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 .
Status of Claims
Claims 1-8 remain pending. Claims 1 and 3-8 have been amended.
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 1-4, and 6-8 are rejected under 35 U.S.C. 103 as being unpatentable over Cao et al “Robot Behavior-Tree-Based Task Generation with Large Language Models” view of Sokabe et al (US 20260042206 A1) and Lee et al “Behavior-Tree Embeddings for Robot Task-Level Knowledge” (Hereinafter referred to as Cao, Sokabe, and Lee respectively)
Regarding Claim 1, Cao teaches a method for generating a behavior tree for controlling a robot device (See at least Cao Page 1 Abstract), the method comprising the following steps:
combining a plurality of predetermined behavior trees and background knowledge into a behavior tree knowledge graph by including the plurality of predetermined behavior trees as sub-graphs in the behavior tree knowledge graph (See at least Cao Page 3 second paragraph, Pages 5-6 section 3.1, and Pages 7-8 section 3.3., the knowledge base/graph has a plurality of predetermined behavior trees, and the behavior trees have a description (e.g. Assembly car, Make coffee), which is interpreted as background knowledge);
representing the behavior tree knowledge graph in a latent space by a knowledge graph embedding method in a form of a plurality of clusters of latent space elements (See at least Cao Pages 7-8 section 3.3., the behavior tree embedding produces vectors in vector space, which is interpreted as representing the behavior trees in a latent space in the form of clusters), wherein each cluster of the plurality of clusters represents a sub-graph of the behavior tree knowledge graph (See at least Cao Pages 7-8 section 3.3., each vector/cluster represents a graph/behavior tree of the knowledge base/graph)…;
extracting, from a prompt describing a desired behavior of the robot device, a prompt representation graph representing terms in the prompt as nodes and relations between the terms indicated in the prompt as edges (See at least Cao Pages 4-7, Sections 2.2. 3, 3.1, and 3.2, the behavior tree constructed from the prompt is interpreted as a prompt representation graph);
supplementing the prompt representation graph with one or more nodes and/or one or more edges according to relations, specified by the behavior tree knowledge graph, between multiple terms indicated in the prompt and/or one or more terms indicated in the prompt and other terms represented in the behavior tree knowledge graph (See at least Cao Pages 6-7, Section 3.2, and Pages 9-10 tables 3-4, the behavior tree/prompt representation graph is constructed using verbs from the verb list in the knowledge base/graph, which supplements the verbs from the prompt based on similarity);
representing the supplemented prompt representation graph as an additional cluster in the latent space by the knowledge graph embedding method (See at least Cao Pages 7-8 section 3.3, the supplemented prompt representation graph/behavior tree is represented as a vector/cluster in latent/vector space by the behavior tree embedding process)…;
determining similarities of the additional cluster with the clusters of the plurality of clusters (See at least Cao Pages 7-8 section 3.3, the supplemented prompt representation graph/behavior tree is represented as a vector/cluster in latent/vector space by the behavior tree embedding process);
selecting one of the sub-graphs of the behavior tree knowledge graph depending on the similarity of the cluster representing the sub-graph with the additional cluster (See at least Cao Pages 7-8 section 3.3, the vector/cluster that is most similar is selected);
generating the behavior tree for controlling the robot device by adjusting the selected sub-graph according to knowledge from the prompt and the behavior tree knowledge graph (See at least Cao Pages 7-8 section 3.3, and Pages 8-11 section 4.2, the behavior tree is generated by adjusting the selected graph/tree according to the prompt and knowledge base/graph), wherein the adjusting includes identifying missing information in the selected sub-graph with respect to the prompt and adjusting the selected sub-graph to include the missing information from the behavior tree knowledge graph or replace information in the sub-graph with the missing information from the prompt or the behavior tree knowledge graph (See at least Cao Pages 8-11 section 4.2, the information in the selected sub-graph regarding the wheel assembly task is replaced with missing information from the prompt regarding the desktop assembly task).
Cao fails to explicitly disclose controlling the robot device according to the generated behavior tree.
