DETAILED ACTION
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 § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1-5, 8-12, 15-19, 22-26 is/are rejected under 35 U.S.C. 103 as being unpatentable over Amazon, What is RAG? in view of Zhang et al. (hereinafter Zhang), A survey on federated learning, further in view of Futia et al. (hereinafter Futia), On the Integration of Knowledge Graphs into Deep Learning Models for a More Comprehensible AI—Three Challenges for Future Research, further in view of Garcez et al. (hereinafter Garcez), The Connectionist Inductive Learning and Logic Programming System.
Regarding Claim 1, Amazon discloses a computing system for operationalizing generative artificial intelligence employing a distributed neuro-symbolic reasoning and action platform, the computing system comprising:
obtaining a plurality of input data, the input data comprising enterprise knowledge and expert knowledge ["specific domains or an organization's internal knowledge base" pg. 2; "external data. It can come from multiple data sources, such as a APIs, databases, or document repositories" pg. 3; "add vast external knowledge sources" pg. 4; "Modern enterprises store vast amounts of information like manuals, FAQs, research reports, customer service guides, and human resource document repositories across various systems" pg. 4];
processing the obtained plurality of input data using an embedding model to create a vectorized dataset ["embedding language models, converts data into numerical representations and stores it in a vector database" pg. 3];
training the machine learning model using the vectorized dataset as a training dataset via federated learning at the plurality of edge devices , wherein each of the plurality of edge devices trains a local instance of the machine learning model on a local knowledge base stored on that edge device, wherein the server aggregates model parameters received from the plurality of edge devices without centrally storing the local knowledge bases, and wherein the machine learning model is configured to learn representations of the vectorized dataset ["embedding language models, converts data into numerical representations and stores it in a vector database" pg. 3];
engineering a prompt ["prompt engineering techniques" pg. 3] using an output of the reasoning task;
submitting the prompt and a plurality of context elements to one or more large language models [#4 in the figure on pg. 4; “RAG model augments the user input (or prompts) by adding the relevant retrieved data in context.” pg. 3 ¶5];
curating one or more responses for the one or more large language models to create a single response or action [#5 in the figure on pg. 4]; and
transmitting the single response or action to a user device [#5 in the figure on pg. 4].
However, Amazon fails to explicitly disclose a plurality of edge devices in communication with a server, the server including one or more hardware processors configured for:
generating a workflow, using a distributed computational graph, for training a machine learning model;
training the machine learning model using the vectorized dataset as a training dataset via federated learning at the plurality of edge devices, wherein each of the plurality of edge devices trains a local instance of the machine learning model on a local knowledge base stored on that edge device, wherein the server aggregates model parameters received from the plurality of edge devices without centrally storing the local knowledge bases, and wherein the machine learning model is configured to learn representations of the vectorized dataset;
applying symbolic reasoning techniques to the symbolic representations to perform a reasoning task by passing the symbolic representations through data transformation nodes constituting the distributed computational graph.
Zhang discloses a plurality of edge devices in communication with a server, the server including one or more hardware processors [“multiple clients collaborate to solve machine learning problems, which is under the coordination of a central aggregator” Abstract; Fig. 1] configured for:
generating a workflow, using a distributed computational graph, for training a machine learning model [“schematic diagram of federated learning” Fig. 1];
training the machine learning model using the vectorized dataset as a training dataset via federated learning at the plurality of edge devices, wherein each of the plurality of edge devices trains a local instance of the machine learning model on a local knowledge base stored on that edge device, wherein the server aggregates model parameters received from the plurality of edge devices without centrally storing the local knowledge bases, and wherein the machine learning model is configured to learn representations [“With the application of federated learning, each device uses local data for local training, then uploads the model to the server for aggregation, and finally the server sends the model update to the participants to achieve the learning goal” Abstract; “In federated learning, compared with distributed machine learning, each worker node is the sole owner of its own data and a training participant of the model” §1.1 ¶4; Fig. 1] of the vectorized dataset;
applying symbolic reasoning techniques to the symbolic representations to perform a reasoning task by passing the symbolic representations through data transformation nodes constituting the distributed computational graph [“multiple clients collaborate to solve machine learning problems, which is under the coordination of a central aggregator” Abstract; Fig. 1].
It would have been obvious to one having ordinary skill in the art, having the teachings of Amazon and Zhang before him before the effective filing date of the claimed invention, to modify the system of Amazon to incorporate the federated learning of Zhang.
Given the advantage of data privacy and distributed computing for increased speed and efficiency, one having ordinary skill in the art would have been motivated to make this obvious modification.
However, Amazon fails to explicitly disclose mapping the learned representations to symbolic concepts or rules to create symbolic representations of the learned representations;
applying symbolic reasoning techniques to the symbolic representations to perform a reasoning task by passing the symbolic representations through data transformation nodes constituting the distributed computational graph;
engineering a prompt using an output of the reasoning task.
