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 .
Claims 1-20 are pending.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. Representative claim 11 recites “receiving, …., a query associated with a supply chain network; representing the supply chain network as at least one graph based on historical transactions in the supply chain network; obtaining at least one machine learning model that is trained based on graph data related to nodes and edges in the at least one graph; generating, …, supply chain prediction data based on the query; and transmitting the supply chain prediction data …”. Therefore, the claim as a whole is directed to “Graphing Supply Chain Data for Making Predictions”, which is an abstract idea because it is a method of organizing human activity including commercial interactions (including agreements in the form of contracts; sales activities or behaviors; and business relations) and mental process, including concepts performed in the human mind (including an observation, evaluation, judgment, opinion). “Graphing Supply Chain Data for Making Predictions” is considered to be is a method of organizing human activity because the claims are directed to human processes of analyzing business relationships that form a supply chain, graphing those relationships, and analyzing that graph to make predictions is a business practice.
“Graphing Supply Chain Data for Making Predictions” is also considered to be is a mental process because the graphing of business relationships in a supply chain may be done with pen and paper, and a response to a query may be a mental process. That is, the analysis of graph data to visualize business relationships allows the predictions of changes in business practices such as seeking redundant suppliers or changing ordering quantities based on time related price fluctuations. As such, the claims are directed to an abstract idea.
Claims 1 and 20 recite substantially similar features to those recited in representative claim 11 and are ineligible based on substantially the same reasons.
This judicial exception is not integrated into a practical application. In particular, claims 1, 11, and 20 recites the following additional element(s): generating, using the at least one machine learning model, supply chain prediction data; claim 1: a non-transitory memory; and at least one processor, and at least one machine learning model; claim 11 and 20: a computing device and at least one machine learning model. These additional elements individually or in combination do not integrate the exception into a practical application. That is, the recitations of additional elements amount merely reciting the words ‘‘apply it’’ (or an equivalent) with the judicial exception, or merely including instructions to implement an abstract idea on a computer, or merely using a computer as a tool to perform an abstract idea (see MPEP 2106.05(f)). That is, the recitations to computer technology are high level recitations that amount to providing a tool to implement the organizing human activity or mental processes to which the claims are directed. Here, there are no details about a particular machine learning model or how the machine learning model operates other than that it is being used to determine the supply chain prediction data. The machine learning model is used to generally apply the abstract idea without placing any limitation on how the machine learning model operates. In addition, the limitation recites only the idea of generating supply chain prediction data using a machine learning model without details on how this is accomplished. The claim omits any details as to how the machine learning model solves a technical problem, and instead recites only the idea of a solution or outcome. Also, the claim invokes a generic machine learning model merely as a tool for making the recited mathematical calculation rather than purporting to improve the technology or a computer. See MPEP 2106.05(f). Therefore, the limitation represents no more than mere instructions to apply the judicial exception on a computer. It can also be viewed as nothing more than an attempt to generally link the use of a judicial exception to a particular technological environment or field of use (see MPEP 2106.05(h)). Accordingly, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. Claims 1, 11, and 20 are directed to an abstract idea.
Claims 1, 11, and 20 do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements, individually and in combination, are merely being used to apply the abstract idea to a technological environment. As noted above, the additional elements are recited at a high level of generality. There are no details about a particular machine learning model or how the machine learning model operates other than that it is being used to determine the supply chain prediction data. The machine learning model is used to generally apply the abstract idea without placing any limitation on how the machine learning model operates. Such high level recitations cannot amount to significantly more. Accordingly, claims 1, 11, and 20 are ineligible.
Dependent claims 2-10 and 12-19 merely further limit the abstract idea and are thereby considered to be ineligible.
Dependent claims 2 and 12 further limit the abstract idea of “Graphing Supply Chain Data for Making Predictions” by introducing the element of product nodes each representing a product offered for sale by a retailer; customer nodes each representing a customer of the retailer; and shipping nodes each representing a supplier, a manufacturer, a distributor, or the retailer’s store, warehouse, fulfillment center, distribution center, or consolidation center, which does not include an improvement to another technology or technical field, an improvement to the functioning of the computer itself, or meaningful limitations beyond generally linking the use of the abstract idea to a particular technological environment. Therefore, dependent claims 2 and 12 are also non-statutory subject matter.
Dependent claims 3 and 13 further limit the abstract idea of “Graphing Supply Chain Data for Making Predictions” by introducing the element of each edge in the at least one graph is between two nodes in the at least one graph, and represents a relationship between the two nodes; the relationship is associated with transaction-related features; and the edges in the at least one graph comprise at least: an order edge between a customer node and product node, wherein the order edge is associated with features related to at least: order placed time, upcoming events, quantity, price, and item indicators, a cost edge between a product node and a shipping node, wherein the cost edge is associated with costs related to at least: shipping, packaging, fulfilment and inventory levels, a deliver edge between a shipping node and a customer node, wherein the deliver edge is associated with features related to at least: distance, destination region, carriers, transport vehicles, and labor capacity, which does not include an improvement to another technology or technical field, an improvement to the functioning of the computer itself, or meaningful limitations beyond generally linking the use of the abstract idea to a particular technological environment. Therefore, dependent claims 3 and 13 are also non-statutory subject matter.
