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 .
Detailed Action
The following action is in response to the communication(s) received on 01/23/2026.
As of the claims filed 01/23/2026:
Claims 1, 2, 10, 14, 17, and 20 have been amended.
Claim 22 has been canceled.
Claims 1-3, 5-12, 14-18, 20, 21, and 23 are pending.
Claims 1, 10, and 17 are independent claims.
Response to Arguments
Applicant’s arguments filed 10/01/2025 have been fully considered, but are not fully persuasive.
With respect to the rejection under 35 USC § 102 and 35 USC § 103:
Applicant asserts that Claim 1 has been amended to indicate novel subject matter with the amendment:
PNG
media_image1.png
500
642
media_image1.png
Greyscale
Examiner respectfully disagrees, as Bosselut [p.3 1st col 1st ¶]; [p.3 2nd col last ¶] (Note: encoding the absolute position of the sequence corresponds to using a node generation technique selected from the group consisting of a sequence-to-sequence paradigm.); and Fig. 1 (the generated edges (dashed lines) correspond to the interconnect of the plurality of nodes based on the node features; outputting information is a technique used by gate recurrent unit, thus COMET corresponds to using an edge generated technique selected from the group consisting of gated recurrent unit-based generation) teach these limitations.
The art rejections regarding Claims 10 and 17 are maintained for the same reasons above.
With respect to the rejection under 35 USC § 101, Applicant asserts the amendments to claim 1, specifically “in a first stage…; in a second stage…” directs the invention towards patent-eligible subject matter. Examiner respectfully disagrees. Similarly to the limitations recited in canceled claim 22, the two specific stages are merely a detail of an abstract idea (generating the plurality of nodes; generating the edges).
Applicant further asserts that the present claims are directed to an improvement to technology or technical field of automatically constructed knowledge graphs (p.10 ¶2-3). Examiner respectfully submits that the improvement is towards constructing knowledge graphs, which is an abstract idea and not a technology.
Applicant further asserts that “multi-stage methodology to construct a knowledge graph” reflects an improvement in technology as discussed in the Specification (p.11 ¶2). As stated above, these steps merely reflect an improvement in the abstract idea of generating a knowledge graph, not a particular technology.
At least for the reasons given above, the claims remain ineligible.
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 1-3, 5-12, 14-18, and 20-23 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Claim 1 recites A... method, thus a process, one of the four statutory categories of patentable subject matter (Step 1). However, Claim 1 further recites:
in a first stage, generating a plurality of nodes corresponding to entities of the input text…; uses the input of node queries to generate the plurality of nodes, which is an evaluation or judgement that can be performed in the human mind;
extracting node features from the plurality of nodes, which is an evaluation or judgement that can be performed in the human mind;
and in a second stage, generating edges based on the node features to interconnect the plurality of nodes… , which is an evaluation or judgement that can be performed in the human mind;
Thus, the claim recites an abstract idea under Step 2A Prong 1.
Under Step 2A Prong 2, the claim does not include any additional elements which integrate the abstract idea into a practical application, since the additional elements consist of:
A computer-implemented method…; …with a pretrained model using a node generation technique selected from the group consisting of a sequence-to-sequence paradigm and query vectors; …using an edge generation technique selected from the group consisting of gated recurrent unit-based generation and classifier-based edge generation, as the performance of an abstract idea on a computer is not more than instructions to "apply it" on a computer, which by MPEP 2106.05(f) cannot integrate an abstract idea into a practical application;
receiving input text in a natural language format, which is merely an insignificant extra-solution activity of data gathering, which by MPEP 2106.05(g) cannot integrate an abstract idea into a practical application.
receives an input of node queries, which is merely an insignificant extra-solution activity of data gathering, which by MPEP 2106.05(g) cannot integrate an abstract idea into a practical application.
Thus, the claim is directed towards an abstract idea.
Further, the additional elements, alone or in combination, do not provide significantly more than the abstract idea itself, because implementation on a computer (MPEP 2106.05(f)), and the activity of data gathering (MPEP 2106.05(g)) cannot provide significantly more, as storing and retrieving information in memory is well understood, routine, and conventional (MPEP 2106.05(d)(II)(iv)) and the combination of additional elements does not provide an inventive concept. Thus, the claim is ineligible.
