Response to Amendment
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 .
This action is responsive to amendment filed on 1/12/26. Claims 1-20 are pending.
The claims comprise multidimensional sample dataset and executes a query on the dataset indicating search embeddings in a plurality of subsets that generates output dataset. The concept comprises application in learning models and is therefore statutory.
Double Patenting
The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the “right to exclude” granted by a patent and to prevent possible harassment by multiple assignees. A nonstatutory double patenting rejection is appropriate where the conflicting claims are not identical, but at least one examined application claim is not patentably distinct from the reference claim(s) because the examined application claim is either anticipated by, or would have been obvious over, the reference claim(s). See, e.g., In re Berg, 140 F.3d 1428, 46 USPQ2d 1226 (Fed. Cir. 1998); In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) or 1.321(d) may be used to overcome an actual or provisional rejection based on nonstatutory double patenting provided the reference application or patent either is shown to be commonly owned with the examined application, or claims an invention made as a result of activities undertaken within the scope of a joint research agreement. See MPEP § 717.02 for applications subject to examination under the first inventor to file provisions of the AIA as explained in MPEP § 2159. See MPEP § 2146 et seq. for applications not subject to examination under the first inventor to file provisions of the AIA . A terminal disclaimer must be signed in compliance with 37 CFR 1.321(b).
The filing of a terminal disclaimer by itself is not a complete reply to a nonstatutory double patenting (NSDP) rejection. A complete reply requires that the terminal disclaimer be accompanied by a reply requesting reconsideration of the prior Office action. Even where the NSDP rejection is provisional the reply must be complete. See MPEP § 804, subsection I.B.1. For a reply to a non-final Office action, see 37 CFR 1.111(a). For a reply to final Office action, see 37 CFR 1.113(c). A request for reconsideration while not provided for in 37 CFR 1.113(c) may be filed after final for consideration. See MPEP §§ 706.07(e) and 714.13.
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/patent/patents-forms. The actual filing date of the application in which the form is filed determines what form (e.g., PTO/SB/25, PTO/SB/26, PTO/AIA /25, or PTO/AIA /26) should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to www.uspto.gov/patents/apply/applying-online/eterminal-disclaimer.
Claims 1-20 are rejected on the ground of nonstatutory double patenting as being unpatentable over claims of U.S. Patent No. 11,886,435. Although the claims at issue are not identical, they are not patentably distinct from each other because they comprise substantially the same subject matter using broader terminology.
Claims 2-11 and 1-19 depend from claims 1 and 12 respectively, and are therefore rejected on the merits.
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) 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Jaech et al (USPN. 10,268,646) in view of Interlandi et al (USPN. 2023/0244662).
Regarding claims 1 and 12, Jaech discloses a method and system, comprising: one or more processors coupled to memory, the one or more processors configured to (figs. 4 and 5):
identify a query for a multi-dimensional sample dataset, the query indicating an operation to perform on a plurality of samples of the multi-dimensional sample dataset, (see figs. 3 and 4, col. 22, line 26 to col. 24, line 9, Match-Tensor architecture, query match matrix based on match tensor where query is the size of an embedded vector), but Jaech does not explicitly teach,
“the operation identifying a first column identifier of the respective first tensor, the query specifying (i) a second column identifier for a second column of an output data structure in which a respective virtual tensor for a subset of samples of the plurality of samples is to be stored, the second column not included in the multi-dimensional sample dataset and ii) one or more filtering criteria”.
However, Interlandi teaches,
“the operation identifying a first column identifier of the respective first tensor, the query specifying (i) a second column identifier for a second column of an output data structure in which a respective virtual tensor for a subset of samples of the plurality of samples is to be stored, the second column not included in the multi-dimensional sample dataset, and ii) one or more filtering criteria” (fig. 2A, query 104 and 122(1), “select a from xyz where b = c”, b=c is an example of filtering criteria, par. 32-33, wherein selected X is from the first column and Y is from the second column, note that table XYZ has at least three columns, see generation of tensors as done in step ii) “Each dimension of the tensor can have a different respective number of elements or values. In some examples, the data formatter may generate a tensor for each column of a database table” and see each column may be represented by multiple tensors below “As an example, a column of integers or Boolean values in the data store 106 may be represented as a one dimension tensor (e.g., a vector), while a column of string values may be represented as a two dimensional tensor (e.g., a matrix)” Interlandi).
It would have been obvious to one of ordinary skill in the art at the effective filing time of the application to integrate a query with filtering criteria in Jaech as done by Interlandi by implementing the filtering on multi-dimensional data system and query of Jaech (figs. 2 and 4, items 200 and 401, Jaech). One would have been motivated to effectively search the Jaech social graph (fig. 2, social graph 200, Jaech).