Dependent claims 4 and 14 further limit the abstract idea of “Graphing Supply Chain Data for Making Predictions” by introducing the element of the supply chain network is represented as the at least one graph based on: obtaining tabular data associated with the historical transactions in the supply chain network; and converting the tabular data for the historical transactions in each day to a daily graph to create daily graphs over a time period, which does not include an improvement to another technology or technical field, an improvement to the functioning of the computer itself, or meaningful limitations beyond generally linking the use of the abstract idea to a particular technological environment. Therefore, dependent claims 4 and 14 are also non-statutory subject matter.
Dependent claims 5 and 15 further limit the abstract idea of “Graphing Supply Chain Data for Making Predictions” by introducing the element of generating an embedding for each node in the daily graph using an auto-encoder; and refining each embedding through message passing in the daily graph, which does not include an improvement to another technology or technical field, an improvement to the functioning of the computer itself, or meaningful limitations beyond generally linking the use of the abstract idea to a particular technological environment. Therefore, dependent claims 5 and 15 are also non-statutory subject matter.
Dependent claims 6 and 16 further limit the abstract idea of “Graphing Supply Chain Data for Making Predictions” by introducing the element of generating a sequence of the daily graphs to create a graph dataset including temporal aspect of the sequence; splitting the graph dataset into a plurality of batches to generate a training dataset; and …, which does not include an improvement to another technology or technical field, an improvement to the functioning of the computer itself, or meaningful limitations beyond generally linking the use of the abstract idea to a particular technological environment. Therefore, dependent claims 6 and 16 are also non-statutory subject matter.
Dependent claims 7 and 17 further limit the abstract idea of “Graphing Supply Chain Data for Making Predictions” by introducing the element of the at least one machine learning model includes a model sequence of machine learning models; the machine learning models in the model sequence include at least one graph neural network (GNN); each of the machine learning models in the model sequence is utilized to perform a corresponding task in the supply chain network; all of the machine learning models in the model sequence are trained based on the same training dataset; and an output of one machine learning model in the model sequence serves as an input to a succeeding machine learning model in the model sequence, which does not include an improvement to another technology or technical field, an improvement to the functioning of the computer itself, or meaningful limitations beyond generally linking the use of the abstract idea to a particular technological environment. Therefore, dependent claims 7 and 17 are also non-statutory subject matter.
Dependent claims 8 and 18 further limit the abstract idea of “Graphing Supply Chain Data for Making Predictions” by introducing the element of: the machine learning models in the model sequence include two GNNs generated based on two graphs respectively; and to connect the two graphs using at least one edge extending between nodes in the two graphs, generate prediction data for a simulated supply chain operation based on a first GNN of the two GNNs, and optimize one or more parameters of a second GNN of the two GNNs based on the prediction data, which does not include an improvement to another technology or technical field, an improvement to the functioning of the computer itself, or meaningful limitations beyond generally linking the use of the abstract idea to a particular technological environment. Therefore, dependent claims 8 and 18 are also non-statutory subject matter.
Dependent claims 9 and 19 further limit the abstract idea of “Graphing Supply Chain Data for Making Predictions” by introducing the element of determining, based on the query, a task in the supply chain network and a reward function associated with the task; training an automated agent to learn a manner of performing simulated actions in a dynamic environment of the supply chain network, wherein each simulated action is assigned a reward value based on the reward function, wherein the automated agent is trained based on a graph neural network in the at least one machine learning model to maximize a cumulative reward value based on the performed simulated actions; determining at least one simulated change of the supply chain network based on the query; and generating the supply chain prediction data based on a simulation of the supply chain network using the graph neural network based on the at least one simulated change, wherein the automated agent is configured to perform, during the simulation, simulated actions according to the learned manner based on modifications of the graph neural network in response to the at least one simulated change, which does not include an improvement to another technology or technical field, an improvement to the functioning of the computer itself, or meaningful limitations beyond generally linking the use of the abstract idea to a particular technological environment. Therefore, dependent claims 9 and 19 are also non-statutory subject matter.
Dependent claim 10 further limits the abstract idea of “Graphing Supply Chain Data for Making Predictions” by introducing the element of determine, based on the simulation …., insight data of the supply chain prediction data, wherein the insight data includes at least one root cause of a performance of the supply chain network in response to the at least one simulated change; and transmit the supply chain prediction data together with the insight data …., which does not include an improvement to another technology or technical field, an improvement to the functioning of the computer itself, or meaningful limitations beyond generally linking the use of the abstract idea to a particular technological environment. Therefore, dependent claim 10 is also non-statutory subject matter.
