DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
This Office Action is in response to correspondence filed 30 October 2025 in reference to application 18/233,323. Claims 1-7, 12-16, and 21-27 are pending and have been examined.
Response to Amendment
The amendment filed 30 October 2025 has been accepted and considered in this office action. Claims 1-4 and 12-15 have been amended, claims 8-11 and 17-20 cancelled, and claims 21-27 added.
Response to Arguments
Applicant's arguments filed 30 October 2025 have been fully considered but they are not persuasive.
Applicant argues, see Remarks page 9, that the rejections under 35 USC 101 should be withdrawn due to the claim amendments. However the added limitations that the audio signals are received from a video conference and that a transcript is generated do not change the fact that the steps could be performed as a mental process. For example, a human is capable of listening to audio recorded in a video conference and generating the transcript as claimed. Therefore the rejections under 35 USC 101 are maintained.
Applicant argues, see Remarks pages 9-10 that Anidjar and Jung fail to teach the limitations of the claims.
Specifically, Applicant argues “Anidjar has no mentioning of a video conference. Thus, Anidjar does not teach or suggest "receiving audio input comprising mixed audio signals provided by one or more client devices of one or more users of a video conference platform that participate in a video conference’.” This particular augment is moot in light of new grounds of rejections presented below in view of Yoshioka et al.
Applicant next argues “Anidjar does not teach or suggest ‘converting the audio input into a plurality of discrete tokens comprising one or more of a plurality of acoustic tokens or a plurality of semantic tokens derived from text data extracted from the mixed audio signals’.” However examiner notes that Anidjar converts the audio to text and then then notates the start and time of each word, See section IV. This listing of the start and stop time of each word indicates the corresponding discrete acoustic token as claimed. Therefore Anidjar teaches this limitation.
Claim Rejections - 35 USC § 101
The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
Claims 1-6, 12-16, and 21-26 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract without significantly more.
Claims 1, 12, and 21 recite receiving audio input comprising mixed audio signals provided by one or more client devices of one or more users of a video conference platform that participate in a video conference; converting the audio input into a plurality of discrete tokens comprising one or more of a plurality of acoustic tokens or a plurality of semantic tokens derived from text data extracted from the mixed audio signals; and determining a plurality of sound sources each corresponding to a subset of discrete tokens of a plurality of subsets of discrete tokens, wherein a transcript of at least one of the plurality of sounds sources is generated for the video conference.
The limitation of r receiving audio input comprising mixed audio signals provided by one or more client devices of one or more users of a video conference platform that participate in a video conference, as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. That is, other than reciting “client devices” in claims 1, 12, and 21 “a memory device” and “a processing device” in claim 12, and “a non-transitory computer readable medium” in claim 21, nothing in the claim element precludes the step from practically being performed in the mind. For example, but for the computer components, “receiving” in the context of this claim encompasses a person listening to an audio signal that was recorded during a video conference.
Similarly, the limitation of converting the audio input into a plurality of discrete tokens comprising one or more of a plurality of acoustic tokens or a plurality of semantic tokens derived from text data extracted from the mixed audio signals, as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. For example, but for the computer components, “converting” in the context of this claim encompasses the person writing the words they heard in the audio input and a start and stop time for each word.
Finally, the limitation of determining, a plurality of sound sources each corresponding to a subset of discrete tokens of a plurality of subsets of discrete tokens, as drafted, wherein a transcript of at least one of the plurality of sounds sources is generated for the video conference, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. This limitation additionally requires using a trained machine learning model. However this trained machine learning model is claimed in such a way that it provides no structural details, and thus can be considered a generic computer component as well. For example, but for the computer components, “determining” in the context of this claim encompasses the person writing the speaker of each token by the token and writing out a transcript.
If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components, then it falls within the “Mental Processes” grouping of abstract ideas. Accordingly, the claims recite an abstract idea.
This judicial exception is not integrated into a practical application. In particular, the claim only additionally recites computer components. The computer components are recited at a high-level of generality such that it amounts no more than mere instructions to apply the exception using a generic computer component. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. The claims are directed to an abstract idea.
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional element of computer components amounts to no more than mere instructions to apply the exception using a generic computer component. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept. The claims are not patent eligible.
