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 .
Response to Amendment
Claims 1, 5, 12 and 16 are amended. Claims 1-20 are presented for examination.
Response to Arguments
Applicant’s arguments filed on 7/23/2026 have been reviewed. Following is the response:
35 U.S.C.§ 103 Rejections
Applicant argues “, the cited portions of Perez set forth that "A full-text indexed search of the index of ontology values is performed to extract candidate matches for each identified mention". Then, "[t]he ranking model is used to rank the candidate (slot, value) pairs by computing a score using the features extracted for each mention." However, the Applicant's attorney respectfully submits that using a "ranking model" to "rank the candidate (slot, value) pairs" extracted from an "index of ontology values" by "computing a score using the features extracted for each mention" fails to teach or suggest that "causing... first neural network model output to be generated" is based on "application of the token embedding and the slot descriptor embedding as input to a first layer of the trained neural network model" and that "determining... that one or more of the tokens correspond to the slot assigned to the selected domain" is "based on application of the first neural network model output as input to one or more additional layers of the trained neural network model or the alternate trained neural network model” Put another way, the Office Action, p. 5, equates Perez's "features extracted from mentions" to independent claim 1's "slot descriptor embeddings", and Perez's "sequence of tokens", to independent claim 1's "token embeddings". However, the Applicant's attorney respectfully submits that the cited portions of Perez fail to teach or suggest that Perez's "features extracted from mentions" and "sequence of tokens" are applied "as input to a first layer of the trained neural network model" to generate "first neural network model output" and that the "first neural network model output" is applied "as input to one or more additional layers of the trained neural network model" to determine "that one or more of the tokens correspond to the slot assigned to the selected domain" as set forth in independent claim 1.”
However, “feature extracted from mentions” and “sequence of tokens” results in slot value pairs which is an input to the ranking model, which is a neural network model ( Para 0065) and the results is the token corresponding to the slot ( Para 0065, 0079-0071, 0116) . Its inherent that in the neural network there will be first layer, hence Perez still reads on claim 1.
Examiner’s Note:
The applicant noted that the amendments find support in paragraphs [0080]–[0083] of the specification. If all concepts from these paragraphs are indeed incorporated into the independent claims, the claims would be allowable over the cited prior art.
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.
And
KSR, 550 U.S. at 418, 82 USPQ2d at 1396. Exemplary rationales that may support a conclusion of obviousness include:
(A) Combining prior art elements according to known methods to yield predictable results;
(B) Simple substitution of one known element for another to obtain predictable results;
(C) Use of known technique to improve similar devices (methods, or products) in the same way;
(D) Applying a known technique to a known device (method, or product) ready for improvement to yield predictable results;
(E) "Obvious to try" – choosing from a finite number of identified, predictable solutions, with a reasonable expectation of success;
(F) Known work in one field of endeavor may prompt variations of it for use in either the same field or a different one based on design incentives or other market forces if the variations are predictable to one of ordinary skill in the art;
(G) Some teaching, suggestion, or motivation in the prior art that would have led one of ordinary skill to modify the prior art reference or to combine prior art reference teachings to arrive at the claimed invention.
See MPEP § 2143 for a discussion of the rationales listed above along with examples illustrating how the cited rationales may be used to support a finding of obviousness. See also MPEP § 2144 - § 2144.09 for additional guidance regarding support for obviousness determination.