Claim 2, dependent upon Claim 1, further recites no additional abstract ideas. However:
Under Step 2A Prong 2, the claim does not include any additional elements which integrate the abstract idea into a practical application, since the additional elements consist of:
the pretrained model is a pretrained language model, which are additional details of the pretrained model, in which the performance of an abstract idea on a computer is not more than instructions to "apply it" on a computer, which by MPEP 2106.05(f) cannot integrate an abstract idea into a practical application.
Thus, the claim is directed towards an abstract idea.
Further, under Step 2B, the additional element does not provide significantly more than the abstract idea itself, because implementation on a computer (MPEP 2106.05(f)) cannot provide significantly more. Thus, the claim is ineligible.
Claim 3, dependent upon Claim 2, further recites no additional abstract ideas. However:
Under Step 2A Prong 2, the claim does not include any additional elements which integrate the abstract idea into a practical application, since the additional elements consist of:
the pretrained language model includes an encoder and a decoder of transformer model, as the performance of an abstract idea on a computer is not more than instructions to "apply it" on a computer, which by MPEP 2106.05(f) cannot integrate an abstract idea into a practical application.
Thus, the claim is directed towards an abstract idea.
Further, under Step 2B, the additional element does not provide significantly more than the abstract idea itself, because implementation on a computer (MPEP 2106.05(f)) cannot provide significantly more. Thus, the claim is ineligible
Claim 5, dependent upon Claim 3, further recites
the edges are generated using the node features, which is an evaluation or judgement that can be performed in the human mind.
Thus, the claim recites an abstract idea under Step 2A Prong 1.
Under Step 2A Prong 2, the claim does not include any additional elements which integrate the abstract idea into a practical application, since the additional elements consist of:
the decoder receives input of learnable node queries, which is merely an insignificant extra-solution activity of data gathering, which by MPEP 2106.05(g) cannot integrate an abstract idea into a practical application;
the decoder directly outputs node features, which is merely an insignificant extra-solution activity of data output, which by MPEP 2106.05(g) cannot integrate an abstract idea into a practical application.
Thus, the claim is directed towards an abstract idea.
Further, the additional elements, alone or in combination, do not provide significantly more than the abstract idea itself, because and the activity of data gathering/output(MPEP 2106.05(g)) cannot provide significantly more, as storing and retrieving information in memory is well understood, routine, and conventional (MPEP 2106.05(d)(II)(iv)), and receiving or transmitting data over a network is well understood, routine, and conventional (MPEP 2106.05(d)(II)(i)) and the combination of additional elements does not provide an inventive concept. Thus, the claim is ineligible.
Claim 6, dependent upon Claim 1, further recites
the edges are generated using gated recurrent unit (GRU) techniques, which is an evaluation or judgement that can be performed in the human mind.
Thus, the claim recites an abstract idea under Step 2A Prong 1.
Under Step 2A Prong 2 and 2B, the claim does not recite any new additional elements which could integrate the abstract idea into a practical application or provide significantly more than the abstract idea itself. Thus, the claim is ineligible.
Claim 7, dependent upon Claim 1, further recites
the edges are generated using classification-based techniques, which is an evaluation or judgement that can be performed in the human mind.
Thus, the claim recites an abstract idea under Step 2A Prong 1.
Under Step 2A Prong 2 and 2B, the claim does not recite any new additional elements which could integrate the abstract idea into a practical application or provide significantly more than the abstract idea itself. Thus, the claim is ineligible.
Claim 8, dependent upon Claim 1, further recites
the edges are selected by down-weighting cross-entropy loss for well-classified samples and increasing cross-entropy for misclassified samples within the input text, which is an evaluation or judgement that can be performed in the human mind.
Thus, the claim recites an abstract idea under Step 2A Prong 1.
Under Step 2A Prong 2 and 2B, the claim does not recite any new additional elements which could integrate the abstract idea into a practical application or provide significantly more than the abstract idea itself. Thus, the claim is ineligible.
