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
2. This office action is in response to the amendment filed on 06/18/2026. Claims 1-20 are pending and have been considered below.
3. The rejections of Claims 11-13 under 35 U.S.C. 101 as directed to non-statutory subject matter (i.e computer program) are moot pursuant to amendments.
Claim Rejections - 35 USC § 112
4. The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 4-6 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 4 depends on claim 1 and recites, in relevant part, “wherein the generating includes generating an aligned knowledge graph…”. However, claim 1 recites two separate “generating” clauses, each directed to a different operation. Accordingly, it is unclear which of the two generating operations recited in claim 1 is being referenced by the “generating” limitation of claim 4.
Specifically, claim 1 first recites “generating, by a processing device, training data using a machine-learning model..” and subsequently recites “generating , by the processing device, a trained machine-learning model using...” Claim 4 does not identify which of these two generating operations is intended to further limit the claimed subject matter. Applicant is invited to amend claim 4 to expressly identify the particular generating operation of claim 1 to which the limitation refers.
Therefore, claim 4 as well as dependent claims 5-6 are also rejected
5. The rejections of claims 1-20 under 35 U.S.C. 101 as directed to abstract ideas without significantly more are moot pursuant to claims amendments and arguments.
Claim Rejections - 35 USC § 103
6. 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 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.
7. Claims 1-10 and 14-20 are rejected under 35 U.S.C. 103 as being unpatentable over Iida et al. (TABBIE: Pretrained Representations of Tabular Data May 2021) in view of Wang et al. (K-ADAPTER: Infusing Knowledge into Pre-Trained Models with Adapters August 2021) and further in view of Kim (US 2022/0284309).
Claim 1. Iida discloses a method comprising: generating, by a processing device, training data using a machine-learning model, the training data formed based (Introduction, fig. 1) based on: a tabular data corpus having a plurality of items of tabular data (fig. 1); and generating, by the processing device, a trained machine-learning model using machine learning based on the training data, the trained machine-learning model configured to generate a tabular-data type prediction based on a subsequent item of tabular data (TABBIE architecture relies on two Transformers that independently encode rows and columns, respectively; their representations are pooled at each layer. This setup reduces the sequence length of each Transformer’s input, which cuts down on its complexity, while also allowing us to easily extract representations of cells, rows, and columns… TABBIE uses a simplified training objective compared to masked language modeling: instead of predicting masked cells, we repurpose ELECTRA’s objective function for tabular pretraining by asking the model to predict whether or not each cell in a table is real or corrupted) (Introduction: col. 2, ll 19 thru p. 2, col. 1; Sections 3.1-3.3).
Iida does not explicitly disclose a knowledge graph. However, Wang discloses (..KADAPTER,
a framework that retains the original parameters of the pre-trained model fixed and supports the development of versatile knowledge-infused model) (abstract). Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iida further in view of Wang to incorporate the above cited features. One would have been motivated to do so in order to improve semantic understanding and model performance.
Iida does not explicitly disclose an aligned knowledge graph generated based, at least in part, on aligning the tabular data corpus with the knowledge graph; and annotations to the aligned knowledge graph.
However, Kim discloses an aligned knowledge graph generated based, at least in part, on aligning the tabular data corpus with the knowledge graph; and annotations to the aligned knowledge graph (abstract, [0045]-[0049]-[0051], figs 5-9, and claims 1 and 15). Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iida further in view of Kim to incorporate the above cited features. One would have been motivated to do so to provide additional semantic information to a tabular-data representation model.
Claim 2. Iida Wang and Kim disclose the method as described in claim 1, Iida further discloses wherein the tabular-data type prediction includes a column type prediction, a relation prediction, an outlier cell prediction, a table classification, column-based embedding retrieval, or entity-based embedding retrieval (Table 1, Fig. 4).
Claim 3. Iida Wang and Kim disclose the method as described in claim 1, Iida further discloses wherein: the tabular data includes a plurality of values arranged along a plurality of axes, respectively (Figs. 1-4); and Wang further discloses the knowledge graph includes a plurality of nodes representative of entities and a plurality of connections between the plurality of nodes representative of respective concepts (Table 1, p. 1406, Related Work). One would have been motivated to do so to improve semantic understanding and model performance.
Claim 4. Iida Wang and Kim disclose the method as described in claim 1, Kim further discloses wherein the generating includes generating an aligned knowledge graph by aligning the tabular data corpus with the knowledge graph and generating samples by selecting entities from the aligned knowledge graph (abstract, [0045]-[0049]-[0051], figs 5-9, and claims 1 and 15). One would have been motivated to do so to provide additional semantic information to a tabular-data representation model.
Claim 5. Iida Wang and Kim disclose the method as described in claim 4, Wang further discloses wherein the generating includes forming a plurality of samples from the aligned knowledge graph (we inject two kinds of knowledge in this work, including (1) factual knowledge obtained from automatically aligned text triplets on Wikipedia and Wikidata and (2) linguistic knowledge obtained via dependency parsing) (abstract). One would have been motivated to do so to improve semantic understanding and model performance.
