent of the initial item and user encodings via backpropagating the gradient induced by the loss function (see below) through the layers of the GNN without changing the weight matrices W.sup.(l). In other words, the initial encodings are adapted, so that the result of the prediction is better. Paragraph 85, “For a backpropagation it is determined which element, e.g., initial encoding has what influence on the value of the loss function. In contrast, for a usual forward pass a prediction would be calculated, a loss function would be determined, and the weights would be adapted to minimize the loss function, by using its gradient, and thus optimize the prediction.[training the recommendation neural network model by updating the user embeddings and the item embeddings via backpropagation based on a weighted sum of gradients of the plurality of losses;].”)
Li, Gao, and Hildebrandt are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, and Hildebrandt to teach the above limitation(s). The motivation for doing so is advantageous because a recommendation model that has “… trainable parameters are tuned by gradient descent of the initial item and user encodings via backpropagating the gradient induced by the loss function (see below) through the layers of the GNN without changing the weight matrices W.sup.(l).” (Hildebrandt, paragraph 84).
Regarding claim 10, Li in view of Gao, in view of Hildebrandt teaches the method of claim 8. Li and Hildebrandt do not explicitly disclose, but Gao does disclose
wherein the plurality of training tasks comprise any one or more of: a task of predicting an item for a user; (Gao, paragraph 2.1.3, line 18, “The sequential recommendation method extracts information from these behavioral sequences and predicts the user’s next interaction item, as shown in Figure 4 [a task of predicting an item for a user;].”)
a task of predicting a social connection for a user;( Gao, paragraph 3:17, line 14, “In some trivial tasks, the recommender system’s data input can be considered a kind of relational data, such as user-item interactions or user-user social relations. Paragraph 2.1.3, line 13, “Hence, social relations are often integrated into recommender systems to enhance the final performance, which is called social recommendation. Figure 3 illustrates the data input of social recommendation, of which user interactions are determined by both preferences and social factors (social influence and social homophily). [a task of predicting a social connection for a user;].”)
and a task of predicting an item category for a user. (Gao, paragraph 2.1.2, line 15, “After the matching stage, multiple sources of candidate items from different channels are merged into one list and then scored by a single ranking model. Specifically, the ranking model ranks these items according to the scores, and the top dozens of items are selected. [and a task of predicting an item category for a user.;].”)
Li, Hildebrandt, and Gao are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Hildebrandt, and Gao to teach the above limitation(s). The motivation for doing so is advantageous because “The sequential recommendation method extracts information from these behavioral sequences and predicts the user’s next interaction item, as shown in Figure 4” (Gao, paragraph 2.1.3, line 18,) and “Hence, social relations are often integrated into recommender systems to enhance the final performance, which is called social recommendation. Figure 3 illustrates the data input of social recommendation, of which user interactions are determined by both preferences and social factors (social influence and social homophily).” (Gao, paragraph 2.1.3, line 13).
Regarding claim 11, Li in view of Gao, in view of Hildebrandt teaches the method of claim 8. Li and Gao do not explicitly disclose, but Hildebrandt does disclose
the operation of training the recommendation neural network model comprising keeping the neural network encoder unchanged during the backpropagation. (Hildebrandt, paragraph 28, “The initial encodings are used as input to the encoder network. The weights of the encoder network are randomly initialized. According to an embodiment they are at least partly frozen or/and set to zero. This further reduces computational complexity. Claim 4, “The method according to claim 1, wherein at least a part of the weights of the encoder network is kept constant after the initialization. Paragraph 83, “According to an embodiment, the weight matrix W.sup.(1), where I denotes an individual layer from 1 to L is randomly initialized according to some distribution (e.g., using Xavier initializations). The weight matrix is not trained further, i.e., no adaptations to W.sub.i.sup.(1), where i=0,1 and 1=1, L are made due to the training, the weights W.sup.(1) are kept constant or are “frozen”, as indicated by the asterisks. Paragraph 84, “The result thereof, at reference sign (3) are final encodings h.sub.u.sup.L h.sub.i.sup.(L) which enter the decoder network DN at reference sign (4). This means that the trainable parameters are tuned by gradient descent of the initial item and user encodings via backpropagating the gradient induced by the loss function (see below) through the layers of the GNN without changing the weight matrices W.sup.(l). In other words, the initial encodings are adapted, so that the result of the prediction is better. Paragraph 85, “For a backpropagation it is determined which element, e.g., initial encoding has what influence on the value of the loss function. In contrast, for a usual forward pass a prediction would be calculated, a loss function would be determined, and the weights would be adapted to minimize the loss function, by using its gradient, and thus optimize the prediction. In other words, the initial encodings are adapted, so that the result of the prediction is better. [wherein the training the recommendation neural network model comprises keeping the neural network encoder unchanged during the backpropagation;].”)
Li, Gao, and Hildebrandt are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, and Hildebrandt to teach the above limitation(s). The motivation for doing so is advantageous because “For a backpropagation it is determined which element, e.g., initial encoding has what influence on the value of the loss function. In contrast, for a usual forward pass a prediction would be calculated, a loss function would be determined, and the weights would be adapted to minimize the loss function, by using its gradient, and thus optimize the prediction. In other words, the initial encodings are adapted, so that the result of the prediction is better.” (Hildebrandt, paragraph 85).