Claim 9, dependent upon Claim 1, further recites
sparsifying an adjacency matrix used to select the edges, which is an evaluation or judgement that can be performed in the human mind;
Thus, the claim recites an abstract idea under Step 2A Prong 1.
Under Step 2A Prong 2, the claim does not include any additional elements which integrate the abstract idea into a practical application, since the additional elements consist of:
language models are trained to generate the edges, the computer-implemented method further comprising training the language models, as the performance of an abstract idea on a computer is not more than instructions to "apply it" on a computer, which by MPEP 2106.05(f) cannot integrate an abstract idea into a practical application.
Thus, the claim is directed towards an abstract idea.
Further, under Step 2B, the additional element does not provide significantly more than the abstract idea itself, because implementation on a computer (MPEP 2106.05(f)) cannot provide significantly more. Thus, the claim is ineligible.
Claims 10-12 and 14-16 recite A system, thus a machine, one of the four statutory categories of patentable subject matter. However, Claims 10-16 recite comprising: a processor; and a memory in communication with the processor, the memory containing instructions that, when executed by the processor, cause the processor to perform precisely the abstract ideas and additional elements of Claims 1-7, respectively. Therefore, Step 2A Prong 1 analysis remains the same. As for Step 2A Prong 2 and Step 2B: performance on a computer cannot integrate an abstract idea into a practical application (Step 2A Prong 2) nor provide significantly more than the abstract idea itself (Step 2B) (MPEP 2106.05(f)), Claims 10-16 are rejected as subject-matter ineligible for reasons set forth in the rejections of Claims 1-7, respectively.
Claims 17 recites A computer program product, thus an article of manufacture, one of the four statutory categories of patentable subject matter. However, Claims 17 recites comprising: a processor; and a memory in communication with the processor, the memory containing instructions that, when executed by the processor, cause the processor to perform precisely the abstract ideas and additional elements of Claim 1. Therefore, Step 2A Prong 1 analysis remains the same. As for Step 2A Prong 2 and Step 2B: performance on a computer cannot integrate an abstract idea into a practical application (Step 2A Prong 2) nor provide significantly more than the abstract idea itself (Step 2B) (MPEP 2106.05(f)), Claims 17 is rejected as subject-matter ineligible for reasons set forth in the rejections of Claim 1.
Claim 18, dependent upon Claim 1, further recites no additional abstract ideas. However:
Under Step 2A Prong 2, the claim does not include any additional elements which integrate the abstract idea into a practical application, since the additional elements consist of:
the pretrained model is a pretrained language model that includes an encoder and a decoder of transformer model, as the performance of an abstract idea on a computer is not more than instructions to "apply it" on a computer, which by MPEP 2106.05(f) cannot integrate an abstract idea into a practical application.
Thus, the claim is directed towards an abstract idea.
Further, under Step 2B, the additional element does not provide significantly more than the abstract idea itself, because implementation on a computer (MPEP 2106.05(f)) cannot provide significantly more. Thus, the claim is ineligible
Claim 20, dependent on 17, recites the computer program product configured to perform precisely the methods of Claim 5, respectively. Thus, Claim 20 is ineligible for reasons set forth in Claim 5, respectively.
Claim 21, dependent upon Claim 1, further recites
pre-processing an input to obtain the input text, which is an evaluation or judgement that can be performed in the human mind.
Thus, the claim recites an abstract idea under Step 2A Prong 1.
Under Step 2A Prong 2 and 2B, the claim does not recite any new additional elements which could integrate the abstract idea into a practical application or provide significantly more than the abstract idea itself. Thus, the claim is ineligible.
Claim 22, dependent upon Claim 1, further recites
the generating the plurality of nodes occurs in a first stage and the generating the edges to interconnect the plurality of nodes occurs in a second stage, which is merely a detail of an abstract idea (generating the plurality of nodes; generating the edges).
Thus, the claim recites an abstract idea under Step 2A Prong 1.
Under Step 2A Prong 2 and 2B, the claim does not recite any new additional elements which could integrate the abstract idea into a practical application or provide significantly more than the abstract idea itself. Thus, the claim is ineligible.
Claim 23, dependent upon Claim 1, further recites
balancing the edges via one selected from the group consisting of a focal loss technique and sparsifying an adjacency matrix of the node features, which is an mathematical concept.