Claim 6. Iida Wang and Kim disclose the method as described in claim 5, Wang further discloses wherein the plurality of samples includes a plurality of triplet sets, each said triplet set defining a first entity, a second entity, and a relationship between the first and second entities (p. 1406, Related work, Table 1; p. 1408, Section 3.3 Factual Adapter). One would have been motivated to do so in order to improve semantic understanding and model performance.
Claim 7. Iida Wang and Kim disclose the method as described in claim 1, Wang further discloses wherein the machine-learning model includes an adapter module having dual-path architecture including a tabular adapter module (knowledge specific adapter) and a knowledge adapter module (p. 1407, Sections 3 and 3.1-3.3, fig. 1b). One would have been motivated to do so to improve semantic understanding and model performance.
Claim 8. Iida Wang and Kim disclose the method as described in claim 7, Wang further discloses wherein the training includes: training the tabular adapter module using a plurality of samples formed from the aligned knowledge graph generated by aligning the tabular data corpus with the knowledge graph (knowledge specific adapter); and training the knowledge adapter module using the knowledge graph (p. 1405, abstract, introduction; fig. 1b, sections 3.1-3.3). One would have been motivated to do so to improve semantic understanding and model performance.
Claim 9. Iida Wang and Kim disclose the method as described in claim 1, Iida further discloses comprising: obtaining a pre-trained machine-learning model; and generating an adapted pre-trained machine-learning model by adding an adapter to the pre-trained machine learning model, and wherein the generating the trained machine-learning model includes training the adapted pre-trained machine learning model (abstract, Sections 2-2.4).
Claim 10. Iida Wang and Kim disclose the method of claim 9, Iida further discloses wherein the training the adapted pre-trained machine learning model includes training the adapter and keeping layers of the pre-trained machine learning model fixed (Sections 2.1, 3.2).
Claim 14. Iida discloses a non-transitory computer-readable storage medium storing executable instructions, which when executed by a processing device, cause the processing device to perform operations comprising: receiving an input including an item of tabular data (fig. 1); and generating a tabular-data type prediction by processing the item of tabular data using a machine-learning model, the machine-learning model including one or more adapter layers trained based on a tabular data corpus having a plurality of items of tabular data (TABBIE architecture relies on two Transformers that independently encode rows and columns, respectively; their representations are pooled at each layer. This setup reduces the sequence length of each Transformer’s input, which cuts down on its complexity, while also allowing us to easily extract representations of cells, rows, and columns… TABBIE uses a simplified training objective compared to masked language modeling: instead of predicting masked cells, we repurpose ELECTRA’s objective function for tabular pretraining by asking the model to predict whether or not each cell in a table is real or corrupted) (Introduction: col. 2, ll 19 thru p. 2, col. 1; abstract, Sections 2-2.4).
Iida does not explicitly disclose a knowledge graph and aligning the tabular data corpus with the knowledge graph. However, Wang discloses (..K-ADAPTER, a framework that retains the original parameters of the pre-trained model fixed and supports the development of versatile knowledge-infused model) (abstract). Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iida further in view of Wang to incorporate the above cited features. One would have been motivated to do so to improve semantic understanding and model performance.
However, Kim discloses aligning the tabular data corpus with the knowledge graph (abstract, [0045]-[0049]-[0051], figs 5-9, and claims 1 and 15). Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Iida further in view of Wang to incorporate the above cited features. One would have been motivated to do so to provide additional semantic information to a tabular-data representation model.
Claim 15. Iida Wang and Kim disclose the non-transitory computer-readable storage medium as described in claim 14, Iida further discloses wherein the tabular-data type prediction includes a column type prediction, a relation prediction, an outlier cell prediction, a table classification, column-based embedding retrieval, or entity-based embedding retrieval (table 1, fig. 4; sections 3.1-4).
Claim 16. Iida Wang and Kim disclose the non-transitory computer-readable storage medium as described in claim 14, Wang further discloses wherein machine-learning model is trained using a plurality of samples obtained from an aligned knowledge graph generated by aligning the tabular data corpus with the knowledge graph (p. 1405, abstract, introduction; fig. 1b). One would have been motivated to do so to improve semantic understanding and model performance.
Claim 17. Iida Wang and Kim disclose the non-transitory computer-readable storage medium as described in claim 16, Wang further discloses wherein the plurality of samples includes a plurality of triplet sets, each said triplet set defining a first entity, a second entity, and a relationship between the first and second entities (p. 1406, Related work, Table 1; p. 1408, Section 3.3 Factual Adapter). One would have been motivated to do so to improve semantic understanding and model performance.
Claim 18. Iida Wang and Kim disclose the non-transitory computer-readable storage medium as described in claim 14, Wang further discloses wherein the machine-learning model includes an adapter module having dual-path architecture including a tabular adapter module (knowledge specific adapter) and a knowledge adapter module (p. 1407, Sections 3 and 3.1, fig. 1b). One would have been motivated to do so to improve semantic understanding and model performance.