Dependent claims 2-10 and 12-19 also do not integrated into a practical application. Dependent claims 6 and 16 recite training the at least one machine learning model based on the training dataset, claims 7 and 17 recite at least one machine learning model includes a model sequence of machine learning models; the machine learning models in the model sequence include at least one graph neural network (GNN); each of the machine learning models in the model sequence is utilized to perform a corresponding task in the supply chain network; all of the machine learning models in the model sequence are trained based on the same training dataset; and an output of one machine learning model in the model sequence serves as an input to a succeeding machine learning model in the model sequence; Claims 9 and 19 recite training an automated agent to learn a manner of performing simulated actions in a dynamic environment of the supply chain network, wherein each simulated action is assigned a reward value based on the reward function, wherein the automated agent is trained based on a graph neural network in the at least one machine learning model to maximize a cumulative reward value based on the performed simulated actions; determining at least one simulated change of the supply chain network based on the query; and generating the supply chain prediction data based on a simulation of the supply chain network using the graph neural network based on the at least one simulated change, wherein the automated agent is configured to perform, during the simulation, simulated actions according to the learned manner based on modifications of the graph neural network in response to the at least one simulated change. These additional elements merely generally link the abstract idea to a particular technological environment or field of use. MPEP 2106.04(d)(I) indicates that generally linking an abstract idea to a particular technological environment or field of use cannot provide a practical application. That is, while the use of additional elements such as graph neural networks is recited, the uses of this technology is seen to be the use of known technology for a specific form of data, not to address any technological problem or provide any technological solution. Accordingly, even in combination, these additional elements do not integrate the abstract idea into a practical application. This has been re-evaluated under the “significantly more” analysis and has also been found insufficient to provide significantly more. MPEP 2106.05(A) indicates that generally linking an abstract idea to a particular technological environment or field of use cannot provide significantly more. As such, the recitations of additional elements individually or in combination do not integrate the exception into a practical application, but rather, the recitation of any recitations of additional elements amounts to merely reciting the words ‘‘apply it’’ (or an equivalent) with the judicial exception, or merely including instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea (see MPEP 2106.05(f)). That is, the claims provide no practical limits or improvements to any technology. Accordingly, dependent claims 2-10 and 12-19 are also ineligible.
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of pre-AIA 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claims 1, 11, and 20 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by U.S. Patent Application Publication No. 20190311312 to Leidner et al.
With regards to claims 1, 11, and 20, Leidner et al. teaches:
a non-transitory memory having instructions stored thereon; and at least one processor operatively coupled to the non-transitory memory (paragraph [0046]), and configured to read the instructions to:
receive, from a computing device, a query associated with a supply chain network (paragraph [0016], “identification module (“IAIM”) for permitting the receipt of a set of information, the IAIM comprising executable code adapted to determine whether the set of information contains data related to one or more of a supplier, a commodity, and a customer; and an instantiated query generation module (“IQGM”) communicatively coupled to the IAIM for generating a query comprising a supplier entry, a commodity entry, and a customer entry, the IQGM comprising a placeholder generation module for inserting a placeholder into the query for one or more of: the supplier entry if the IAIM determines a supplier absence in the set of information; the commodity entry if the IAIM determines a commodity absence in the set of information; and the customer entry if the IAIM determines a customer absence in the set of information; a transceiver for sending the query and receiving a set of supply chain information;”),
represent the supply chain network as at least one graph based on historical transactions in the supply chain network (paragraph [0014], “The mined information is then parsed and processed to generate a supply chain graph. Once a diagrammatic depiction of the supply chain graph for a company or industry is generated, value can be derived from the graph for analyzing and forecasting supply and demand of the resources, e.g., commodities, represented in the graph. Furthermore, the supply chain graph can be used to predict risk throughout the supply chain.”),
obtain at least one machine learning model that is trained based on graph data related to nodes and edges in the at least one graph (paragraph [0038], “Quantitative analysis, techniques or mathematics and models associated with modules 124 to 128 in conjunction with computer science are processed by processor 121 of server 120 thereby rendering server 120 into a special purpose computing machine use to transform records and data related to commodity transactions found in documents and other information in SCGS databases 110 into supply chain graph representations and to arrive at predictive behavior, and potentially predictive representations, for use by business analysts.”; paragraph [0055], “The set of triples is then used by SCGSGM 390 to construct a supply chain graph. In one exemplary implementation, the supply chain graph is generated by turning each of the supplier entity and customer entity into two graph nodes. The nodes are connected by a vertex labeled with the commodity type. The supply chain graph may use additional nodes and vertices by additional triples from the set of triples to the graph. The process of adding triples to the graph builds out a comprehensive supply chain graph that provides a user with an enhanced tool and experience in analyzing and forecasting supply and demand. The resulting supply chain graph signal generated by SCGSGM 390 is sent to transmitter 380 to be sent as supply chain graph signal 304.”),