Thus, the claim recites an abstract idea under Step 2A Prong 1.
Under Step 2A Prong 2 and 2B, the claim does not recite any new additional elements which could integrate the abstract idea into a practical application or provide significantly more than the abstract idea itself. Thus, the claim is ineligible.
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1-3, 5-7, and 10-12, and 14-20 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Bosselut et al., “COMET : Commonsense Transformers for Automatic Knowledge Graph Construction” (hereinafter Bosselut)
Regarding Claim 1, Bosselut teaches:
A computer-implemented method using a multi-stage methodology to construct a knowledge graph comprising: (Bosselut [p.2 2nd col 1st ¶] Second, we develop a framework for using large-scale transformer language models to learn to produce commonsense knowledge tuples2
2Code is available at https://github.com/atcbosselut/comet-commonsense
[p.2 2nd col middle ¶] COMET learns to adapt the language model representations learned from pretraining to add novel nodes and edges to the seed knowledge graph.)) (Note: code available on GitHub corresponds to a method that is implemented by a computer; adding novel nodes and edges to the seed knowledge graph corresponds to a multi-stage methodology to construct a knowledge graph)
receiving input text in a natural language format; (Bosselut [p.4 1st col last ¶] COMET relies on a seed set of knowledge tuples from an existing KB to learn to produce commonsense knowledge. In this work, we use ATOMIC and ConceptNet as knowledge seed sets, but other commonsense knowledge resources could have been used as well as COMET is domain-agnostic.) (Note: knowledge seed set corresponds to the input text)
in a first stage, generating a plurality of nodes corresponding to entities of the input text…
(Bosselut
PNG
media_image2.png
514
469
media_image2.png
Greyscale
) (Note: the solid circles correspond to the first stage and generating a plurality of nodes corresponding to entities of the input text; the generated edges (dashed lines) correspond to the interconnect of the plurality of nodes.)
with a pretrained model…, (Bosselut [p.2 2nd col middle ¶] COMET learns to adapt the language model representations learned from pretraining to add novel nodes and edges to the seed knowledge graph.)
… using a node generation technique selected from the group consisting of a sequence-to-sequence paradigm and query vectors, wherein the pretrained model receives an input of node queries, and wherein the pretrained model uses the input of node queries…
(Bosselut [p.3 1st col 1st ¶] More specifically, the problem assumes COMET is given a training knowledge base of natural language tuples in {s, r, o} format, where s is the phrase subject of the tuple, r is the relation of the tuple, and o is the phrase object of the tuple. For example, a ConceptNet tuple relating to “taking a nap" would be: (s=“take a nap", r=Causes, o=“have energy"). The task is to generate o given s and r as inputs.
PNG
media_image3.png
504
913
media_image3.png
Greyscale
[p.3 2nd col last ¶] Since the transformer (a self-attention model) has no concept of ordering of tokens, a position embedding pt is initialized for each absolute position in the sequence. For any input word xt ∈ X, our encoding of the input is the sum of its word embedding, et with a position embedding encoding its absolute position in the sequence X: h 0 t = et + pt (10) where pt is the position embedding for time step t, and h 0 is the input to the first transformer layer.)) (Note: the square vocab layer corresponds to the node features directly outputted by the decoder (the adjacent row below the vocab layer; encoding the absolute position of the sequence corresponds to using a node generation technique selected from the group consisting of a sequence-to-sequence paradigm.)
…to generate the plurality of nodes; (Bosselut
PNG
media_image2.png
514
469
media_image2.png
Greyscale
) (Note: each circle corresponds to each node)
extracting node features from the plurality of nodes; (Bosselut
PNG
media_image2.png
514
469
media_image2.png
Greyscale
[p.2 left last ¶] We summarize our contributions in this work as follows. First, we develop a generative approach to knowledge base construction. A model must learn to produce new nodes and identify edges between existing nodes by generating phrases that coherently complete an existing seed phrase and relation type
[p.3 1st col 1st ¶] More specifically, the problem assumes COMET is given a training knowledge base of natural language tuples in {s, r, o} format, where s is the phrase subject of the tuple, r is the relation of the tuple, and o is the phrase object of the tuple. For example, a ConceptNet tuple relating to “taking a nap" would be: (s=“take a nap", r=Causes, o=“have energy"). The task is to generate o given s and r as inputs.