However, Sokabe teaches controlling the robot device according to the generated behavior tree (See at least Sokabe Paragraph 0074, the robot is controlled according to the generated behavior tree).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify the teachings disclosed in Cao with Sokabe to control the robot device according to the generated behavior tree. This modification, as taught by Sokabe, would allow the robot to execute the tasks in the generated behavior tree, thus, improving robot control.
Even though Cao teaches representing the behavior tree knowledge graph and supplemented prompt representation graph in latent space, modified Cao fails to explicitly disclose each node of the sub-graph comprises a latent space element that is an embedding of the node in the latent space; and
forming the additional cluster to comprise, for each node of the supplemented prompt representation graph, a latent space element that is an embedding of the node in the latent space.
However, Lee teaches each node comprises a latent space element that is an embedding of the node in the latent space (See at least Lee Page 12076 Section B World-Embedding-Based Approach, the nodes for the behavior tree are embedded to obtain vectors, which is interpreted as a latent space element).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify the teachings disclosed in modified Cao with Lee to comprise a latent space element that is an embedding of the node in the latent space for the sub-graph and the supplemented prompt representation graph. This modification, as taught by Lee, would capture the rich semantic information of Action nodes, allows a node to capture the structural information within its k-hop neighbors after the kth message passing, and is well-suited for the modularity of a behavior tree (See at least Lee Page 12076 Section B World-Embedding-Based Approach).
Regarding Claim 2, modified Cao teaches the extracting of the prompt representation graph includes using a machine learning model configured to operate on natural language text which is supplied with the prompt as input (See at least Cao Pages 4-5, Sections 2.2. and 3, the GPT-model, which is a machine learning model, operates on the natural language text input).
Regarding Claim 3, modified Cao teaches the selecting of the sub-graph includes selecting that sub-graph of the behavior tree knowledge graph whose cluster by which it is represented has a highest similarity with the additional cluster (See at least Cao Pages 7-8 section 3.3, the vector/cluster with the highest similarity is selected).
Regarding Claim 4, modified Cao teaches generating the prompt such that the prompt representation graph includes one or more nodes and/or one or more edges for terms and/or relations not included in the behavior tree knowledge graph (See at least Cao Pages 6-7, Section 3.2, and Pages 9-10 tables 3-4, there are terms in the prompt that are not in the verb list included in the knowledge base/graph).
Regarding Claim 6, Cao teaches a method for controlling a robot device (See at least Cao Page 1 Abstract), comprising:
describing a desired behavior of the robot device in a text prompt (See at least Cao Pages 4-6, Sections 2.2. 3, and 3.1, the desired behavior of assembling a computer is described in a text prompt);
generating a behavior tree for controlling a robot device including:
combining a plurality of predetermined behavior trees and background knowledge into a behavior tree knowledge graph by including the plurality of predetermined behavior trees as sub-graphs in the behavior tree knowledge graph (See at least Cao Page 3 second paragraph, Pages 5-6 section 3.1, and Pages 7-8 section 3.3., the knowledge base/graph has a plurality of predetermined behavior trees, and the behavior trees have a description (e.g. Assembly car, Make coffee), which is interpreted as background knowledge),
representing the behavior tree knowledge graph in a latent space by a knowledge graph embedding method in a form of a plurality of clusters of latent space elements (See at least Cao Pages 7-8 section 3.3., the behavior tree embedding produces vectors in vector space, which is interpreted as representing the behavior trees as clusters in a latent space), wherein each cluster of the plurality of clusters represents a sub-graph of the behavior tree knowledge graph (See at least Cao Pages 7-8 section 3.3., each vector/cluster represents a graph/behavior tree of the knowledge base/graph),