generate, using the at least one machine learning model, supply chain prediction data based on the query (paragraph [0042], “The historical database or corpus may be separate from or derived from SCGS database set 110, which may comprise continuous feeds and may be updated, e.g., in near or close to real time, allowing the SCGS to automatically and timely analyze content, update supply chain visualizations based on “new” content, and generate commodity trade or predictive signals in close to real-time, i.e., within approximately one second.”; paragraph [0051], “The methods and systems of the present invention, described in detail hereafter, may be employed in providing remote users, such as investors, access to a searchable database. In particular, remote users may search a database using search queries based on company RIC, a commodity listing, stock or other name to retrieve and view predictive analysis and/or suggested action as discussed hereinbelow.”), and
transmit the supply chain prediction data to the computing device (paragraph [0051], “The methods and systems of the present invention, described in detail hereafter, may be employed in providing remote users, such as investors, access to a searchable database. In particular, remote users may search a database using search queries based on company RIC, a commodity listing, stock or other name to retrieve and view predictive analysis and/or suggested action as discussed hereinbelow. … Client side application software may be stored on machine-readable medium and comprising instructions executed, for example, by the processor 220 of computer 211, and presentation of web-based interface screens facilitate the interaction between user system 210 and central system 211, such as tools for further analyzing the data streams and other data and reports received via network 226 and stored locally or accessed remotely.”).
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 2 and 12 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Patent Application Publication No. 20190311312 to Leidner et al. as applied to claims 1, 11, and 20 above, and further in view of “Towards knowledge graph reasoning for supply chain risk management using graph neural networks” by Kosasih et al.
With regards to claims 2 and 12, Leidner et al. teaches: the nodes in the at least one graph (paragraph [0017], “The set of supply chain visualization information may comprise a set of triples comprising a supplying entity, a commodity type, and a supplied entity. The supply chain visualization signal may comprise at least one graph node and at least one connecting vertex. The supply chain visualization signal may comprise a first node representing a supplying entity and a second node representing a customer entity and a connecting vertex representing a commodity type.”) comprise at least:
…
customer nodes each representing a customer of the retailer (paragraph [0032], “The set of triples returned using the instantiated query q′ can be used to construct a supply chain graph by turning each triple into two graph nodes, in this example for SupplyingEntity and SuppliedEntity respectively, which are connected using a vertex labeled with CommodityType.”); and
shipping nodes each representing a supplier, a manufacturer, a distributor, or the retailer’s store, warehouse, fulfillment center, distribution center, or consolidation center (paragraph [0032], “The set of triples returned using the instantiated query q′ can be used to construct a supply chain graph by turning each triple into two graph nodes, in this example for SupplyingEntity and SuppliedEntity respectively, which are connected using a vertex labeled with CommodityType. In the present example nodes are created for “Marathon Oil Corporation” (SupplyingEntity) and “BP” (SuppliedEntity) with an arc (directed line) connecting the nodes labeled “oil” (CommodityType).”), but fails to explicitly teach product nodes each representing a product offered for sale by a retailer.
However Kosasih et al. teaches product nodes each representing a product offered for sale by a retailer (Section 4 “The core GNN engine is built using a widely used software package called PyTorch Geometric (Fey and Lenssen Citation2019). Meanwhile, the supply chain dataset is queried from different databases, containing information about different companies, their products, certifications, etc., and how they are dependent with each other.”, Section 5.1 “The Automotive dataset is from Marklines, a supplier information company located in Japan. The dataset contains five types of entities: company, country, certification, product and capability. … Properties of the training and testing datasets are shown in Table 4. As it can be seen, the graph has a large number of ‘makes product’ and ‘supplier to’ edges, however, ‘located in’ and ‘has certificate’ edges are relatively smaller in number.”).
This part of Kosasih et al. is applicable to the system of Leidner et al. as they both share characteristics and capabilities, namely, they are directed to supply chain graph generation. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Leidner et al. to include the product nodes as taught by Kosasih et al. One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to modify Leidner et al. to utilize different data types and attributes from a taxonomy of items that are present in the data in order to predict missing links and add them in the graph so knowledge gaps are closed (see Section 3 of Kosasih et al.).
Claims 3 and 13 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Patent Application Publication No. 20190311312 to Leidner et al. and “Towards knowledge graph reasoning for supply chain risk management using graph neural networks” by Kosasih et al. as applied to claims 2 and 12 above, and further in view of U.S. Patent No. 12511614 to Mohanty et al.