) (Note: the phrases correspond to the nodes; the relation r corresponds to the node features extracted from the plurality of nodes; the generated/identified edges correspond to the generated node features)
and in a second stage, generating edges based on the node features to interconnect the plurality of nodes using an edge generation technique selected from the group consisting of gated recurrent unit-based generation and classifier-based edge generation. (Bosselut
PNG
media_image2.png
514
469
media_image2.png
Greyscale
) (Note: the solid circles correspond to generating a plurality of nodes corresponding to entities of the input text; the generated edges (dashed lines) correspond to the interconnect of the plurality of nodes based on the node features; outputting information is a technique used by gate recurrent unit, thus COMET corresponds to using an edge generated technique selected from the group consisting of gated recurrent unit-based generation.)
Regarding Claim 2, Bosselut respectively teaches and incorporates the claimed limitations and rejections of Claim 1. Bosselut further teaches:
The computer-implemented method of claim 1, wherein the pretrained model is a pretrained language model. (Bosselut [p.2 2nd col middle ¶] COMET learns to adapt the language model representations learned from pretraining to add novel nodes and edges to the seed knowledge graph.)
Regarding Claim 3, Bosselut respectively teaches and incorporates the claimed limitations and rejections of Claim 2. Bosselut further teaches:
The computer-implemented method of claim 2, wherein the pretrained language model includes an encoder and a decoder of transformer model. (Bosselut [p.3 2nd col 2nd ¶] As shown in Figure 2(c), we follow Radford et al. (2018) and use the output of the previous layer’s transformer block as the query input for the multi-headed attention of the next block.
PNG
media_image3.png
504
913
media_image3.png
Greyscale
[p.3 2nd col last ¶] Since the transformer (a self-attention model) has no concept of ordering of tokens, a position embedding pt is initialized for each absolute position in the sequence. For any input word xt ∈ X, our encoding of the input is the sum of its word embedding, et with a position embedding encoding its absolute position in the sequence X: h 0 t = et + pt (10) where pt is the position embedding for time step t, and h 0 is the input to the first transformer layer.) (Note: the top square transformer block row corresponds to the decoder.)
Regarding Claim 5, Bosselut respectively teaches and incorporates the claimed limitations and rejections of Claim 3. Bosselut further teaches:
The computer-implemented method of claim 3, wherein: the decoder receives input of learnable node queries; and the decoder directly outputs node features; (Bosselut [p.3 1st col 1st ¶] More specifically, the problem assumes COMET is given a training knowledge base of natural language tuples in {s, r, o} format, where s is the phrase subject of the tuple, r is the relation of the tuple, and o is the phrase object of the tuple. For example, a ConceptNet tuple relating to “taking a nap" would be: (s=“take a nap", r=Causes, o=“have energy"). The task is to generate o given s and r as inputs.
PNG
media_image3.png
504
913
media_image3.png
Greyscale
) (Note: the square vocab layer corresponds to the node features directly outputted by the decoder (the adjacent row below the vocab layer)
Regarding Claim 6, Bosselut respectively teaches and incorporates the claimed limitations and rejections of Claim 1. Bosselut further teaches:
The computer-implemented method of claim 1, wherein the edges are generated using gated recurrent unit (GRU) techniques. (Bosselut
PNG
media_image3.png
504
913
media_image3.png
Greyscale
) (Note: outputting information is a technique used by GRUs, thus COMET corresponds to using GRU techniques.)