extracting, from the text prompt, a prompt representation graph representing terms in the prompt as nodes and relations between the terms indicated in the prompt as edges (See at least Cao Pages 4-7, Sections 2.2. 3, 3.1, and 3.2, the behavior tree constructed from the prompt is interpreted as a prompt representation graph),
supplementing the prompt representation graph with one or more nodes and/or one or more edges according to relations, specified by the behavior tree knowledge graph, between multiple terms indicated in the prompt and/or one or more terms indicated in the prompt and other terms represented in the behavior tree knowledge graph (See at least Cao Pages 6-7, Section 3.2, and Pages 9-10 tables 3-4, the behavior tree/prompt representation graph is constructed using verbs from the verb list in the knowledge base/graph, which supplements the verbs from the prompt based on similarity),
representing the supplemented prompt representation graph as an additional cluster in the latent space by the knowledge graph embedding method (See at least Cao Pages 7-8 section 3.3, the supplemented prompt representation graph/behavior tree is represented as a vector/cluster in latent/vector space by the behavior tree embedding process),
determining similarities of the additional cluster with the clusters of the plurality of clusters (See at least Cao Pages 7-8 section 3.3, the similarity between vectors/clusters is determined),
selecting one of the sub-graphs of the behavior tree knowledge graph depending on the similarity of the cluster representing the sub-graph with the additional cluster (See at least Cao Pages 7-8 section 3.3, the vector/cluster that is most similar is selected), and
generating the behavior tree for controlling the robot device by adjusting the selected sub-graph according to knowledge from the prompt and the behavior tree knowledge graph (See at least Cao Pages 7-8 section 3.3, and Pages 8-11 section 4.2, the behavior tree is generated by adjusting the selected graph/tree according to the prompt and knowledge base/graph), wherein the adjusting includes identifying missing information in the selected sub-graph with respect to the prompt and adjusting the selected sub-graph to include the missing information from the behavior tree knowledge graph or replace information in the sub-graph with the missing information from the prompt or the behavior tree knowledge graph (See at least Cao Pages 8-11 section 4.2, the information in the selected sub-graph regarding the wheel assembly task is replaced with missing information from the prompt regarding the desktop assembly task).
Cao fails to explicitly disclose controlling the robot device according to the generated behavior tree.
However, Sokabe teaches controlling the robot device according to the generated behavior tree (See at least Sokabe Paragraph 0074, the robot is controlled according to the generated behavior tree).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify the teachings disclosed in Cao with Sokabe to control the robot device according to the generated behavior tree. This modification, as taught by Sokabe, would allow the robot to execute the tasks in the generated behavior tree, thus, improving robot control.
Even though Cao teaches representing the behavior tree knowledge graph and supplemented prompt representation graph in latent space, modified Cao fails to explicitly disclose each node of the sub-graph comprises a latent space element that is an embedding of the node in the latent space; and
forming the additional cluster to comprise, for each node of the supplemented prompt representation graph, a latent space element that is an embedding of the node in the latent space.
However, Lee teaches each node comprises a latent space element that is an embedding of the node in the latent space (See at least Lee Page 12076 Section B World-Embedding-Based Approach, the nodes for the behavior tree are embedded to obtain vectors, which is interpreted as a latent space element).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify the teachings disclosed in modified Cao with Lee to comprise a latent space element that is an embedding of the node in the latent space for the sub-graph and the supplemented prompt representation graph. This modification, as taught by Lee, would capture the rich semantic information of Action nodes, allows a node to capture the structural information within its k-hop neighbors after the kth message passing, and is well-suited for the modularity of a behavior tree (See at least Lee Page 12076 Section B World-Embedding-Based Approach).