Jaech in view of Interlandi teach
maintain the multidimensional sample dataset comprising a plurality of samples, each plurality sample of the plurality of samples comprising a respective first tensor identified by a first column identifier and comprising one or more respective embeddings, the first column identifier corresponding to a first column in the multidimensional sample dataset in which the respective first tensor of each sample of the plurality of samples is stored (fig. 1, pars. 19 and 28, data store 106 and database table store tensors comprising vectors multidimensional, Interlandi);
execute the query to:
generate the respective virtual tensor for each sample of the plurality of samples, the respective virtual tensor not stored in the multi-dimensional sample dataset, the respective virtual tensor generated by executing the operation specified in the query using the respective first tensor of each sample of the plurality of samples as input, the respective virtual tensor generated to store a respective result of the operation (pars. 40 and 42, “generate one or more DNN data structures 126 corresponding to the query data associated with the data store 106” note that DNN structures are tensors representing information stored in the data store 106 but not the data itself stored in the data store 106, Interlandii),
select the subset of samples from the plurality of samples based on applying the one or more filtering criteria to at least the respective virtual tensor generated for each sample of the plurality of samples to generate the output data structure comprising the subset of samples, each sample of the subset of samples comprising (figs. 2 and 3, items 302-310 and pars. 42 and 43, the created data structures 126 are created from subset of samples matching/relevant to the query, Interlandi):
(i) the respective first tensor identified by the first column identifier in the first column, and (ii) the respective virtual tensor identified by the second column identifier specified in the query, the output data structure generated such that the respective virtual tensor is accessible using the second column identifier (col. 28, line 44- col. 29, line 11, match similar queries to retrieved objects based on embeddings of some objects, note that virtual tensors are equated to subset of sampled data, Jaech, in addition note Interlandi above, query data and formatter generate tensor representations for the desired result where columns meet the required conditions, pars. 32-33, select b=c, from columns x,y,z Interlandi); and
provide the output data structure as a response to the query (figs. 7, items 750 and 780, results in the form of a multi dimensional tensor for identified objects on interface, Jaech).
2. Jaech in view of Interlandi teach, wherein the operation comprises determining one of a Euclidean distance or a cosine similarity between an embedding identified by the query and the one or more respective embeddings in each the plurality of samples (col. 28, line 44- col. 29, line 11, match similar queries to retrieved objects based on degree of similarity and cosine function, Jaech).
3. Jaech in view of Interlandi teach wherein the one or more filtering criteria indicates a second operation to rank each of the subset of samples according to results of a similarity comparison between an embedding identified by the query and the one or more respective embeddings of each sample of the subset of samples (fig. 7, item 770, ranking respective sample objects and relevant citations, Jaech, and par. 32, multiple conditions, Interlandi).
4. Jaech in view of Interlandi teach wherein the one or more filtering criteria indicates a second operation a number of samples of the plurality of samples include into the output data structure (col. 27, lines 48-67, adjusting size of number of embeddings or number of matching strings, Jaech).
5. Jaech in view of Interlandi teach wherein the one or more processors are configured to generate the one or more respective embeddings in each of the plurality of samples using a machine learning model (col. 38, lines 30-64, embeddings are used in machine learning, Jaech).
6. Jaech in view of Interlandi teach, wherein the one or more processors are configured to store, in the respective virtual tensor of each sample of the subset of samples, similarity comparison between the first tensor identified by the operation indicated in the query and the one or more respective embeddings in the sample (fig. 1, item 160, a similarity score is computing representing a degree of similarity between matched objects and query matrix or values, see also relevant content and col. 18, lines 26-49, structured queries bringing together virtual indexes of content based on relation of content to various social graph elements, Jaech).
7. Jaech in view of Interlandi teach wherein the one or more processors are configured to: receive, from a user interface, the query comprising at least one structured query language (SQL) keyword and determine the operation of the query based on the SQL keyword (fig. 7 and col. 17, line 50 to col. 18, line 48, structured query input and output/result, Jaech).
8. Jaech in view of Interlandi teach wherein the user interface is configured to display the respective one or more respective embeddings of each sample of the plurality of samples in response to an interaction (fig. 7, and (col. 28, line 44- col. 29, line 11, match similar queries to retrieved objects based on embeddings and display/results, Jaech).
9. Jaech in view of Interlandi teach wherein the one or more processors are configured to identify the subset of samples according to a match between an embedding identified by the query and the one or more respective embeddings of each sample of the plurality of samples, the match established within a predetermined threshold (col. 28, line 44- col. 29, line 11, match similar queries to retrieved objects based on embeddings within thresholds, Jaech).