With regards to claims 3 and 13, Leidner et al. fails to explicitly teach, but Mohanty et al. teaches:
each edge in the at least one graph is between two nodes in the at least one graph, and represents a relationship between the two nodes (Col. 11, lines 63-67, “While constructing the graphical model, probability module 202 may generate edges connecting each node in the graph, with further refinement removing edges when learning module 208 calculates that they do not represent relationships present in supply chain data 262.”);
the relationship is associated with transaction-related features (col 18, lines 22-37, “Database 154 of supply chain planner 150 may comprise one or more databases or other data storage arrangement at one or more locations, local to, or remote from, server 152. Database 154, for example, transaction data 260, supply chain data 262, product data 264, inventory data 266, inventory policies 268, store data 270, customer data 272, supply chain models 274, and levers 276. Although database 154 is illustrated and described as comprising transaction data 260, supply chain data 262, product data 264, inventory data 266, inventory policies 266, store data 270, customer data 272, supply chain models 274, and levers 276, embodiments contemplate any suitable number or combination of data, located at one or more locations, local to, or remote from, supply chain planner 150, according to particular needs.”); and
the edges in the at least one graph comprise at least: an order edge between a customer node and a product node (col. 11, lines 60-67, “Probability module 202 may construct a graphical model in which each node represents one of the identified attributes. While constructing the graphical model, probability module 202 may generate edges connecting each node in the graph, with further refinement removing edges when learning module 208 calculates that they do not represent relationships present in supply chain data 262.”), wherein
the order edge is associated with features related to at least: order placed time, upcoming events, quantity, price, and item indicators (col. 17, line 60, through col. 18, line 1, “According to one embodiment, historical data 248 comprises historic sales patterns, prices, promotions, weather conditions and other factors influencing demand of one or more items sold in one or more stores over a time period, such as, for example, one or more days, weeks, months, years, including, for example, a day of the week, a day of the month, a day of the year, week of the month, week of the year, month of the year, special events, paydays, and the like.”),
a cost edge between a product node and a shipping node, wherein the cost edge is associated with costs related to at least: shipping, packaging, fulfilment and inventory levels (col. 14, lines 31-52, “Limitations on supplying materials and items to particular buffers may represent transportation limitations (e.g. cost, time, available transportation options) or outputs of various operations (such as, for example, different production processes, which produce different items, each of which may be represented by a different SKU, and which each may be stored at different buffers). Although the limitation of the flow of items between nodes of supply chain network model 220 is described as cost, timing, transportation, or production limitations, embodiments contemplate any suitable flow of items or limitations of the flow of items between any one or more different nodes of a supply chain network, according to particular needs.”),
a deliver edge between a shipping node and a customer node, wherein the deliver edge is associated with features related to at least: distance, destination region, carriers, transport vehicles, and labor capacity (col. 13, lines 33-41, “Various transportation or manufacturing processes are modelled as edges connecting the nodes. Each edge may represent the flow, transportation, or assembly of materials (such as items or resources) between the nodes by, for example, production processing or transportation.”).
This part of Mohanty et al. is applicable to the system of Leidner et al. as they both share characteristics and capabilities, namely, they are directed to supply chain graph generation. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Leidner et al. to include the data set based edge types as taught by Mohanty et al. One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to modify Leidner et al. to incorporating the one or more features in the of available list of data and determine one or more inferences pertaining to the one or more supply chain entity target variables (see abstract of Mohanty et al.).
Claims 4-6 and 14-16 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Patent Application Publication No. 20190311312 to Leidner et al. as applied to claims 1, 11, and 20 above, and further in view of U.S. Patent Application Publication No. 20200233864 to Jin et al.
With regards to claims 4 and 14, Leidner et al. teaches: the supply chain network is represented as the at least one graph based on:
obtaining tabular data associated with the historical transactions in the supply chain network (paragraph [0038], “Quantitative analysis, techniques or mathematics and models associated with modules 124 to 128 in conjunction with computer science are processed by processor 121 of server 120 thereby rendering server 120 into a special purpose computing machine use to transform records and data related to commodity transactions found in documents and other information in SCGS databases 110 into supply chain graph representations and to arrive at predictive behavior, and potentially predictive representations, for use by business analysts.”); and
converting the tabular data for the historical transactions … to create … graphs over a time period (paragraph [0038], “Quantitative analysis, techniques or mathematics and models associated with modules 124 to 128 in conjunction with computer science are processed by processor 121 of server 120 thereby rendering server 120 into a special purpose computing machine use to transform records and data related to commodity transactions found in documents and other information in SCGS databases 110 into supply chain graph representations and to arrive at predictive behavior, and potentially predictive representations, for use by business analysts.”), but fails to explicitly teach creating daily graphs over a time period.