Regarding claim 14, Li in view of Gao, in view of Hildebrandt teaches the method of claim 8. Li and Gao do not explicitly disclose, but Hildebrandt does disclose
wherein one or more losses from the plurality of losses are computed further based on information including any one or more of: user profile information; (Hildebrandt paragraph 145, “For the prediction extracted by the decoder network a loss function is calculated which depends on the training objective. According to an embodiment, as training objective a “personalized ranking may be employed, i.e., a ranking of recommended items for a specific user [wherein one or more losses from the plurality of losses are computed further based on information including any one or more of: user profile information;].”)
and user social connection information (Hildebrandt, paragraph 32, “The loss function is defined such that higher scores are achieved if there is an observed interaction between a user and an item than if there is none. [a task of predicting a social connection for a user;].”)
Li, Gao, and Hildebrandt are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, and Hildebrandt to teach the above limitation(s). The motivation for doing so is advantageous because “The loss function is defined such that higher scores are achieved if there is an observed interaction between a user and an item than if there is none.” (Hildebrandt, paragraph 32).
Regarding claim 15, Li teaches
A non-transitory machine-readable medium comprising a plurality of machine-executable instructions which, when executed by one or more processors, are adapted to cause the one or more processor to perform operations comprising: (Li, paragraph 6, “A method, apparatus, and non-transitory computer readable medium for item recommendation are described. One or more embodiments of the method, apparatus, and non-transitory computer readable medium include receiving input indicating a relationship between a user and a first content item; generating a knowledge graph based on the input. Paragraph 76-77, “Thus, the functions described herein may be implemented in hardware or software and may be executed by a processor, firmware, or any combination thereof. If implemented in software executed by a processor, the functions may be stored in the form of instructions or code on a computer-readable medium…Computer-readable media includes both non-transitory computer storage media and communication media including any medium that facilitates transfer of code or data. A non-transitory storage medium may be any available medium that can be accessed by a computer, [A non-transitory machine-readable medium comprising a plurality of machine-executable instructions which, when executed by one or more processors, are adapted to cause the one or more processor to perform operations comprising:].”).
transforming the … into a user-item graph; (Li, paragraph 7, “One or more embodiments of the method, apparatus, and non-transitory computer readable medium include receiving training data including relationships between a plurality of users and a plurality of content items; generating a knowledge graph based on the training data, wherein the knowledge graph represents the relationships between the plurality of users and the plurality of content items; [transforming the … into a user-item graph]”)
encoding, by a neural network encoder, the user-item graph into user embeddings and item embeddings; (Li, paragraph 23, “In some embodiments, a multi-modal graph encoder of the item recommendation apparatus generates a first feature embedding representing a user and a second feature embedding representing a content item based on the knowledge graph [encoding, by a neural network encoder, the user-item graph into user embeddings and item embeddings;]”)
generating a plurality of losses according to a plurality of training tasks performed based on the user embeddings and, item embeddings; (Li, paragraph 53, “Some examples of the apparatus and method further include a training component configured to compute a loss function based on the first feature embedding and the second feature embedding and to update parameters of the multi-modal graph encoder based on the loss function [generating a plurality of losses according to a plurality of training tasks performed based on the user embeddings and, item embeddings;].”)
training the recommendation neural network model by updating the user embeddings and the item embeddings via, … based on a weighted sum of gradients of the plurality of losses; (Li, pg. 21 col. 63, “In some examples, training component 520 computes a loss function based on the first feature embedding and the second feature embedding. Training component 520 updates parameters of the multi-modal graph encoder 540 based on the loss function. In some examples, training component 520 identifies a first content item and a second content item. Training component 520 determines that a user prefers the first content item over the second content item using similarity scores for the first content item and the second content item. Training component 520 computes a ranking loss based on the determination, where the loss function includes the ranking loss [training the recommendation neural network model by updating the user embeddings and the item embeddings via, …based on a weighted sum of gradients of the plurality of losses;].”)
and generating, by a neural network decoder, one or more recommended items for a given user based on the updated user embeddings and the updated item embeddings. (Li, paragraph 8, “and a recommendation component configured to compare the first feature embedding to the second feature embedding to obtain similarity scores between the users and the content items and to identify recommended content items to the users based on the similarity scores. [and generating, by a neural network decoder, one or more recommended items for a given user based on the updated user embeddings and the updated item embeddings].”)
Li does not explicitly disclose receiving and transforming time series data, but Gao does disclose
…time series data… (Gao, paragraph 2.1.3, line 18, “…uses the user’s historical behavior sequence to learn timestamp-aware sequential patterns to recommend the next item that the user may be interested in; […time series data…].”)
…time series data… (Gao, paragraph 2.1.3, line 18, “…uses the user’s historical behavior sequence to learn timestamp-aware sequential patterns to recommend the next item that the user may be interested in; […time series data…].”)