10. Jaech in view of Interlandi teach wherein the one or more processors are configured to: detect that the query identifies an attribute indicative of one of a textual item, a graphic feature or metadata corresponding to a search input provided by a user and generate the output data structure according to the attribute (col. 30, lines 6-46, each identified object of the query match may be retrieved thus identify the title of a video clip, as textual, and an object match for the identified video, as image, Jaech).
11. Jaech in view of Interlandi teach wherein the one or more processors are configured to use the output data structure as an input to train one or more machine learning (ML) models (par. 37, lines 26-45, model learns from interactions).
Regarding claim 20, Jaech discloses a medium comprising: one or more processors coupled to memory, the one or more processors configured to (figs. 4 and 5):
identify a query for a multi-dimensional sample dataset, the query indicating an operation to perform on a plurality of samples of the multi-dimensional sample dataset, (see figs. 3 and 4, col. 22, line 26 to col. 24, line 9, Match-Tensor architecture, query match matrix based on match tensor where query is the size of an embedded vector), but Jaech does not explicitly teach,
“the operation identifying a first column identifier of the respective first tensor, the query specifying (i) a second column identifier for a second column of an output data structure in which a respective virtual tensor for a subset of samples of the plurality of samples is to be stored, and ii) one or more filtering criteria”.
However, Interlandi teaches,
“the operation identifying a first column identifier of the respective first tensor, the query specifying (i) a second column identifier for a second column of an output data structure in which a respective virtual tensor for a subset of samples of the plurality of samples is to be stored, and ii) one or more filtering criteria” (fig. 2A, query 104 and 122(1), “select a from xyz where b = c”, b=c is an example of filtering criteria, par. 32-33, wherein selected X is from the first column and Y is from the second column, note that table XYZ has at least three columns, see generation of tensors as done in step ii) “Each dimension of the tensor can have a different respective number of elements or values. In some examples, the data formatter may generate a tensor for each column of a database table” and see each column may be represented by multiple tensors below “As an example, a column of integers or Boolean values in the data store 106 may be represented as a one dimension tensor (e.g., a vector), while a column of string values may be represented as a two dimensional tensor (e.g., a matrix)” Interlandi).
It would have been obvious to one of ordinary skill in the art at the effective filing time of the application to integrate a query with filtering criteria in Jaech as done by Interlandi by implementing the filtering on multi-dimensional data system and query of Jaech (figs. 2 and 4, items 200 and 401, Jaech). One would have been motivated to effectively search the Jaech social graph (fig. 2, social graph 200, Jaech).
Jaech in view of Interlandi teach
maintain the multidimensional sample dataset comprising a plurality of samples, each plurality sample of the plurality of samples comprising a respective first tensor identified by a first column identifier and comprising one or more respective embeddings, the first column identifier corresponding to a first column in the multidimensional sample dataset in which the respective first tensor of each sample of the plurality of samples is stored (fig. 1, pars. 19 and 28, data store 106 and database table store tensors comprising vectors multidimensional, Interlandi);
execute the query to:
apply the operation to the respective first tensor of each sample of the plurality of samples to generate the respective virtual tensor for each sample of the plurality of samples (col. 38, Tables 1 and lines 14-16 and 40-50, labels and results using match-tensor model comprising embeddings, modified Jaech), the respective virtual tensor storing a respective result of the operation (col. 25, line 41 to col. 26, line 27, generate query match matrix for search query, where the matrix corresponds to n-dimensional embeddings for the query terms in an n-dimensional or subset embedding space searched such a “Ariana TV shows”, col. 26, lines 37-67. Note that system produces object match matrix for the first object by encoding generated term embeddings with neural network based on neighboring words contextually, Jaech, in view of Interlandi fig. 2A, item 204, tensor representations);
select the subset of samples from the plurality of samples based on applying the one or more filtering criteria (figs. 2 and 3, items 302-310 and pars. 40 and 42, Interlandi) to at least the respective virtual tensor generated for each sample of the plurality of samples (par. 42, item 308, Interlandi),
generate the output data structure comprising the subset of samples, each sample of the subset of samples comprising the respective first tensor identified by the first column identifier in the first column and the respective virtual tensor identified by the second column identifier in the second column, the output data structure generated such that the respective virtual tensor is accessible using the second column identifier (col. 28, line 44- col. 29, line 11, match similar queries to retrieved objects based on embeddings of some objects, note that virtual tensors are equated to subset of sampled data, Jaech, in addition note Interlandi) query data and formatter generate tensor representations for the desired result where columns meet the required conditions (pars. 32-33, select b=c, from columns x,y,z Interlandi); and
provide the output data structure as a response to the query (figs. 7, items 750 and 780, results in the form of a multi dimensional tensor for identified objects on interface, Jaech).