However Jin et al. teaches converting the tabular data for the historical transactions in each day to a daily graph to create daily graphs over a time period (paragraph [0077], “For example, anomalous subgraph detection can be performed on real-world graphs by constructing two graphs G1 and G2 (e.g., using two consecutive daily datasets).”; paragraph [0079], “Daily graphs were generated with nodes representing entities such as keywords or hashtags appearing in Twitter, and edges denoting their coexistence (co-mentions) on a particular day. As with anomalous subgraph detection, event detection can be performed by constructing consecutive (e.g., daily) graphs Gt−1 and Gt, learning a latent summary * from Gt−1, and inductively learning node embeddings for Gt using the latent summary *.”).
This part of Jin et al. is applicable to the system of Leidner et al. as they both share characteristics and capabilities, namely, they are directed to graph generating and analysis. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Leidner et al. to include the daily tabular graph data as taught by Jin et al. One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to modify Leidner et al. in order to effectively capture higher-order structural information of different node-centric subgraphs with increasing size (see paragraphs [0004]-[0006] of Jin et al.).
With regards to claims 5 and 15, Leidner et al. fails to explicitly teach creating daily graphs over a time period. However Jin et al. teaches
generating an embedding for each node in the daily graph using an auto-encoder (paragraph [0041], “Any number of graph-based tasks can be performed using a latent summary, or a portion thereof. For example, inductive summarization component 245 can automatically compose new relational functions to capture structural features that are transferable. Thus, the feature matrices learned on one graph can be transferred to another graph for inductive learning tasks. In another example, node embedding component 250 can derive node embeddings from a latent summary on-the-fly. In this manner, the node embeddings need not be stored, but rather, can be generated as needed, thereby saving storage space.”; paragraph [0076], “A latent summary can be used to perform any type of graph-based task such as node classification, node clustering, link prediction, entity resolution, anomaly and event detection, and others. … The latent summarization system can learn node embeddings for each dataset G(V, E) by creating a subgraph G′(V, E′) that includes all nodes of G, but randomly excludes some subset of edges (e.g., 40%). Node embeddings can be derived from G′ (e.g., based on its latent summary) and some subset of edges (e.g., 10%|E|) can be randomly selected for training data.”); and
refining each embedding through message passing in the daily graph (paragraph [0079], “As with anomalous subgraph detection, event detection can be performed by constructing consecutive (e.g., daily) graphs Gt−1 and Gt, learning a latent summary * from Gt−1, and inductively learning node embeddings for Gt using the latent summary *. Node embeddings for consecutive graphs (e.g., days) can be compared to identify abrupt changes of graph structures. For example, the Frobenius norm can be computed for each graph, and any measure of deviation may used to identify deviating graphs (e.g., standard deviation, top-n deviating graphs, etc.).”).
This part of Jin et al. is applicable to the system of Leidner et al. as they both share characteristics and capabilities, namely, they are directed to graph generating and analysis. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Leidner et al. to include the daily tabular graph data as taught by Jin et al. One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to modify Leidner et al. in order to effectively capture higher-order structural information of different node-centric subgraphs with increasing size (see paragraphs [0004]-[0006] of Jin et al.).
With regards to claims 6 and 16, Leidner et al. fails to explicitly teach, but Jin et al. teaches: the at least one machine learning model is trained based on:
generating a sequence of the daily graphs to create a graph dataset including temporal aspect of the sequence (paragraph [0077], “In some embodiments, a latent summary can be used to perform inductive anomaly detection (e.g., by inductive anomaly detector 160 of FIG. 1), such as anomalous subgraph detection and real-world graph event detection. For example, anomalous subgraph detection can be performed on real-world graphs by constructing two graphs G1 and G2 (e.g., using two consecutive daily datasets)”);
splitting the graph dataset into a plurality of batches to generate a training dataset (paragraph [0076], “Generally, training, test, and ground truth data sets can be generated. An edge ei j can be represented by concatenating node-embeddings of its source and destination nodes: emb(ei j)=[emb(i), emb(j)]. The latent summarization system can learn node embeddings for each dataset G(V, E) by creating a subgraph G′(V, E′) that includes all nodes of G, but randomly excludes some subset of edges (e.g., 40%). Node embeddings can be derived from G′ (e.g., based on its latent summary) and some subset of edges (e.g., 10%|E|) can be randomly selected for training data. Out of the removed edges, some subset (e.g., 25% (10%|E|)) can be used as missing links for testing.”); and
training the at least one machine learning model based on the training dataset (paragraph [0076], “Node embeddings can be derived from G′ (e.g., based on its latent summary) and some subset of edges (e.g., 10%|E|) can be randomly selected for training data. Out of the removed edges, some subset (e.g., 25% (10%|E|)) can be used as missing links for testing. Fake edges can be randomly created for the training and testing datasets. As such, the predictive algorithm can be trained to predict missing links (e.g., recommendations) or incorrect links.”).
This part of Jin et al. is applicable to the system of Leidner et al. as they both share characteristics and capabilities, namely, they are directed to graph generating and analysis. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Leidner et al. to include the daily tabular graph data as taught by Jin et al. One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to modify Leidner et al. in order to effectively capture higher-order structural information of different node-centric subgraphs with increasing size (see paragraphs [0004]-[0006] of Jin et al.).