Claims 1-4 and 12-14 are rejected under 35 U.S.C. 103 as being unpatentable over Perez (US 20180121415) and further in view of Gillespie ( US 10229680)
Regarding claim 1, Perez teaches a method implemented by one or more processors, comprising: receiving natural language input generated based on user interface input during a human-to-automated assistant dialog ( input utterance, Para 0029); generating a token embedding of tokens determined based on the natural language input ( token, Para 0073, 0077, 0117); selecting a domain based on the human-to-automated assistant dialog( domain specific corpus, Para 0082); determining at least one slot descriptor embedding for at least one textual descriptor of a slot assigned to the selected domain ( detect mentions in the text, Para 0076-0077), wherein the at least one textual slot descriptor embedding (value, topic and slot is based on the “mentions” which is an input to the language model ( for e.g. n-gram ontology model ) is determined based on the application of at least one slot descriptor as input to a trained neural network model (tuple is determined based on mentions using ontology ( n-gram model), Para 0078; which is then input to the ranking model to determine the slot filling and ranking model is a neural network model; neural network can be any combination ( alternative) of neural network look at the e.g. Para 0065) causing, based on application of the token embedding and the slot descriptor embedding as input to a first layer of the trained neural network model or an alternate trained neural network model, first layer output to be generated (( determine slots based on mentions and token, Fig 2, Para 0078-0080; which is an input the ranking model to rank and updating the belief of the dialog state, Par 0110; slots goes to the ranking model which is a neural network( inherently has layers), Para 0065 alternate model is presented as an OR statement, hence the trained model is the model which is being used) wherein the token embedding is applied as input to the first layer of the trained neural network model or the alternate trained neural network model based on being determined based on the tokens determined based on the natural language input (tokenized sequence, Para 0078), and wherein the slot descriptor embedding is applied as input to the first layer of the trained neural network model or the alternate trained neural network model based on being of the slot assigned to the selected domain ( slot, value based on topic, Para 0078-0082);; determining, based on application of the first layer output as input to one or more additional layers of the trained neural network model or the alternate trained neural network model, that one or more of the tokens correspond to the slot assigned to the selected domain ( rank the slots, Para 0079-0081, Para 0109-0110) ; generating an command that includes a slot value for the slot that is based on one or more of the tokens determined to correspond to the slot ( ( task wished by user, Para 0054; info for the slot, Para 0120); and wherein the agent command causes the agent to generate responsive content and transmit the responsive content over one or more networks ( execute task, Fig 2, Para 0053-0054)
Perez does not explicitly teach generating an agent command ; and transmitting the agent command to an agent over one or more networks, transmitting the agent command to an agent over one or more networks, wherein the agent command causes the agent to generate responsive content and transmit the responsive content over one or more networks
However, Gillespie teaches generating an agent command (providing the content, contextual metadata may be generated by the domain using formatting logic. The contextual metadata may be generated such that text corresponding to the content being displayed is formatted into slots and values associated with those slots, where the slots correspond to the slots associated with the domain's intent, Col 38, line 20-38) ; and transmitting the agent command to an agent over one or more networks, transmitting the agent command to an agent over one or more networks, wherein the agent command causes the agent to generate responsive content and transmit the responsive content over one or more networks ( sending out the data to be executed, Claim 1; for e.g. an orchestrator of the speech-processing system may send a request to a multi-domain functionality system that inquires which domain is currently responsible for providing the displayed content to the electronic device. After determining the particular domain, the orchestrator may receive or otherwise cause the natural language understanding system to receive contextual metadata representing content displayed on the client device by the domain., Col 1, line 50-60)
It would have been obvious having the teachings of Perez to further include the concept of Gillespie before effective filing date to resolve one or more declared slots for a particular intent and invoke a particular domain to execute a particular task for e.g. play a song etc. ( Col 38, line 20-38, Gillespie)
Regarding claim 2, Perez as above in claim 1, teaches wherein selecting the domain comprises selecting the agent based on the human-to-automated assistant dialog, and wherein the at least one slot descriptor embedding is determined based on the at least one slot descriptor or the slot descriptor embedding being assigned to the agent ( assigned to a domain, Para 0076)
Regarding claim 3, Perez as above in claim 1, teaches , further comprising: receiving the responsive content generated by the agent ( task executed, Fig 2, Perez; play a song e.g. Col 38, line 20-38)
Regarding claim 4, Gillespie as above in claim 3, teaches : transmitting, to a client device at which the user interface input was provided, output that is based on the responsive content generated by the agent ( an orchestrator of the speech-processing system may send a request to a multi-domain functionality system that inquires which domain is currently responsible for providing the displayed content to the electronic device…. For e.g. song name slot, Col 1, line 50-67, Col 2, line 1-10)
Regarding claim 12, arguments analogous to claim 1, are applicable. In addition, Perez teaches a system ( fig 1)
Regarding claim 13, arguments analogous to claim 2, are applicable.