Li and Gao are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li and Gao to teach the above limitation(s). The motivation for doing so is advantageous because a recommendation model that “… uses the user’s historical behavior sequence to learn timestamp-aware sequential patterns to recommend the next item that the user may be interested in.” (Gao, paragraph 2.1.3, line 18).
Li and Gao do not explicitly disclose backpropagation, but Hildebrandt does disclose
backpropagation in “training the recommendation neural network model by updating the user embeddings and the item embeddings via backpropagation based on a weighted sum of gradients of the plurality of losses;” (Hildebrandt, paragraph 84,The result thereof, at reference sign (3) are final encodings h.sub.u.sup.L h.sub.i.sup.(L) which enter the decoder network DN at reference sign (4). This means that the trainable parameters are tuned by gradient descent of the initial item and user encodings via backpropagating the gradient induced by the loss function (see below) through the layers of the GNN without changing the weight matrices W.sup.(l). In other words, the initial encodings are adapted, so that the result of the prediction is better. Paragraph 85, “For a backpropagation it is determined which element, e.g., initial encoding has what influence on the value of the loss function. In contrast, for a usual forward pass a prediction would be calculated, a loss function would be determined, and the weights would be adapted to minimize the loss function, by using its gradient, and thus optimize the prediction.[training the recommendation neural network model by updating the user embeddings and the item embeddings via backpropagation based on a weighted sum of gradients of the plurality of losses;].”)
Li, Gao, and Hildebrandt are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, and Hildebrandt to teach the above limitation(s). The motivation for doing so is advantageous because a recommendation model that has “… trainable parameters are tuned by gradient descent of the initial item and user encodings via backpropagating the gradient induced by the loss function (see below) through the layers of the GNN without changing the weight matrices W.sup.(l). ” (Hildebrandt, paragraph 84).
Regarding claim 17, Li in view of Gao, in view of Hildebrandt teaches the method of claim 15. Li and Hildebrandt do not explicitly disclose, but Gao does disclose
wherein the plurality of training tasks comprise any one or more of: a task of predicting an item for a user; (Gao, paragraph 2.1.3, line 18, “The sequential recommendation method extracts information from these behavioral sequences and predicts the user’s next interaction item, as shown in Figure 4 [a task of predicting an item for a user;].”)
a task of predicting a social connection for a user; (Gao, paragraph 3:17, line 14, “In some trivial tasks, the recommender system’s data input can be considered a kind of relational data, such as user-item interactions or user-user social relations. [a task of predicting a social connection for a user;].”) (Gao, paragraph 2.1.3, line 13, “Hence, social relations are often integrated into recommender systems to enhance the final performance, which is called social recommendation. Figure 3 illustrates the data input of social recommendation, of which user interactions are determined by both preferences and social factors (social influence and social homophily). [a task of predicting a social connection for a user;].”)
and a task of predicting an item category for a user. (Gao, paragraph 2.1.2, line 15, “After the matching stage, multiple sources of candidate items from different channels are merged into one list and then scored by a single ranking model. Specifically, the ranking model ranks these items according to the scores, and the top dozens of items are selected. [and a task of predicting an item category for a user.;].”)
Li, Gao, and Hildebrandt are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, and Hildebrandt to teach the above limitation(s). The motivation for doing so is advantageous because “The sequential recommendation method extracts information from these behavioral sequences and predicts the user’s next interaction item, as shown in Figure 4” (Gao, paragraph 2.1.3, line 18,) and “Hence, social relations are often integrated into recommender systems to enhance the final performance, which is called social recommendation. Figure 3 illustrates the data input of social recommendation, of which user interactions are determined by both preferences and social factors (social influence and social homophily).” (Gao, paragraph 2.1.3, line 13).
Regarding claim 18, Li in view of Gao, in view of Hildebrandt teaches the method of claim 15. Li and Gao do not explicitly disclose, but Hildebrandt does disclose
wherein the training the recommendation neural network model comprises keeping the neural network encoder unchanged during the backpropagation. (Hildebrandt, paragraph 28, “The initial encodings are used as input to the encoder network. The weights of the encoder network are randomly initialized. According to an embodiment they are at least partly frozen or/and set to zero. This further reduces computational complexity. Claim 4, “The method according to claim 1, wherein at least a part of the weights of the encoder network is kept constant after the initialization. Paragraph 83, “According to an embodiment, the weight matrix W.sup.(1), where I denotes an individual layer from 1 to L is randomly initialized according to some distribution (e.g., using Xavier initializations). The weight matrix is not trained further, i.e., no adaptations to W.sub.i.sup.(1), where i=0,1 and 1=1, L are made due to the training, the weights W.sup.(1) are kept constant or are “frozen”, as indicated by the asterisks. Paragraph 84, “The result thereof, at reference sign (3) are final encodings h.sub.u.sup.L h.sub.i.sup.(L) which enter the decoder network DN at reference sign (4). This means that the trainable parameters are tuned by gradient descent of the initial item and user encodings via backpropagating the gradient induced by the loss function (see below) through the layers of the GNN without changing the weight matrices W.sup.(l). In other words, the initial encodings are adapted, so that the result of the prediction is better. Paragraph 85, “For a backpropagation it is determined which element, e.g., initial encoding has what influence on the value of the loss function. In contrast, for a usual forward pass a prediction would be calculated, a loss function would be determined, and the weights would be adapted to minimize the loss function, by using its gradient, and thus optimize the prediction. In other words, the initial encodings are adapted, so that the result of the prediction is better. [wherein the training the recommendation neural network model comprises keeping the neural network encoder unchanged during the backpropagation;].”)