Claims 2, 13 and 22 additionally recite that the tokens are semantic tokens, providing input comprising the audio input; and obtaining, one or more outputs identifying the plurality of semantic tokens. The recitation that the tokens are sematic tokens do not change the fact that a human can generate semantic tokens. Additionally, a person can perform providing input comprising the audio input; and obtaining one or more outputs identifying the plurality of semantic tokens by listening to the audio input and transcribing the semantic tokens. The claim additionally recites a second machine learning model. However like the other machine learning model this is recited as such a high level that it also constitutes generic computer components. Similar to above, no additional limitations are recited that amount to significantly more than the abstract idea, or provide a practical application. These claims are not patent eligible.
Claims 3, 14, and 23 additionally recite that the tokens are acoustic tokens, providing input comprising the audio input; and obtaining, one or more outputs identifying the plurality of acoustic tokens. The recitation that the tokens are acoustic tokens do not change the fact that a human can generate acoustic tokens. Additionally, a person can perform providing input comprising the audio input; and obtaining one or more outputs identifying the plurality of acoustic tokens by listening to the audio input and labeling the acoustic sounds. The claim additionally recites a second machine learning model. However like the other machine learning model this is recited as such a high level that it also constitutes generic computer components. Similar to above, no additional limitations are recited that amount to significantly more than the abstract idea, or provide a practical application. These claims are not patent eligible.
Claims 4, 15, and 24 additionally recite providing, first input comprising the plurality of discrete tokens and second input comprising another plurality of discrete tokens, wherein each of the plurality of discrete tokens and the other plurality of discrete tokens comprises at least one of: a plurality of acoustic tokens and a plurality of semantic tokens. However a human can gather acoustic labeled tokens and semantic tokens such as words and use those features to determine the sound sources. Similar to above, no additional limitations are recited that amount to significantly more than the abstract idea, or provide a practical application. These claims are not patent eligible.
Claims 5 and 25 additionally recite providing, input comprising at least one or more of: one or more transcripts corresponding to the audio input, one or more audio descriptions corresponding to the audio input, one or more class identities corresponding to the audio input, and one or more captions corresponding to the audio input. However a human can gather acoustic labeled tokens and transcripts such as words and use those features to determine the sound sources. Similar to above, no additional limitations are recited that amount to significantly more than the abstract idea, or provide a practical application. These claims are not patent eligible.
Claims 6, 16, and 26 additionally recite obtaining one or more outputs identifying one or more transcripts corresponding to the audio input. However a human can obtain a transcription by writing out the transcript of the audio marked with speakers. Similar to above, no additional limitations are recited that amount to significantly more than the abstract idea, or provide a practical application. These claims are not patent eligible.
EXAMINER NOTE: Claims 7 and 27 are NOT rejected as being ineligible because they contains steps that cannot be meaningfully completed in the mind and not mere execution of an abstract idea with generic computer components and therefore includes more than the abstract ideas of claims 1 and 21 from which claims 7 and 27 depend.
Claim Rejections - 35 USC § 103
Claim(s) 1, 4-6, 12, 15-16, 21, and 24-26 is/are rejected under 35 U.S.C. 103 as being unpatentable over Anidjar et al. (Hybrid Speech and Text Analysis Methods for Speaker Change Detection) in view of Yoshioka et al. (US PAP 2020/0349950).
Consider claim 1, Anidjar teaches A method comprising:
receiving audio input comprising mixed audio signals (section III, input data, Figure 1, audio recordings, Section IV datasets);
converting the audio input into a plurality of discrete tokens comprising one or more of a plurality of acoustic tokens or a plurality of semantic tokens derived from text data extracted from the mixed audio signals (section III, figure 1, converting audio recordings to text, section IV determining start and end time of each word to determine each discrete acoustic token); and
determining, using a trained machine learning model, a plurality of sound sources each corresponding to a subset of discrete tokens of a plurality of subsets of discrete tokens (Section VI, section VII determining speaker labels based on audio features and text tokens).
Anidjar does not specifically teach receiving audio input provided by one or more client devices of one or more users of a video conference platform that participate in a video conference;
wherein a transcript of at least one of the plurality of sounds sources is generated for the video conference.
In the same field of speaker diarization, Yoshioka teaches receiving audio input provided by one or more client devices of one or more users of a video conference platform that participate in a video conference (0051, receiving audio streams from devices 0001, 0042 may be video conferencing devices);
wherein a transcript of at least one of the plurality of sounds sources is generated for the video conference (0051, generating a transcript of the meeting).
Therefore it would have been obvious to one of ordinary skill in the art at the time of effective filing to use client devices as source devices for video conferences as taught by Yoshioka in the system of Anidjar in order generate accurate meeting transcripts (Yoshioka 0001-02).