Futia discloses mapping the learned representations to symbolic concepts or rules to create symbolic representations of the learned representations [“predicted output of deep learning models can be mapped into entities of KGs [Knowledge Graphs] or concepts and relationships of ontologies (knowledge matching)” pg. 2 ¶2; Fig. 1];
applying symbolic reasoning techniques to the symbolic representations to perform a reasoning task by passing the symbolic representations [“KGs and ontologies are natively built to be queried and therefore they are able to provide answers to user requests (interactive explanations) and to provide a symbolic level to interpret the behaviour and the results of a deep learning model” pg. 2 ¶2; Fig. 1] through data transformation nodes constituting the distributed computational graph;
engineering a prompt using an output of the reasoning task [“provide answers” pg. 2 ¶2; Fig. 1].
It would have been obvious to one having ordinary skill in the art, having the teachings of Amazon, Zhang, and Futia before him before the effective filing date of the claimed invention, to modify the combination to incorporate the symbolic reasoning of Futia.
Given the advantage of explainable learning, one having ordinary skill in the art would have been motivated to make this obvious modification.
However, Amazon fails to explicitly disclose iteratively refining, by the machine learning model, the learned representations based on an output of the reasoning task.
Garcez discloses iteratively refining, by the machine learning model, the learned representations based on an output of the reasoning task [“The Connectionist Inductive Learning and Logic Programming (C-IL2P) system is a massively parallel computational model based on a feedforward artificial neural network that integrates inductive learning from examples and background knowledge, with deductive learning from Logic Programming.” §1 ¶8; “The connectionist inductive learning and logic programming system” Fig. 1; “the knowledge extracted can be more easily analyzed by an expert that decides if it should feed the system once more, closing the learning cycle” §1 ¶8].
It would have been obvious to one having ordinary skill in the art, having the teachings of Amazon, Zhang, Futia, and Gracez before him before the effective filing date of the claimed invention, to modify the combination to incorporate the neural-symbolic system using iterative refinement of Gracez.
Given the advantage of increased model accuracy, one having ordinary skill in the art would have been motivated to make this obvious modification.
Regarding Claim 2, Amazon, Zhang, Futia, and Gracez disclose the computing system of claim 1. Amazon further discloses wherein the plurality of context elements comprises retrieval-augmented generation (RAG) information [“Retrieval-Augmented Generation (RAG)” pg. 1].
Regarding claim 3, Amazon, Zhang, Futia, and Gracez disclose the computing system of claim 2. Amazon further discloses wherein the RAG information is obtained from a RAG marketplace [“How can AWS support your Retrieval-Augmented Generation requirements? Amazon Bedrock is a fully-managed service” pg. 4].
Regarding Claim 4, Amazon, Zhang, Futia, and Gracez disclose the computing system of claim 1. Amazon further discloses wherein the plurality of context elements comprises information associated with the user device [“RAG model augments the user input (or prompts) by adding the relevant retrieved data in context.” pg. 3 ¶5]
Regarding Claim 5, Amazon, Zhang, Futia, and Gracez disclose the computing system of claim 1. Amazon further discloses wherein the plurality of context elements comprises information associated with an action a user of the user device is performing during interaction with the platform [“user query” pg. 3 ¶2; “If an employee searches, "How much annual leave do I have?"” pg. 3 ¶4].
Claims 8-12 are rejected on the same grounds as claims 1-5, respectively.
Claims 15-19 are rejected on the same grounds as claims 1-5, respectively.
Claims 22-26 are rejected on the same grounds as claims 1-5, respectively.
Claim(s) 6, 13, 20, 27 is/are rejected under 35 U.S.C. 103 as being unpatentable over Amazon, Zhang, Futia, and Gracez, in view of Lucas et al. (hereinafter Lucas), U.S. Patent Application Publication 2020/0176098.
Regarding Claim 6, Amazon, Zhang, Futia, and Gracez disclose the computing system of claim 1.
However, Amazon fails to explicitly disclose wherein the expert knowledge comprises scored datasets or scored model output.
Lucas discloses wherein the expert knowledge comprises scored datasets or scored model output [“Some MLA may identify features of importance and identify a coefficient, or weight, to them. The coefficient may be multiplied with the occurrence frequency of the feature to generate a score, and once the scores of one or more features exceed a threshold, certain classifications may be predicted by the MLA.” ¶63; “individual words may be provided a weighting factor for probability of occurrence across a massive training set” ¶103].
It would have been obvious to one having ordinary skill in the art, having the teachings of Amazon, Zhang, Futia, Gracez, and Lucas before him before the effective filing date of the claimed invention, to modify the combination to incorporate the feature weighting of Lucas.
Given the advantage of weighting features for a more accurate outcome, one having ordinary skill in the art would have been motivated to make this obvious modification.