Li, Gao, and Hildebrandt are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, and Hildebrandt to teach the above limitation(s). The motivation for doing so is advantageous because “For a backpropagation it is determined which element, e.g., initial encoding has what influence on the value of the loss function. In contrast, for a usual forward pass a prediction would be calculated, a loss function would be determined, and the weights would be adapted to minimize the loss function, by using its gradient, and thus optimize the prediction. In other words, the initial encodings are adapted, so that the result of the prediction is better.” (Hildebrandt, paragraph 85).
Claims 2, 9, and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Li in view of Gao, in view of Hildebrandt, and in further view of Janakiraman et al. (USPGPUB No. US 2021/0287273) (hereafter referred to as Janakiraman).
Regarding claim 2, Li in view of Gao, in view of Hildebrandt teaches the method of claim 1. Li , Gao, and Hildebrandt do not explicitly disclose, but Janakiraman does disclose
wherein the transforming the time series data into a user-item graph comprises: creating a first node representing at least one user and a second node representing at least one item; (Janakiraman paragraph 4, “The repository further stores a bipartite graph including user nodes corresponding to the users, item nodes corresponding to the items, and edges each connecting a user node and an item node [creating a first node representing at least one user and a second node representing at least one item;].”)
adding an edge between the first node and the second node representing that the at least one user has interacted with the at least one item; (Janakiraman paragraph 14, “The recommendation may be based on scores calculated for paths through a bipartite graph that represents interactions between users and items. A bipartite graph is a graph whose nodes may be divided into two disjoint sets (e.g., user nodes and item nodes) such that each edge connects a node in one set to a node in the other set [adding an edge between the first node and the second node representing that the at least one user has interacted with the at least one item;].”)
and associating the edge with a score representing a preference level of the at least one user towards the at least one item. (Janakiraman paragraph 3, “The method further includes calculating, using the edge weights calculated for the edges, scores for the paths, and recommending, to the target user and using the scores for the paths, a recommended item. [and associating the edge with a score representing a preference level of the at least one user towards the at least one item;].”)
Li, Gao, Hildebrandt, and Janakiraman are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, Hildebrandt, and Janakiraman to teach the above limitation(s). The motivation for doing so is advantageous because “The recommendation may be based on scores calculated for paths through a bipartite graph that represents interactions between users and items. A bipartite graph is a graph whose nodes may be divided into two disjoint sets (e.g., user nodes and item nodes) such that each edge connects a node in one set to a node in the other set.” (Janakiraman paragraph 14).
Regarding claim 9, Li in view of Gao, in view of Hildebrandt teaches the method of claim 8. Li, Gao, and Hildebrandt do not explicitly disclose, but Janakiraman does disclose
the operation of transforming the time series data into a user-item graph further comprising: creating a first node representing at least one user and a second node representing at least one item; (Janakiraman paragraph 4, “The repository further stores a bipartite graph including user nodes corresponding to the users, item nodes corresponding to the items, and edges each connecting a user node and an item node [creating a first node representing at least one user and a second node representing at least one item;].”)
adding an edge between the first node and the second node representing that the at least one user has interacted with the at least one item; (Janakiraman paragraph 14, “The recommendation may be based on scores calculated for paths through a bipartite graph that represents interactions between users and items. A bipartite graph is a graph whose nodes may be divided into two disjoint sets (e.g., user nodes and item nodes) such that each edge connects a node in one set to a node in the other set [adding an edge between the first node and the second node representing that the at least one user has interacted with the at least one item;].”)
and associating the edge with a score representing a preference level of the at least one user towards the at least one item. (Janakiraman paragraph 3, “The method further includes calculating, using the edge weights calculated for the edges, scores for the paths, and recommending, to the target user and using the scores for the paths, a recommended item. [and associating the edge with a score representing a preference level of the at least one user towards the at least one item;].”)
Li, Gao, Hildebrandt, and Janakiraman are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, Hildebrandt, and Janakiraman to teach the above limitation(s). The motivation for doing so is advantageous because “The recommendation may be based on scores calculated for paths through a bipartite graph that represents interactions between users and items. A bipartite graph is a graph whose nodes may be divided into two disjoint sets (e.g., user nodes and item nodes) such that each edge connects a node in one set to a node in the other set.” (Janakiraman paragraph 14).