Claims 9, 10, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Patent Application Publication No. 20190311312 to Leidner et al. as applied to claims 1, 11, and 20 above, and further in view of U.S. Patent Application Publication No. 20220187847 to Cella et al.
With regards to claims 9 and 19, Leidner et al. fails to explicitly teach, but Cella et al. teaches: wherein the supply chain prediction data is generated based on:
determining, based on the query, a task in the supply chain network and a reward function associated with the task (paragraph [2390], “Reinforcement Learning is a machine learning technique for learning optimal behavior in an environment by taking actions and getting feedback, similar to how humans and animals learn by interacting with their environments”);
training an automated agent to learn a manner of performing simulated actions in a dynamic environment of the supply chain network (paragraph [0139], “This value proposition may be amplified when highly configurable robots are designed with the latest functionality and enabled with a high level of artificial intelligence; when the platform is equipped with intelligence and computing capabilities that integrate data from a wide range of sources, including deployed robots, value chain network (VCN) entities involved in a wide range of supply chain activities (such as picking, packing, moving, storing, warehousing, transporting and/or delivering among others) and demand-related activities (such as marketing, selling, advertising, forecasting, pricing, positioning, placing, designing, and others), ERP systems, smart contracts, and the like; and when the platform learns from and manages performance based on operational outcomes.”), wherein each simulated action is assigned a reward value based on the reward function, wherein the automated agent is trained based on a graph neural network in the at least one machine learning model to maximize a cumulative reward value based on the performed simulated actions (paragraph [2390], “The typical reinforcement learning approach includes an agent (say robot control system 12150) that observes its environment, evaluates its current state (e.g., robot velocity, distance to an object in front), and selects an action (e.g., provide control instruction to actuator or motor, adjust velocity, change direction and the like). Upon carrying out an action, the agent is presented with, in addition to its new state, a reward (e.g., +10 for allowing sufficient space between the robot and an obstacle in front of it and −10 for allowing insufficient space) which provides some indication of the success of the action. The goal for a reinforcement learning agent is to learn an optimal policy or behavior that maximizes the expected cumulative reward.”);
determining at least one simulated change of the supply chain network based on the query (paragraph [2391], “Reinforcement learning system 12668 includes one or more reinforcement learning algorithms for evaluating various states, actions and rewards in determining optimal policy for executing one or more tasks by the MPR 12100.”; paragraph [2392], “RPA system 12652 enables MPR 12100 automate workflows as well as any repetitive tasks and processes. In embodiments, the RPA system 12652 may monitor human interaction with various systems to learn patterns and processes performed by humans in performance of respective tasks. In embodiments, an RPA system 12652 may learn to perform certain tasks based on the learned patterns and processes, such that the tasks may be performed by the RPA system 12652 in lieu or in support of a human decision maker.”); and
generating the supply chain prediction data based on a simulation of the supply chain network using the graph neural network based on the at least one simulated change (paragraph [2394], “For example, the analytics system 12660 may perform data analytics on thermal and vibration data generated by the MPR 12100 over a period of time for anomaly, detection, system failure detection, predictive maintenance and for avoiding costly downtime and disruption of operation of the MPR 12100. In another example, the analytics system 12660 may analyze sensor data of the MPR 12100 to generate insights about things like general health of the MPR 12100 efficiency of one or more tasks performed by the MPR 12100, optimal positions and setting for the MPR 12100 and so on.”),
wherein the automated agent is configured to perform, during the simulation, simulated actions according to the learned manner based on modifications of the graph neural network in response to the at least one simulated change (paragraph [2395], “Neural networks represent a system of interconnected “neurons” which send messages to each other. The connections have numeric weights that can be tuned based on experience, making neural nets adaptive to inputs and capable of learning.”).
This part of Cella et al. is applicable to the system of Leidner et al. as they both share characteristics and capabilities, namely, they are directed to graph generating and data analysis. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Leidner et al. to include the reinforcement learning agent based on a reward function as taught by Cella et al. One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to modify Leidner et al. in order to learn an optimal policy or behavior that maximizes the expected cumulative reward (see paragraphs [2390] of Cella et al.).