Consider claim 4, Anidjar teaches The method of claim 1, further comprising:
providing, to the trained machine learning model, first input comprising the plurality of discrete tokens and second input comprising another plurality of discrete tokens, wherein each of the plurality of discrete tokens and the other plurality of discrete tokens comprises at least one of: a plurality of acoustic tokens and a plurality of semantic tokens (Section III, figure 1, speech and word embeddings are provided to the network model, Section V, acoustic embeddings correspond to word tokens as well).
Consider claim 5, Anidjar teaches the method of claim 1, further comprising:
providing, to the trained machine learning model, input comprising at least one or more of: one or more transcripts corresponding to the audio input, one or more audio descriptions corresponding to the audio input, one or more class identities corresponding to the audio input, and one or more captions corresponding to the audio input ((section III, figure 1, converting audio recordings to text transcription).
Consider claim 6, Anidjar teaches The method of claim 1, further comprising: obtaining, from the trained machine learning model, one or more outputs identifying one or more transcripts corresponding to the audio input (Section VII C, speaker diarization of transcripts).
Consider claim 12, Anidjar teaches A system that:
receiving audio input comprising mixed audio signals (section III, input data, Figure 1, audio recordings, Section IV datasets);
converting the audio input into a plurality of discrete tokens comprising one or more of a plurality of acoustic tokens or a plurality of semantic tokens derived from text data extracted from the mixed audio signals (section III, figure 1, converting audio recordings to text, section IV determining start and end time of each word to determine each discrete acoustic token); and
determining, using a trained machine learning model, a plurality of sound sources each corresponding to a subset of discrete tokens of a plurality of subsets of discrete tokens (Section VI, section VII determining speaker labels based on audio features and text tokens).
Anidjar does not specifically teach
a memory device; and
a processing device coupled to the memory device,
receiving audio input provided by one or more client devices of one or more users of a video conference platform that participate in a video conference;
wherein a transcript of at least one of the plurality of sounds sources is generated for the video conference.
In the same field of speaker diarization, Yoshioka teaches
a memory device (0025-26 computer storage and memory); and
a processing device coupled to the memory device (0025, processors),
receiving audio input provided by one or more client devices of one or more users of a video conference platform that participate in a video conference (0051, receiving audio streams from devices 0001, 0042 may be video conferencing devices);
wherein a transcript of at least one of the plurality of sounds sources is generated for the video conference (0051, generating a transcript of the meeting).
Therefore it would have been obvious to one of ordinary skill in the art at the time of effective filing to use computer components and to use client devices as source devices for video conferences as taught by Yoshioka in the system of Anidjar in order to make use of a generic and off the shelf and widely available computer components to implement the system and in order generate accurate meeting transcripts (Yoshioka 0001-02).
Claim 15 contains similar limitations as claim 4 and is therefore rejected for the same reasons.
Claim 16 contains similar limitations as claim 6 and is therefore rejected for the same reasons.
Consider claim 21, Anidjar teaches a system that:
receiving audio input comprising mixed audio signals (section III, input data, Figure 1, audio recordings, Section IV datasets);
converting the audio input into a plurality of discrete tokens comprising one or more of a plurality of acoustic tokens or a plurality of semantic tokens derived from text data extracted from the mixed audio signals (section III, figure 1, converting audio recordings to text, section IV determining start and end time of each word to determine each discrete acoustic token); and
determining, using a trained machine learning model, a plurality of sound sources each corresponding to a subset of discrete tokens of a plurality of subsets of discrete tokens (Section VI, section VII determining speaker labels based on audio features and text tokens).
Anidjar does not specifically teach A non-transitory computer readable medium comprising instructions, which when executed by a processing device, cause the processing device to perform operations;
receiving audio input provided by one or more client devices of one or more users of a video conference platform that participate in a video conference;
wherein a transcript of at least one of the plurality of sounds sources is generated for the video conference.
In the same field of speaker diarization, Yoshioka teaches A non-transitory computer readable medium comprising instructions, which when executed by a processing device, cause the processing device to perform operations (0025-26 computer storage and memory);
receiving audio input provided by one or more client devices of one or more users of a video conference platform that participate in a video conference (0051, receiving audio streams from devices 0001, 0042 may be video conferencing devices);
wherein a transcript of at least one of the plurality of sounds sources is generated for the video conference (0051, generating a transcript of the meeting).