Regarding claim 16, Li in view of Gao, in view of Hildebrandt teaches the method of claim 15. Li, Gao, and Hildebrandt do not explicitly disclose, but Janakiraman does disclose
wherein the transforming the time series data into a user-item graph comprises: creating a first node representing at least one user and a second node representing at least one item; (Janakiraman paragraph 4, “The repository further stores a bipartite graph including user nodes corresponding to the users, item nodes corresponding to the items, and edges each connecting a user node and an item node [creating a first node representing at least one user and a second node representing at least one item;].”)
adding an edge between the first node and the second node representing that the at least one user has interacted with the at least one item; (Janakiraman paragraph 14, “The recommendation may be based on scores calculated for paths through a bipartite graph that represents interactions between users and items. A bipartite graph is a graph whose nodes may be divided into two disjoint sets (e.g., user nodes and item nodes) such that each edge connects a node in one set to a node in the other set [adding an edge between the first node and the second node representing that the at least one user has interacted with the at least one item;].”)
and associating the edge with a score representing a preference level of the at least one user towards the at least one item. (Janakiraman paragraph 3, “The method further includes calculating, using the edge weights calculated for the edges, scores for the paths, and recommending, to the target user and using the scores for the paths, a recommended item. [and associating the edge with a score representing a preference level of the at least one user towards the at least one item;].”)
Li, Gao, Hildebrandt, and Janakiraman are considered analogous to the claimed invention because they both are in the same field of endeavor (recommendation neural network model). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, Hildebrandt, and Janakiraman to teach the above limitation(s). The motivation for doing so is advantageous because “The recommendation may be based on scores calculated for paths through a bipartite graph that represents interactions between users and items. A bipartite graph is a graph whose nodes may be divided into two disjoint sets (e.g., user nodes and item nodes) such that each edge connects a node in one set to a node in the other set.” (Janakiraman paragraph 14).
Claims 5, 6, 12, 13, 19, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Li in view of Gao, in view of Hildebrandt, in further view of Chen et al (“GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networks”).
Regarding claim 5, Li in view of Gao, in view of Hildebrandt teaches the method of claim 1. Li, Gao, and Hildebrandt do not explicitly disclose, but Chen does disclose
the training the recommendation neural network model comprises: computing the plurality of gradients corresponding to the plurality of losses; (Chen, paragraph 1, line 33 “In practice, the multitask loss function is often assumed to be linear in the single task losses
PNG
media_image1.png
14
102
media_image1.png
Greyscale
, where the sum runs over all T tasks. In our case, we propose an adaptive method, and so
PNG
media_image2.png
14
16
media_image2.png
Greyscale
can vary at each training step
PNG
media_image3.png
16
92
media_image3.png
Greyscale
This linear form of the loss function is convenient for implementing gradient balancing, as
PNG
media_image2.png
14
16
media_image2.png
Greyscale
very directly and linearly couples to the backpropagated gradient magnitudes from each task. The challenge is then to find the best value for each
PNG
media_image2.png
14
16
media_image2.png
Greyscale
at each training step t that balances the contribution of each task for optimal model training. To optimize the weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
for gradient balancing, we propose a simple algorithm that penalizes the network when backpropagated gradients from any task are too large or too small. [operation of training the recommendation neural network comprising: computing the plurality of gradients corresponding to the plurality of losses; the training the recommendation neural network model comprises: computing the plurality of gradients corresponding to the plurality of losses;]”.)
and modifying a magnitude of at least one gradient in the plurality of task gradients based on a current training epoch number (Chen, paragraph 1, line 27 “task imbalances impede proper training because they manifest as imbalances between backpropagated gradients. A task that is too dominant during training, for example, will necessarily express that dominance by inducing gradients which have relatively large magnitudes. We aim to mitigate such issues at their root by directly modifying gradient magnitudes through tuning of the multitask loss function [and modifying a magnitude of at least one gradient in the plurality of task gradients based on a current training epoch number.].”)
Li, Gao, Hildebrandt, and Chen are considered analogous to the claimed invention because they both are in the same field of endeavor (loss functions). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, Hildebrandt, and Chen to teach the above limitation(s). The motivation for doing so is advantageous because “task imbalances impede proper training because they manifest as imbalances between backpropagated gradients. A task that is too dominant during training, for example, will necessarily express that dominance by inducing gradients which have relatively large magnitudes…to mitigate such issues at their root by directly modifying gradient magnitudes through tuning of the multitask loss function.” (Chen, paragraph 1, line 27).
Regarding claim 6, Li in view of Gao, in view of Hildebrandt teaches the method of claim 5. Li, Gao, and Hildebrandt do not explicitly disclose, but Chen does disclose
the training the recommendation neural network model comprises: generating a plurality of gradient weights from magnitudes of the plurality of gradients, including the at least one gradient with the modified magnitude; (Chen, paragraph 1, line 35, “In our case, we propose an adaptive method, and so
PNG
media_image2.png
14
16
media_image2.png
Greyscale
can vary at each training step
PNG
media_image3.png
16
92
media_image3.png
Greyscale
This linear form of the loss function is convenient for implementing gradient balancing, as
PNG
media_image2.png
14
16
media_image2.png
Greyscale
very directly and linearly couples to the backpropagated gradient magnitudes from each task. The challenge is then to find the best value for each
PNG
media_image2.png
14
16
media_image2.png
Greyscale
at each training step t that balances the contribution of each task for optimal model training. To optimize the weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
for gradient balancing, we propose a simple algorithm that penalizes the network when backpropagated gradients from any task are too large or too small. The correct balance is struck when tasks are training at similar rates; if task i is training relatively quickly, then its weight
PNG
media_image4.png
22
28
media_image4.png
Greyscale
should decrease relative to other task weights
PNG
media_image5.png
16
55
media_image5.png
Greyscale
to allow other tasks more influence on GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networks training [the training the recommendation neural network model comprises: generating a plurality of gradient weights from magnitudes of the plurality of gradients, including the at least one gradient with the modified magnitude;]”.)