With regards to claim 10, Leidner et al. fails to explicitly teach, but Cella et al. teaches: the at least one processor is configured to:
determine, based on the simulation and the graph neural network, insight data of the supply chain prediction data (paragraph [2390], “Upon carrying out an action, the agent is presented with, in addition to its new state, a reward (e.g., +10 for allowing sufficient space between the robot and an obstacle in front of it and −10 for allowing insufficient space) which provides some indication of the success of the action. The goal for a reinforcement learning agent is to learn an optimal policy or behavior that maximizes the expected cumulative reward.”), wherein the insight data includes at least one root cause of a performance of the supply chain network in response to the at least one simulated change (paragraph [2394], “In embodiments, an analytics system 12660 is configured to perform various analytical processes on data output from the MPR 12100 or one or more components or subsystems. For example, the analytics system 12660 may perform data analytics on thermal and vibration data generated by the MPR 12100 over a period of time for anomaly, detection, system failure detection, predictive maintenance and for avoiding costly downtime and disruption of operation of the MPR 12100.”); and
transmit the supply chain prediction data together with the insight data to the computing device (paragraph [2392], “RPA system 12652 enables MPR 12100 automate workflows as well as any repetitive tasks and processes. In embodiments, the RPA system 12652 may monitor human interaction with various systems to learn patterns and processes performed by humans in performance of respective tasks. In embodiments, an RPA system 12652 may learn to perform certain tasks based on the learned patterns and processes, such that the tasks may be performed by the RPA system 12652 in lieu or in support of a human decision maker.”).
This part of Cella et al. is applicable to the system of Leidner et al. as they both share characteristics and capabilities, namely, they are directed to graph generating and data analysis. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Leidner et al. to include the reinforcement learning agent based on a reward function as taught by Cella et al. One of ordinary skill in the art before the effective filing date of the claimed invention would have been motivated to modify Leidner et al. in order to learn an optimal policy or behavior that maximizes the expected cumulative reward (see paragraphs [2390] of Cella et al.).
Subject Matter Free of Prior Art
The cited art of record fails to teach or suggest, either alone or in combination, the features found within dependent claims 7, 8, 17 and 18. In particular, the cited art does not teach or suggest to: “the machine learning models in the model sequence include at least one graph neural network (GNN); each of the machine learning models in the model sequence is utilized to perform a corresponding task in the supply chain network; all of the machine learning models in the model sequence are trained based on the same training dataset; and an output of one machine learning model in the model sequence serves as an input to a succeeding machine learning model in the model sequence” as recited in claim 7, and the corresponding claim language of the other dependent claims. The combination of features as claimed would not have been obvious to one of ordinary skill in the art because any combination of the evidence at hand to reach the combination of features as claimed would require a substantial reconstruction of Applicant’s claimed invention relying on improper hindsight bias. It is hereby asserted by the Examiner that, in light of the above and in further deliberation over all of the evidence at hand, that the claims are allowable over the previously cited art as set forth in the parent application as the evidence at hand does not anticipate the claims and does not render obvious any further modification of the references to a person of ordinary skill in the art.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
“A Machine Learning Approach for Predicting Hidden Links in Supply Chain with Graph Neural Networks” by Edward Kosasih et al. discusses the use of an automated method to detect potential links that are unknown to the buyer with Graph Neural Network (GNN). Using a real automotive network as a test case, the method provides a complimentary, AI based approach to improve supply chain visibility. Additionally, Integrated Gradient, a widely used technique in machine learning is used to improve the explainability of our approach by highlighting input features that influence GNN’s decisions, and a GNN for link prediction is used to improve supply chain visibility through the use of machine learning.
U.S. Patent Application Publication No. 20230129665 to Kumar et al. discusses a system to receive training data including, for each of a plurality of training timesteps, training forecast states associated with respective training-phase agents included in a training supply chain graph, train a reinforcement learning simulation of the training supply chain graph using the training data via policy gradient reinforcement learning, and receive runtime forecast states associated with respective runtime agents included in a runtime supply chain graph. For a runtime agent, at the trained reinforcement learning simulation, the processor may generate a respective runtime action output associated with a corresponding runtime forecast state of the runtime agent based at least in part on the runtime forecast states.
U.S. Patent Application Publication No. 20220318831 to Marvaniya et al. discusses obtaining a spatiotemporal query related to a demand of at least one product in a supply chain; analyzing the spatiotemporal query to identify one or more parameters affecting the demand of the at least one product, wherein the one or more parameters comprise at least one of one or more climate parameters and one or more disruptive event parameters; generating a knowledge graph comprising information indicating an impact on the demand of the at least one product for at least a portion of the one or more parameters; and outputting, to a user interface, an explanation of a predicted demand forecast for the at least one product based at least in part on the knowledge graph.
U.S. Patent Application Publication No. 20220318711 to Recasens et al. discusses receiving training data representing historic consumer demand for products, detecting changepoints in that data that may be associated with disruptive events, identifying relevant data for modeling, performing clustering, processing configuration information, training one or more machine learning models that are capable of evaluating other received data more accurately, and outputting results to a user display device.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Joshua D Schneider whose telephone number is (571)270-7120. The examiner can normally be reached on Monday - Friday, 9am-5pm.
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, Jessica Lemieux can be reached on (571)270-3445. 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.
/J.D.S./Examiner, Art Unit 3626
/JESSICA LEMIEUX/Supervisory Patent Examiner, Art Unit 3626