Regarding Claim 7, Bosselut respectively teaches and incorporates the claimed limitations and rejections of Claim 1. Bosselut further teaches:
The computer-implemented method of claim 1, wherein the edges are generated using classification-based techniques. (Bosselut [p.2 2nd col 1st¶] The results indicate that COMET is able to produce high quality tuples as human judges find that 77.5% of generated tuples for ATOMIC events and 91.7% of generated tuples for ConceptNet relations are correct.) (Note: producing high-quality tuples corresponds to a classification-based technique (producing a label))
Independent Claim 10 recites A system comprising: a processor; and a memory in communication with the processor, the memory containing instructions that, when executed by the processor, cause the processor (Bosselut [p.2 2nd col 1st ¶] Second, we develop a framework for using large-scale transformer language models to learn to produce commonsense knowledge tuples2
2Code is available at https://github.com/atcbosselut/comet-commonsense) (Note: code available on GitHub corresponds to a method that is implemented by a computer) to perform precisely the methods of Claim 1. Thus, Claim 10 is rejected for reasons set forth in Claim 1.
Claims 11, 12, and 14-16, dependent on 10, also recite the system configured to perform precisely the methods of Claims 2, 3, and 5-7, respectively. Thus, Claims 11-16 are rejected for reasons set forth in Claims 2, 3, and 5-7, respectively.
Independent Claim 17 recites a computer program product, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer (Bosselut [p.2 2nd col 1st ¶] Second, we develop a framework for using large-scale transformer language models to learn to produce commonsense knowledge tuples2
2Code is available at https://github.com/atcbosselut/comet-commonsense) (Note: code available on GitHub corresponds to a method that is implemented by a computer) to perform precisely the methods of Claim 1. Thus, Claim 17 is rejected for reasons set forth in Claim 1.
Regarding Claim 18, Bosselut respectively teaches and incorporates the claimed limitations and rejections of Claim 17. Bosselut further teaches:
The computer program product of claim 17, wherein the plurality of nodes is generated using a pretrained language model (Bosselut [p.2 2nd col middle ¶] COMET learns to adapt the language model representations learned from pretraining to add novel nodes and edges to the seed knowledge graph.)
that includes an encoder and a decoder of transformer model. (Bosselut [p.3 2nd col 2nd ¶] As shown in Figure 2(c), we follow Radford et al. (2018) and use the output of the previous layer’s transformer block as the query input for the multi-headed attention of the next block.
PNG
media_image3.png
504
913
media_image3.png
Greyscale
[p.3 2nd col last ¶] Since the transformer (a self-attention model) has no concept of ordering of tokens, a position embedding pt is initialized for each absolute position in the sequence. For any input word xt ∈ X, our encoding of the input is the sum of its word embedding, et with a position embedding encoding its absolute position in the sequence X: h 0 t = et + pt (10) where pt is the position embedding for time step t, and h 0 is the input to the first transformer layer.) (Note: the top square transformer block row corresponds to the decoder.)
Claims 19 and 20, dependent on 17, also recite the computer program product configured to perform precisely the methods of Claims 4 and 5, respectively. Thus, Claims 19 and 20 are rejected for reasons set forth in Claims 4 and 5, respectively.
Regarding Claim 21, Bosselut respectively teaches and incorporates the claimed limitations and rejections of Claim 1. Bosselut further teaches:
The computer-implemented method of claim 1, further comprising: pre-processing an input to obtain the input text (Bosselut [p.4 1st col last ¶] COMET relies on a seed set of knowledge tuples from an existing KB to learn to produce commonsense knowledge. In this work, we use ATOMIC and ConceptNet as knowledge seed sets, but other commonsense knowledge resources could have been used as well as COMET is domain-agnostic.) (Note: knowledge seed set from an existing KB corresponds to the preprocessed input for obtaining the input text)
Regarding Claim 22, Bosselut respectively teaches and incorporates the claimed limitations and rejections of Claim 1. Bosselut further teaches:
The computer-implemented method of claim 1, wherein the generating the plurality of nodes occurs in a first stage and the generating the edges to interconnect the plurality of nodes occurs in a second stage. (Bosselut
PNG
media_image2.png
514
469
media_image2.png
Greyscale
) (Note: the solid lines generated from the existing knowledge base corresponds to the first stage and the dashed lines generating the novel edges correspond to the second stage.)
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 8 is rejected under 35 U.S.C. 103 as being unpatentable over Bosselut in view of Lin et al, “Focal Loss for Dense Object Detection” (hereinafter Lin).