Methods claims 13-19 comprise substantially the same subject matter as system claims 2-11 and are therefore rejected on the merits.
Response to Arguments
Applicant's amendment filed on 1/12/26 has been considered but is not persuasive. Please see below for details.
Applicant alleges the Double Patenting rejection should be withdrawn as “respective virtual tensor and “operation” are not disclosed.
This is an Obviousness Double Patenting rejection. The claimed respective virtual tensor refers to a generated tensor as claimed “range of values… for inclusion in the second group” in the Double Patenting Obviousness of claim 1. The Tensor data is generated in part based on the query condition (claim 1, Double Patenting). As such, the rejection is maintained.
Applicant alleges the prior art does not teach the amended “query… respective virtual tensor… is to be stored”.
Examiner disagrees.
Jaech in view of Interlandi teaches,
“the operation identifying a first column identifier of the respective first tensor, the query specifying (i) a second column identifier for a second column of an output data structure in which a respective virtual tensor for a subset of samples of the plurality of samples is to be stored, the second column not included in the multi-dimensional sample dataset, and ii) one or more filtering criteria” (fig. 2A, query 104 and 122(1), “select a from xyz where b = c”, b=c is an example of filtering criteria, par. 32-33, wherein selected X is from the first column and Y is from the second column, note that table XYZ has at least three columns, see generation of tensors as done in step ii) “Each dimension of the tensor can have a different respective number of elements or values. In some examples, the data formatter may generate a tensor for each column of a database table” and see each column may be represented by multiple tensors below “As an example, a column of integers or Boolean values in the data store 106 may be represented as a one dimension tensor (e.g., a vector), while a column of string values may be represented as a two dimensional tensor (e.g., a matrix)” Interlandi).
It would have been obvious to one of ordinary skill in the art at the effective filing time of the application to integrate a query with filtering criteria in Jaech as done by Interlandi by implementing the filtering on multi-dimensional data system and query of Jaech (figs. 2 and 4, items 200 and 401, Jaech). One would have been motivated to effectively search the Jaech social graph (fig. 2, social graph 200, Jaech)”.
The combination teaches a query identifying multiple tensors in multiple columns, as each column may represent a tensor.
Applicant alleges the prior art does not teach “generating a respective virtual tensor according to an operation specified in a query”.
Examiner disagrees. The claimed subject matter comprises a multi-dimensional sample dataset which is used to create a table of tensors that can be queried by more than one column.
Similarly, Jaech in view of Interlandi teach,
“generate the respective virtual tensor for each sample of the plurality of samples, the respective virtual tensor not stored in the multi-dimensional sample dataset, the respective virtual tensor generated by executing the operation specified in the query using the respective first tensor of each sample of the plurality of samples as input, the respective virtual tensor generated to store a respective result of the operation (pars. 40 and 42, “generate one or more DNN data structures 126 corresponding to the query data associated with the data store 106” note that DNN structures are tensors representing information stored in the data store 106 but not the data itself stored in the data store 106, Interlandii)”.
The data structure 126 is the created data table comprising a respective virtual tensor with multiple columns based on the query input condition and does not comprise the data from data store 106 but represents the data stored in data store 106 using tensors.
Applicant alleges the output data structure is not generated as claimed.
Examiner is not persuaded. As noted above with reference to data structure 126, the data structure generated does not comprise the data from data store 106 as alleged by the Applicant, instead, the data created in data structure 126 is representative of data store 106 and comprises tensors stored in a plurality of columns.
In conclusion, Examiner notes that independent claim 20 has not been amended to include the alleged subject matter and therefore also remains rejected as no specific allegations regarding claim 20 were submitted.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure in the field of learning models.
USPN. 2021/0124739 : figs 2-3, embeddings and associations.
USPN. 2023/0325391 : Abstract and figs. 1-2, embeddings and tensors calculation.
USPN. 2023/0306087: fig. 1 and par. 24, matrix/tensor query, tensor to tensor ranking
THIS ACTION IS MADE FINAL. 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 MARCIN R FILIPCZYK whose telephone number is (571)272-4019. The examiner can normally be reached M-F 7-4 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, Kavita Stanley can be reached at 571-272-8352 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.
April 8, 2026
/MARCIN R FILIPCZYK/Primary Examiner, Art Unit 2153
Prosecution Insights