Claims 13, 20 and 27 are each rejected by claim 6.
Claim(s) 7, 14, 21, 28 is/are rejected under 35 U.S.C. 103 as being unpatentable over Amazon, Zhang, Futia, Gracez, and Lucas, in view of Hall, Knowledge Marketplaces: The New Learning Hubs For Business Leaders.
Regarding Claim 7, Amazon, Zhang, Futia, Gracez, and Lucas disclose the computing system of claim 6.
However, Amazon fails to explicitly disclose wherein the expert knowledge is obtained from an expert knowledge marketplace.
Hall discloses wherein the expert knowledge is obtained from an expert knowledge marketplace [“a concept known as knowledge marketplaces, companies are building networks of market experts to help leaders in virtually every area of their business” pg. 2].
It would have been obvious to one having ordinary skill in the art, having the teachings of Amazon, Zhang, Futia, Gracez, Lucas and Hall before him before the effective filing date of the claimed invention, to modify the combination to incorporate the knowledge marketplace of Hall.
Given the advantage of acquiring expertise and insights for more accurate results, one having ordinary skill in the art would have been motivated to make this obvious modification.
Claim 14 is rejected on the same grounds as claim 7.
Claim 21 is rejected on the same grounds as claim 7.
Claim 28 is rejected on the same grounds as claim 7.
Examiner’s Note
The Examiner respectfully requests of the Applicant in preparing responses, to fully consider the entirety of the reference(s) as potentially teaching all or part of the claimed invention. It is noted, REFERENCES ARE RELEVANT AS PRIOR ART FOR ALL THEY CONTAIN. “The use of patents as references is not limited to what the patentees describe as their own inventions or to the problems with which they are concerned. They are part of the literature of the art, relevant for all they contain.” In re Heck, 699 F.2d 1331, 1332-33, 216 USPQ 1038, 1039 (Fed. Cir. 1983) (quoting In re Lemelson, 397 F.2d 1006, 1009, 158 USPQ 275, 277 (CCPA 1968)). A reference may be relied upon for all that it would have reasonably suggested to one having ordinary skill in the art, including non-preferred embodiments (see MPEP 2123). The Examiner has cited particular locations in the reference(s) as applied to the claim(s) above for the convenience of the Applicant. Although the specified citations are representative of the teachings of the art and are applied to the specific limitations within the individual claim(s), typically other passages and figures will apply as well.
Additionally, any claim amendments for any reason should include remarks indicating clear support in the originally filed specification.
Response to Arguments
Regarding the 101 rejection, the rejection is withdrawn due to amendments and Applicant’s arguments. Examiner agrees a practical application is found in the current claims.
Regarding the 103 rejections, Applicant's arguments have been fully considered but have been found unpersuasive. Applicant argues that 1) Zhang fails to disclose through data transformation nodes constituting the distributed computational graph, 2) the combination of Zhang and Futia does not arrive at the limitation based on viewing the individual components separately, 3) Futia fails to disclose applying symbolic reasoning techniques to the symbolic representations to perform a reasoning task by passing the symbolic representations in a process that utilizes data transformation nodes of a graph of any kind, and 4) none of the cited references disclose the new limitation of iteratively refining, by the machine learning model, the learned representations based on an output of the reasoning task. Examiner disagrees for at least the following reasons.
First, Zhang discloses a distributed computation graph in, for example, Figure 1. The schematic diagram is a distributed computation graph as it shows the needed computations in a distributed manner in graph form. Additionally, the remote devices used for the local computation are the data transformation nodes.
Second, in response to Applicant's arguments against the references individually, one cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986). Zhang disclose the federated learning aspects, while Futia discloses the symbolic reasoning. It is the combination of the references which disclose to a person having ordinary skill in the art the idea of using federated learning of Zhang for the symbolic reasoning of Futia.
Third, again Applicant is arguing the references in isolation. The data transformation nodes of a distributed computational graph are disclosed by Zhang. It is the combination of the two references which disclose the limitation.
Fourth, Applicant's arguments with respect to the claims have been considered but are moot because the arguments do not apply to the references being used in the current rejection of the limitations.
For at least these reasons, the prior art rejections are maintained.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
The following references were found during the examination of this patent application and were found to be relevant to patentability. Applicant is advised to review these references prior to responding to this Office action.
Yu et al. (A survey on neural-symbolic learning systems) discloses combining neural systems and symbolic systems to create neural-symbolic learning systems that possess powerful perception and cognition.
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ROBERT H BEJCEK II whose telephone number is (571)270-3610. The examiner can normally be reached Monday - Friday: 9:00am - 5:00pm.
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 T. 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.
/R.B./ Examiner, Art Unit 2148
/MICHELLE T BECHTOLD/ Supervisory Patent Examiner, Art Unit 2148