Regarding Claims 7-8, Cao teaches …generating a behavior tree for controlling a robot device… (See at least Cao Page 1 Abstract), perform the following steps:
combining a plurality of predetermined behavior trees and background knowledge into a behavior tree knowledge graph by including the plurality of predetermined behavior trees as sub-graphs in the behavior tree knowledge graph (See at least Cao Page 3 second paragraph, Pages 5-6 section 3.1, and Pages 7-8 section 3.3., the knowledge base/graph has a plurality of predetermined behavior trees, and the behavior trees have a description (e.g. Assembly car, Make coffee), which is interpreted as background knowledge);
representing the behavior tree knowledge graph in a latent space by a knowledge graph embedding method in a form of a plurality of clusters of latent space elements (See at least Cao Pages 7-8 section 3.3., the behavior tree embedding produces vectors in vector space, which is interpreted as representing the behavior trees in a latent space in the form of clusters), wherein each cluster of the plurality of clusters represents a sub-graph of the behavior tree knowledge graph (See at least Cao Pages 7-8 section 3.3., each vector/cluster represents a graph/behavior tree of the knowledge base/graph);
extracting, from a prompt describing a desired behavior of the robot device, a prompt representation graph representing terms in the prompt as nodes and relations between the terms indicated in the prompt as edges (See at least Cao Pages 4-7, Sections 2.2. 3, 3.1, and 3.2, the behavior tree constructed from the prompt is interpreted as a prompt representation graph);
supplementing the prompt representation graph with one or more nodes and/or one or more edges according to relations, specified by the behavior tree knowledge graph, between multiple terms indicated in the prompt and/or one or more terms indicated in the prompt and other terms represented in the behavior tree knowledge graph (See at least Cao Pages 6-7, Section 3.2, and Pages 9-10 tables 3-4, the behavior tree/prompt representation graph is constructed using verbs from the verb list in the knowledge base/graph, which supplements the verbs from the prompt based on similarity);
representing the supplemented prompt representation graph as an additional cluster in the latent space by the knowledge graph embedding method (See at least Cao Pages 7-8 section 3.3, the supplemented prompt representation graph/behavior tree is represented as a vector/cluster in latent/vector space by the behavior tree embedding process);
determining similarities of the additional cluster with the clusters of the plurality of clusters (See at least Cao Pages 7-8 section 3.3, the supplemented prompt representation graph/behavior tree is represented as a vector/cluster in latent/vector space by the behavior tree embedding process);
selecting one of the sub-graphs of the behavior tree knowledge graph depending on the similarity of the cluster representing the sub-graph with the additional cluster (See at least Cao Pages 7-8 section 3.3, the vector/cluster that is most similar is selected);
generating the behavior tree for controlling the robot device by adjusting the selected sub-graph according to knowledge from the prompt and the behavior tree knowledge graph (See at least Cao Pages 7-8 section 3.3, and Pages 8-11 section 4.2, the behavior tree is generated by adjusting the selected graph/tree according to the prompt and knowledge base/graph), wherein the adjusting includes identifying missing information in the selected sub-graph with respect to the prompt and adjusting the selected sub-graph to include the missing information from the behavior tree knowledge graph or replace information in the sub-graph with the missing information from the prompt or the behavior tree knowledge graph (See at least Cao Pages 8-11 section 4.2, the information in the selected sub-graph regarding the wheel assembly task is replaced with missing information from the prompt regarding the desktop assembly task).
Cao fails to explicitly disclose a data processing device configured to generate a behavior tree for controlling a robot device,
a non-transitory computer-readable medium on which are stored instructions generating a behavior tree for controlling a robot device, the instructions are executed by a computer, and
controlling the robot device according to the generated behavior tree.
However, Sokabe teaches a data processing device configured to generate a behavior tree for controlling a robot device (See at least Sokabe Paragraphs 0040-0041 and 0074, the computer for generating a behavior tree is interpreted as the data processing device),
a non-transitory computer-readable medium on which are stored instructions generating a behavior tree for controlling a robot device, the instructions are executed by a computer (See at least Sokabe Paragraphs 0040-0041, 0046-0047, and 0074, the computer executes a program/instructions stored in a non-transitory recording medium to generate a behavior tree), and
controlling the robot device according to the generated behavior tree (See at least Sokabe Paragraph 0074, the robot is controlled according to the generated behavior tree).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify the teachings disclosed in Cao with Sokabe to have a data processing device configured to generate a behavior tree for controlling a robot device, a non-transitory computer-readable medium on which are stored instructions that are executed by a computer for generating a behavior tree for controlling a robot device, and controlling the robot device according to the generated behavior tree. This modification, as taught by Sokabe, would allow the system in Cao to autonomously perform the behavior tree generation for controlling the robot by executing the instructions/programs stored in the memory by the computer/data processing device (See at least Sokabe Paragraphs 0040-0041, 0046-0047, and 0074), which is routine and well-understood in the art, and allow the robot to execute the tasks in the generated behavior tree, thus, improving robot control.