Regarding claim 14, arguments analogous to claim 3, are applicable.
Regarding claim 15, arguments analogous to claim 4, are applicable.
Claims 5-6 and 16-17 are rejected under 35 U.S.C. 103 as being unpatentable over Perez (US 20180121415) and further in view of Gillespie ( US 10229680) and further in view of Deoras ( US 20150066496 )
Regarding claim 5, Perez modified by Gillespie as above in claim 1, does not explicitly teaches wherein application of the token embedding and the slot descriptor embedding to the first layer of the trained neural network model or the alternate trained neural network model, comprises: applying both the token embedding and the slot descriptor embedding to a combining layer of the trained neural network model or the alternate trained neural network model
However, Deoras teaches applying both the token embedding and the slot descriptor embedding to a combining layer of the trained neural network model or the alternate trained neural network model(combined training of token and context, RNN and DNN, Para 0004, 0007, 0064 Fig 9-10)
It would have been obvious having the teachings of Perez and Gillespie to further include the concept of Deoras to have more improved slot filing ( Para 0006, Deoras)
Regarding claim 6, Deoras as above in claim 5, teaches wherein the combining layer is a feed forward layer ( feed forward, Para 0006)
Regarding claim 16, arguments analogous to claim 5, are applicable.
Regarding claim 17, arguments analogous to claim 6, are applicable.
Claims 7-11 and 18-20 are rejected under 35 U.S.C. 103 as being unpatentable over Perez (US 20180121415) and further in view of Gillespie ( US 10229680) and further in view of Hashimoto ( US 20180121799)
Regarding claim 7, Perez modified by Gillespie as above in claim 1, does not explicitly teach , wherein generating the token embedding of the tokens of the natural language input comprises: applying the tokens to a memory layer of the trained neural network model or the alternate trained neural network model to generate the token embedding
However, Hashimoto teaches wherein generating the token embedding of the tokens of the natural language input comprises: applying the tokens to a memory layer of the trained neural network model or the alternate trained neural network model to generate the token embedding ( token embedding using the bi-directional LSTM ( stacked) network, Para 0186)
It would have been obvious having the teachings of Perez and Gillespie to further include the teachings of Hashimoto before effective filing to improve the processing of words
Regarding claim 8, Hashimoto as above in claim 7, teaches wherein the memory layer is a bi-directional memory layer comprising a plurality of memory units ( stacked, Para 0186)
Regarding claim 9, Hashimoto as above in claim 7, wherein generating the token embedding of the tokens of the natural language input further comprises: applying one or more annotations of one or more of the tokens to the memory layer to generate the token embedding ( annotated words ( tokens) , Para 0099, 0142)
Regarding claim 10, Hashimoto as above in claim 7, teaches wherein the combining layer is downstream from the memory layer, and upstream from one or more additional layers of the neural network model or the alternate trained neural network model ( fig 4a; stacked lstm )
Regarding claim 11, Hashimoto as above in claim 10, teaches wherein the one or more additional layers include at least one of: an additional memory layer; and an affine layer ( lstm inherently have affine layer , Fig 4)
Regarding claim 18, arguments analogous to claim 7, are applicable.
Regarding claim 19, arguments analogous to claim 8, are applicable.
Regarding claim 20, arguments analogous to claim 9, are applicable.
Conclusion
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 Richa Sonifrank whose telephone number is (571)272-5357. The examiner can normally be reached M-T 7AM - 5:30PM.
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, Phan Hai can be reached at (571)272-6338. 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.
/Richa Sonifrank/ Primary Examiner, Art Unit 2654
Prosecution Insights