and computing the weighted sum of gradients of the plurality of losses based on the generated plurality of gradient weights. (Chen, paragraph 3.2, line 23 “Equation 1 gives a target for each task i’s gradient norms, and we update our loss weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
to move gradient norms towards this target for each task. GradNorm is then implemented as an
PNG
media_image6.png
24
16
media_image6.png
Greyscale
loss function
PNG
media_image7.png
14
28
media_image7.png
Greyscale
between the actual and target gradient norms at each timestep for each task, summed over all tasks:
PNG
media_image8.png
50
320
media_image8.png
Greyscale
where the summation runs through all T tasks. When differentiating this loss
PNG
media_image7.png
14
28
media_image7.png
Greyscale
, we treat the target gradient norm
PNG
media_image9.png
20
100
media_image9.png
Greyscale
as a fixed constant to prevent loss weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
from spuriously drifting towards zero.
PNG
media_image7.png
14
28
media_image7.png
Greyscale
is then differentiated only with respect to the
PNG
media_image2.png
14
16
media_image2.png
Greyscale
, as the
PNG
media_image4.png
22
28
media_image4.png
Greyscale
directly control gradient magnitudes per task. The computed gradients
PNG
media_image10.png
16
52
media_image10.png
Greyscale
are then applied via standard update rules to update each
PNG
media_image2.png
14
16
media_image2.png
Greyscale
(as shown in Figure 1) [ and computing the weighted sum of gradients of the plurality of losses based on the generated plurality of gradient weights.].”)
Li, Gao, Hildebrandt, and Chen are considered analogous to the claimed invention because they both are in the same field of endeavor (loss functions). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, Hildebrandt, and Chen to teach the above limitation(s). The motivation for doing so is advantageous because “To optimize the weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
for gradient balancing, we propose a simple algorithm that penalizes the network when backpropagated gradients from any task are too large or too small.” (Chen, paragraph 1, line 43).
Regarding claim 12, Li in view of Gao, in view of Hildebrandt teaches the method of claim 8. Li, Gao, and Hildebrandt do not explicitly disclose, but Chen does disclose
operation of training the recommendation neural network comprising: computing the plurality of gradients corresponding to the plurality of losses; (Chen, paragraph 1, line 33 “In practice, the multitask loss function is often assumed to be linear in the single task losses
PNG
media_image1.png
14
102
media_image1.png
Greyscale
, where the sum runs over all T tasks. In our case, we propose an adaptive method, and so
PNG
media_image2.png
14
16
media_image2.png
Greyscale
can vary at each training step
PNG
media_image3.png
16
92
media_image3.png
Greyscale
This linear form of the loss function is convenient for implementing gradient balancing, as
PNG
media_image2.png
14
16
media_image2.png
Greyscale
very directly and linearly couples to the backpropagated gradient magnitudes from each task. The challenge is then to find the best value for each
PNG
media_image2.png
14
16
media_image2.png
Greyscale
at each training step t that balances the contribution of each task for optimal model training. To optimize the weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
for gradient balancing, we propose a simple algorithm that penalizes the network when backpropagated gradients from any task are too large or too small. [operation of training the recommendation neural network comprising: computing the plurality of gradients corresponding to the plurality of losses; the training the recommendation neural network model comprises: computing the plurality of gradients corresponding to the plurality of losses;]”.)
and modifying a magnitude of at least one gradient in the plurality of task gradients based on a current training epoch number (Chen, paragraph 1, line 27 “task imbalances impede proper training because they manifest as imbalances between backpropagated gradients. A task that is too dominant during training, for example, will necessarily express that dominance by inducing gradients which have relatively large magnitudes. We aim to mitigate such issues at their root by directly modifying gradient magnitudes through tuning of the multitask loss function [and modifying a magnitude of at least one gradient in the plurality of task gradients based on a current training epoch number.].”)
Li, Gao, Hildebrandt, and Chen are considered analogous to the claimed invention because they both are in the same field of endeavor (loss functions). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, Hildebrandt, and Chen to teach the above limitation(s). The motivation for doing so is advantageous because “task imbalances impede proper training because they manifest as imbalances between backpropagated gradients. A task that is too dominant during training, for example, will necessarily express that dominance by inducing gradients which have relatively large magnitudes…to mitigate such issues at their root by directly modifying gradient magnitudes through tuning of the multitask loss function.” (Chen, paragraph 1, line 27).