Even though Cao teaches representing the behavior tree knowledge graph and supplemented prompt representation graph in latent space, modified Cao fails to explicitly disclose each node of the sub-graph comprises a latent space element that is an embedding of the node in the latent space; and
forming the additional cluster to comprise, for each node of the supplemented prompt representation graph, a latent space element that is an embedding of the node in the latent space.
However, Lee teaches each node comprises a latent space element that is an embedding of the node in the latent space (See at least Lee Page 12076 Section B World-Embedding-Based Approach, the nodes for the behavior tree are embedded to obtain vectors, which is interpreted as a latent space element).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify the teachings disclosed in modified Cao with Lee to comprise a latent space element that is an embedding of the node in the latent space for the sub-graph and the supplemented prompt representation graph. This modification, as taught by Lee, would capture the rich semantic information of Action nodes, allows a node to capture the structural information within its k-hop neighbors after the kth message passing, and is well-suited for the modularity of a behavior tree (See at least Lee Page 12076 Section B World-Embedding-Based Approach).
Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Cao in view of Sokabe and Lee, and in further view of Zhang et al (US 20240265275 A1) (Hereinafter referred to as Zhang)
Regarding Claim 5, Cao fails to explicitly disclose combining the generated behavior tree with the behavior tree knowledge graph.
However, Zhang teaches combining the generated behavior tree with the behavior tree knowledge graph (See at least Zhang Paragraphs 0038, 0080-0081, and 0100, the generated behavior tree is constructed into a knowledge graph instance for storage in the knowledge graph library).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify the teachings disclosed in modified Cao with Zhang to combine the generated behavior tree with the behavior tree knowledge graph. This modification, as taught by Zhang, would allow the system to store and utilize the generated behavior tree to recommend control nodes and/or action nodes, perform error checks, and intelligently map node properties (See at least Zhang Paragraphs 0080-0081, 0100, and 0116).
Response to Arguments
Applicant's arguments filed 08/05/2026 have been fully considered but they are not persuasive. Applicant argues, on page 9 of the remarks, that Cao does not disclose “wherein the adjusting includes identifying missing information in the selected sub-graph with respect to the prompt and adjusting the selected sub-graph to include the missing information from the behavior tree knowledge graph or replace information in the sub-graph with the missing information from the prompt or the behavior tree knowledge graph”. Applicant states “insofar as these portions of Cao describe that a source behavior tree and task description are fed into an LLM to generate a complete target behavior tree, they do not disclose the language of amended claim 1”.
However, Examiner disagrees. Cao teaches, in Section 3.3., that a sub-graph that is most similar to the additional cluster is selected. The selected sub-graph is then adjusted according to the knowledge from the prompt. In section 4.2, the selected sub-graph is the wheel-assembly task. The wheel-assembly task is a behavior tree within the knowledge base. The wheel-assembly task is adjusted according to the “Desktop assembly” prompt. The wheel-assembly task is adjusted by replacing information in the sub-graph with information from the prompt “Desktop assembly”. The information in the “Desktop assembly” includes information that is missing in the selected sub-graph, such as attaching the CPU. Figure 5 in Cao shows how the selected sub-graph is replaced with missing information from the prompt.
PNG
media_image1.png
504
998
media_image1.png
Greyscale
Thus, for these reasons, the claims still stand rejected over the prior art.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ESVINDER SINGH whose telephone number is (571)272-7875. The examiner can normally be reached Monday-Friday: 9 am-5 pm 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, Abby Lin can be reached at 571-270-3976. 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.
/ESVINDER SINGH/Primary Examiner, Art Unit 3657
Prosecution Insights