Therefore it would have been obvious to one of ordinary skill in the art at the time of effective filing to use computer components and to use client devices as source devices for video conferences as taught by Yoshioka in the system of Anidjar in order to make use of a generic and off the shelf and widely available computer components to implement the system and in order generate accurate meeting transcripts (Yoshioka 0001-02).
Claim 24 contains similar limitations as claim 4 and is therefore rejected for the same reasons.
Claim 25 contains similar limitations as claim 5and is therefore rejected for the same reasons.
Claim 26 contains similar limitations as claim 6 and is therefore rejected for the same reasons.
Claim(s) 2, 3, 13, 14, 22, and 23 is/are rejected under 35 U.S.C. 103 as being unpatentable over Anidjar and Yoshioka as applied to claims 1 above, and further in view of Dong (CIF: Continuous Integrate-and-Fire for End-to-End Speech Recognition).
Consider claim 2, Anidjar and Yoshioka teach wherein the plurality of discrete tokens comprises a plurality of semantic tokens (Anidjar Section III and IV text tokens), and wherein converting the audio input into the plurality of discrete tokens comprises:
providing, to a second model, input comprising the audio input (Section III, using text2speech engine); and
obtaining, from the second model, one or more outputs identifying the plurality of semantic tokens (Section III, using text2speech engine).
Anidjar and Yoshioka do not specifically teach the second model is a machine learning model.
In the same field of speech processing, Dong teaches teach the second model is a machine learning model (Figure 2 and section 3, using encoder-decoder networks for speech recognition).
It would have been obvious to one of ordinary skill in the art at the time of effective filing to use a learning model for recognition as taught by Dong in the system of Anidjar and Yoshioka in order to increase the accuracy of the speech recognition model (Dong Abstract).
Consider claim 3, Anidjar and Yoshioka teach wherein the plurality of discrete tokens comprises a plurality of acoustic tokens (Anidjar Section III and IV acoustic features for each word), and wherein converting the audio input into the plurality of discrete tokens comprises:
providing, to a second model, input comprising the audio input (Section III, using text2speech engine, section V extracting acoustic features for each word); and
obtaining, from the second model, one or more outputs identifying the plurality of semantic tokens (Section III, using text2speech engine, section V extracting acoustic features for each word).
Anidjar and Yoshioka do not specifically teach the second model is a machine learning model.
In the same field of speech processing, Dong teaches teach the second model is a machine learning model (Figure 2 and section 3, using encoder-decoder networks for speech recognition).
It would have been obvious to one of ordinary skill in the art at the time of effective filing to use a learning model for recognition as taught by Dong in the system of Anidjar and Yoshioka in order to increase the accuracy of the speech recognition model (Dong Abstract).
Claim 13 contains similar limitations as claim 2 and is therefore rejected for the same reasons.
Claim 14 contains similar limitations as claim 3 and is therefore rejected for the same reasons.
Claim 22 contains similar limitations as claim 2 and is therefore rejected for the same reasons.
Claim 23 contains similar limitations as claim 3 and is therefore rejected for the same reasons.
Allowable Subject Matter
Claims 7 and 27 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
Consider claim 7, Anidjar and Jong teach The method of claim 1, but do not specifically teach “providing, to a third trained machine learning model, input comprising a plurality of waveforms corresponding to the audio input, wherein the plurality of waveforms are generated using a time-domain convolutional neural network, and wherein the plurality of waveforms pertains to a first sound source of the plurality of sound sources; obtaining, from the third trained machine learning model, one or more outputs identifying a first plurality of acoustic tokens corresponding to the plurality of waveforms; providing, to the trained machine learning model, second input comprising the first plurality of acoustic tokens corresponding to the plurality of waveforms; and obtaining, from the trained machine learning model, one or more outputs identifying (i) a second plurality of acoustic tokens, wherein the second plurality of acoustic tokens comprise the first plurality of acoustic tokens with a removal of one or more distortions or artifacts from the first plurality of acoustic tokens.” Rather in Anidjar acoustic tokens are generated based on the boundaries of detected words. Therefore claim 7 contains allowable subject matter.
Claim 27 contains similar limitations as claim 7 and therefore contains allowable subject matter as well.
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 DOUGLAS C GODBOLD whose telephone number is (571)270-1451. The examiner can normally be reached 6:30am-5pm Monday-Thursday.
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, Andrew Flanders can be reached at (571)272-7516. 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.
DOUGLAS GODBOLD
Examiner
Art Unit 2655
/DOUGLAS GODBOLD/Primary Examiner, Art Unit 2655