Regarding claim 13, Li in view of Gao, in view of Hildebrandt teaches the method of claim 8. Li, Gao, and Hildebrandt do not explicitly disclose, but Chen does disclose
the operating of training the recommendation neural network comprising: generating a plurality of gradient weights from magnitudes of the plurality of gradients, including the at least one gradient with the modified magnitude; (Chen, paragraph 1, line 35, “In our case, we propose an adaptive method, and so
PNG
media_image2.png
14
16
media_image2.png
Greyscale
can vary at each training step
PNG
media_image3.png
16
92
media_image3.png
Greyscale
This linear form of the loss function is convenient for implementing gradient balancing, as
PNG
media_image2.png
14
16
media_image2.png
Greyscale
very directly and linearly couples to the backpropagated gradient magnitudes from each task. The challenge is then to find the best value for each
PNG
media_image2.png
14
16
media_image2.png
Greyscale
at each training step t that balances the contribution of each task for optimal model training. To optimize the weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
for gradient balancing, we propose a simple algorithm that penalizes the network when backpropagated gradients from any task are too large or too small. The correct balance is struck when tasks are training at similar rates; if task i is training relatively quickly, then its weight
PNG
media_image4.png
22
28
media_image4.png
Greyscale
should decrease relative to other task weights
PNG
media_image5.png
16
55
media_image5.png
Greyscale
to allow other tasks more influence on GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networks training [the training the recommendation neural network model comprises: generating a plurality of gradient weights from magnitudes of the plurality of gradients, including the at least one gradient with the modified magnitude;]”.)
and computing the weighted sum of gradients of the plurality of losses based on the generated plurality of gradient weights. (Chen, paragraph 3.2, line 23 “Equation 1 gives a target for each task i’s gradient norms, and we update our loss weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
to move gradient norms towards this target for each task. GradNorm is then implemented as an
PNG
media_image6.png
24
16
media_image6.png
Greyscale
loss function
PNG
media_image7.png
14
28
media_image7.png
Greyscale
between the actual and target gradient norms at each timestep for each task, summed over all tasks:
PNG
media_image8.png
50
320
media_image8.png
Greyscale
where the summation runs through all T tasks. When differentiating this loss
PNG
media_image7.png
14
28
media_image7.png
Greyscale
, we treat the target gradient norm
PNG
media_image9.png
20
100
media_image9.png
Greyscale
as a fixed constant to prevent loss weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
from spuriously drifting towards zero.
PNG
media_image7.png
14
28
media_image7.png
Greyscale
is then differentiated only with respect to the
PNG
media_image2.png
14
16
media_image2.png
Greyscale
, as the
PNG
media_image4.png
22
28
media_image4.png
Greyscale
directly control gradient magnitudes per task. The computed gradients
PNG
media_image10.png
16
52
media_image10.png
Greyscale
are then applied via standard update rules to update each
PNG
media_image2.png
14
16
media_image2.png
Greyscale
(as shown in Figure 1) [ and computing the weighted sum of gradients of the plurality of losses based on the generated plurality of gradient weights.].”)
Li, Gao, Hildebrandt, and Chen are considered analogous to the claimed invention because they both are in the same field of endeavor (loss functions). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, Hildebrandt, and Chen to teach the above limitation(s). The motivation for doing so is advantageous because “To optimize the weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
for gradient balancing, we propose a simple algorithm that penalizes the network when backpropagated gradients from any task are too large or too small.” (Chen, paragraph 1, line 43).
Regarding claim 19, Li in view of Gao, in view of Hildebrandt teaches the method of claim 15. Li, Gao, and Hildebrandt do not explicitly disclose, but Chen does disclose
the training the recommendation neural network model comprises: computing the plurality of gradients corresponding to the plurality of losses; (Chen, paragraph 1, line 33 “In practice, the multitask loss function is often assumed to be linear in the single task losses
PNG
media_image1.png
14
102
media_image1.png
Greyscale
, where the sum runs over all T tasks. In our case, we propose an adaptive method, and so
PNG
media_image2.png
14
16
media_image2.png
Greyscale
can vary at each training step
PNG
media_image3.png
16
92
media_image3.png
Greyscale
This linear form of the loss function is convenient for implementing gradient balancing, as
PNG
media_image2.png
14
16
media_image2.png
Greyscale
very directly and linearly couples to the backpropagated gradient magnitudes from each task. The challenge is then to find the best value for each
PNG
media_image2.png
14
16
media_image2.png
Greyscale
at each training step t that balances the contribution of each task for optimal model training. To optimize the weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
for gradient balancing, we propose a simple algorithm that penalizes the network when backpropagated gradients from any task are too large or too small. [operation of training the recommendation neural network comprising: computing the plurality of gradients corresponding to the plurality of losses; the training the recommendation neural network model comprises: computing the plurality of gradients corresponding to the plurality of losses;]”.)
and modifying a magnitude of at least one gradient in the plurality of task gradients based on a current training epoch number (Chen, paragraph 1, line 27 “task imbalances impede proper training because they manifest as imbalances between backpropagated gradients. A task that is too dominant during training, for example, will necessarily express that dominance by inducing gradients which have relatively large magnitudes. We aim to mitigate such issues at their root by directly modifying gradient magnitudes through tuning of the multitask loss function [and modifying a magnitude of at least one gradient in the plurality of task gradients based on a current training epoch number.].”)