Claim 19. Iida Wang and Kim disclose the non-transitory computer-readable storage medium as described in claim 18, Wang further discloses wherein: the tabular adapter module is trained using a plurality of samples formed from an aligned knowledge graph generated by aligning the tabular data corpus with the knowledge graph (knowledge specific adapter); and the knowledge adapter module is trained using the knowledge graph (p. 1405, abstract, introduction; fig. 1b). One would have been motivated to do so to improve semantic understanding and model performance.
Claim 20. Iida Wang and Kim disclose the non-transitory computer-readable storage medium of claim 18, Wang further discloses wherein the machine-learning model is a transformer and layers of the adapter module are disposed between transformer layers of the transformer (p. 1407 Section 3.1). One would have been motivated to do so to improve semantic understanding and model performance.
8. Claims 11-13 are rejected under 35 U.S.C. 103 as being unpatentable over Wang et al. (K-ADAPTER: Infusing Knowledge into Pre-Trained Models with Adapters August 2021) in view of Iida et al. (TABBIE: Pretrained Representations of Tabular Data May 2021) and further in view of Kim (US 2022/0284309).
Claim 11. Wang discloses a machine-learning system executed on one or more processors comprising: a transformer machine learning model having a plurality of transformer layers configured to implement (fig. 1b, Sections 3.1-3.2) relevance of a result of a first adapter layer and a result of a second adapter layer, respectively (p. 1407, §3, that multiple kinds of knowledge are injected into different compact neural models, i.e., adapters, independently. Each adapter is associated with a different knowledge type and produces a corresponding representation… Figure 1(b), p. 1407, expressly depicts multiple adapter paths associated with different knowledge sources, including separate adapter processing associated with factual and linguistic knowledge.) including:
and a dual-path architecture including: the first adapter layer trained using training data including a plurality of samples (p. 1407, §3, Wang explains that K-ADAPTER injects multiple kinds of knowledge into different compact neural models independently, rather than directly modifying the pretrained model. Wang further explains that each kind of knowledge is injected into a corresponding knowledge-specific adapter and that the resulting representations are disentangled); and the second adapter layer trained using the aligned knowledge graph (p. 1407, §3, Wang states that each adapter is pretrained independently on a different task, and specifically identifies: a factual adapter trained on relation classification; and a linguistic adapter trained on dependency-relation prediction) (pp. 1407–1408, §3).
Wang does not explicitly disclose Tabular module, by calculating weights indicating an amount of relevance of a result of a first adapter layer and a result of a second adapter layer, respectively, and formed from an aligned knowledge graph, the aligned knowledge graph generated by aligning a plurality of items included in a tabular data corpus with a knowledge graph.
However, Iida discloses Tabular module (Introduction: col. 2, ll 19 thru p. 2, col. 1), a self-attention mechanism (In §2.1, Iida states that the row Transformer "uses self-attention to produce contextualized output representations" and that the column Transformer similarly produces contextualized representations. Iida, Equation 1, lines corresponding to the row and column Transformer operation.), by calculating weights indicating an amount of relevance of a result of a first adapter layer and a result of a second adapter layer, respectively (self-attention-based contextualization of cells within rows and columns §2.1),. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Wang further in view of Iida to incorporate the above cited features. One would have been motivated to do so to improve prediction accuracy.
However, Kim discloses formed from an aligned knowledge graph, the aligned knowledge graph generated by aligning a plurality of items included in a tabular data corpus with a knowledge graph (abstract, [0045]-[0049]-[0051], figs 5-9, and claims 1 and 15). Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Wang further in view of Kim to incorporate the above cited features. One would have been motivated to do so to provide additional semantic information to a tabular-data representation model.
Claim 12. Wang Iida and Kim disclose the system as described in claim 11, Iida further discloses wherein the transformer layers remain fixed during the training of the first adapter layer and the second adapter layer (Sections 2.1, 3.2). One would have been motivated to do so to improve prediction accuracy.
Claim 13. Wang Iida and Kim disclose the system as described in claim 11, Wang further discloses wherein the tabular adapter module is trained using a plurality of samples formed from the aligned knowledge graph, the plurality of samples including a plurality of triplet sets, each said triplet set defining a first entity, a second entity, and a relationship between the first and second entities (p. 1406, Related work, Table 1; p. 1408, Section 3.3 Factual Adapter).
Response to Arguments
9. Applicant’s arguments and amendments filed on 6/18/2026 have been fully considered but are not persuasive.
Conclusion
10. 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.
11. The prior art made of record and not relied upon is considered pertinent to applicant's disclosure (See PTO-892).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Phenuel S. Salomon whose telephone number is (571) 270-1699. The examiner can normally be reached on Mon-Fri 7:00 A.M. to 4:00 P.M. (Alternate Friday Off) EST.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Usmaan Saeed can be reached on (571) 272-4046. The fax phone number for the organization where this application or proceeding is assigned is 571-273-3800.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/PHENUEL S SALOMON/Primary Examiner, Art Unit 2146
Prosecution Insights