Regarding Claim 8, Bosselut respectively teaches and incorporates the claimed limitations and rejections of Claim 1. Bosselut does not teach, but Lin further teaches:
The computer-implemented method of claim 1, wherein the edges are selected by down-weighting cross-entropy loss for well-classified samples and increasing cross-entropy for misclassified samples within the input text. (Lin
PNG
media_image4.png
403
410
media_image4.png
Greyscale
) (Note: the loss value being higher (left side of graph; pt <= 0.5) corresponds to increasing the cross-entropy for misclassified samples; the reduced relative loss for well-classified examples (pt >0.5) corresponds to the down-weighting cross-entropy loss.)
Bosselut and Lin are analogous to the present invention because both are from the same field of endeavor of improving the performance of neural networks. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to apply Lin’s focal loss to Bosselut’s knowledge graph. The motivation would be to “[enable] training highly accurate dense object detectors in the presence of vast numbers of easy background examples.” (Lin [Fig.1]).
Claims 9 is rejected under 35 U.S.C. 103 as being unpatentable over Bosselut in view of Ye et al, “Sparse Graph Attention Networks” (hereinafter Ye).
Regarding Claim 9, Bosselut respectively teaches and incorporates the claimed limitations and rejections of Claim 1. Bosselut does not teach, but Ye further teaches:
The computer-implemented method of claim 1, wherein language models are trained to generate the edges, the computer-implemented method further comprising training the language models by sparsifying an adjacency matrix used to select the edges. (Ye [Abstract] In this paper, we propose Sparse Graph Attention Networks (SGATs) that learn sparse attention coefficients under an L0-norm regularization, and the learned sparse attentions are then used for all GNN layers, resulting in an edge sparsified graph.
PNG
media_image5.png
228
697
media_image5.png
Greyscale
)
Bosselut and Ye are analogous to the present invention because both are from the same field of endeavor of generating an informed graph. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to apply Ye’s sparsification method to Bosselut’s knowledge graph. The motivation would be to “identify noisy/task-irrelevant edges, and thus perform feature aggregation on most informative neighbors” (Ye [Abstract]).
Claims 23 is rejected under 35 U.S.C. 103 as being unpatentable over Bosselut/Lin, further in view of Ye.
Regarding Claim 23, Bosselut respectively teaches and incorporates the claimed limitations and rejections of Claim 1. Bosselut does not teach, but Lin further teaches:
The computer-implemented method of claim 1, further comprising: balancing the edges via one selected from the group consisting of a focal loss technique (Lin
PNG
media_image4.png
403
410
media_image4.png
Greyscale
) (Note: the loss value being higher (left side of graph; pt <= 0.5) corresponds to increasing the cross-entropy for misclassified samples; the reduced relative loss for well-classified examples (pt >0.5) corresponds to the focal loss technique.)
Bosselut and Lin are analogous to the present invention because both are from the same field of endeavor of improving the performance of neural networks. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to apply Lin’s focal loss to Bosselut’s knowledge graph. The motivation would be to “[enable] training highly accurate dense object detectors in the presence of vast numbers of easy background examples.” (Lin [Fig.1]).
Bosselut/Lin does not teach, but Ye further teaches:
and sparsifying an adjacency matrix of the node features. (Ye [Abstract] In this paper, we propose Sparse Graph Attention Networks (SGATs) that learn sparse attention coefficients under an L0-norm regularization, and the learned sparse attentions are then used for all GNN layers, resulting in an edge sparsified graph. [p.4 left 3rd ¶]
PNG
media_image6.png
571
358
media_image6.png
Greyscale
)
Bosselut/Lin and Ye are analogous to the present invention because both are from the same field of endeavor of generating an informed graph. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to apply Ye’s balancing method to Bosselut/Lin’s knowledge graph. The motivation would be to “identify noisy/task-irrelevant edges, and thus perform feature aggregation on most informative neighbors” (Ye [Abstract]).
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 JOSEP HAN whose telephone number is (703)756-1346. The examiner can normally be reached Mon-Fri 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, Kakali Chaki can be reached on (571) 272-3719. 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.H./Examiner, Art Unit 2122
/KAKALI CHAKI/Supervisory Patent Examiner, Art Unit 2122
Prosecution Insights