Li, Gao, Hildebrandt, and Chen are considered analogous to the claimed invention because they both are in the same field of endeavor (loss functions). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, Hildebrandt, and Chen to teach the above limitation(s). The motivation for doing so is advantageous because “task imbalances impede proper training because they manifest as imbalances between backpropagated gradients. A task that is too dominant during training, for example, will necessarily express that dominance by inducing gradients which have relatively large magnitudes…to mitigate such issues at their root by directly modifying gradient magnitudes through tuning of the multitask loss function.” (Chen, paragraph 1, line 27).
Regarding claim 20, Li in view of Gao, in view of Hildebrandt teaches the method of claim 19. Li, Gao, and Hildebrandt do not explicitly disclose, but Chen does disclose
the training the recommendation neural network model comprises: generating a plurality of gradient weights from magnitudes of the plurality of gradients, including the at least one gradient with the modified magnitude; (Chen, paragraph 1, line 35, “In our case, we propose an adaptive method, and so
PNG
media_image2.png
14
16
media_image2.png
Greyscale
can vary at each training step
PNG
media_image3.png
16
92
media_image3.png
Greyscale
This linear form of the loss function is convenient for implementing gradient balancing, as
PNG
media_image2.png
14
16
media_image2.png
Greyscale
very directly and linearly couples to the backpropagated gradient magnitudes from each task. The challenge is then to find the best value for each
PNG
media_image2.png
14
16
media_image2.png
Greyscale
at each training step t that balances the contribution of each task for optimal model training. To optimize the weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
for gradient balancing, we propose a simple algorithm that penalizes the network when backpropagated gradients from any task are too large or too small. The correct balance is struck when tasks are training at similar rates; if task i is training relatively quickly, then its weight
PNG
media_image4.png
22
28
media_image4.png
Greyscale
should decrease relative to other task weights
PNG
media_image5.png
16
55
media_image5.png
Greyscale
to allow other tasks more influence on GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networks training [the training the recommendation neural network model comprises: generating a plurality of gradient weights from magnitudes of the plurality of gradients, including the at least one gradient with the modified magnitude;]”.)
and computing the weighted sum of gradients of the plurality of losses based on the generated plurality of gradient weights. (Chen, paragraph 3.2, line 23 “Equation 1 gives a target for each task i’s gradient norms, and we update our loss weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
to move gradient norms towards this target for each task. GradNorm is then implemented as an
PNG
media_image6.png
24
16
media_image6.png
Greyscale
loss function
PNG
media_image7.png
14
28
media_image7.png
Greyscale
between the actual and target gradient norms at each timestep for each task, summed over all tasks:
PNG
media_image8.png
50
320
media_image8.png
Greyscale
where the summation runs through all T tasks. When differentiating this loss
PNG
media_image7.png
14
28
media_image7.png
Greyscale
, we treat the target gradient norm
PNG
media_image9.png
20
100
media_image9.png
Greyscale
as a fixed constant to prevent loss weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
from spuriously drifting towards zero.
PNG
media_image7.png
14
28
media_image7.png
Greyscale
is then differentiated only with respect to the
PNG
media_image2.png
14
16
media_image2.png
Greyscale
, as the
PNG
media_image4.png
22
28
media_image4.png
Greyscale
directly control gradient magnitudes per task. The computed gradients
PNG
media_image10.png
16
52
media_image10.png
Greyscale
are then applied via standard update rules to update each
PNG
media_image2.png
14
16
media_image2.png
Greyscale
(as shown in Figure 1) [ and computing the weighted sum of gradients of the plurality of losses based on the generated plurality of gradient weights.].”)
Li, Gao, Hildebrandt, and Chen are considered analogous to the claimed invention because they both are in the same field of endeavor (loss functions). It would have been obvious for a person having ordinary skill in the art before the effective filing date of the claimed invention to combine Li, Gao, Hildebrandt, and Chen to teach the above limitation(s). The motivation for doing so is advantageous because “To optimize the weights
PNG
media_image4.png
22
28
media_image4.png
Greyscale
for gradient balancing, we propose a simple algorithm that penalizes the network when backpropagated gradients from any task are too large or too small.” (Chen, paragraph 1, line 43).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Tan et al., Patent Application Publication (20240062071) also discloses recommendation model for users, embeddings, encoders, backpropagation, weights, and losses.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MARIETOU SADIYA SECK whose telephone number is (571)270-5987. The examiner can normally be reached Monday-Friday (8:30 am - 5:00pm) EST.
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, Michelle Bechtold can be reached at (571) 431-0762. 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.
/MSS/Examiner, Art Unit 2148
/MICHELLE T BECHTOLD/Supervisory Patent Examiner, Art Unit 2148
Prosecution Insights