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
Introduction
This office action is in response to Applicant’s submission filed on 6/18/2024. As such, claims 1-20 have been examined.
Claim Objections
Claim 17 is objected to because of the following informalities: in line 3-4, there is a repetition of the phrase “or before a last word of the sentence”, that should be deleted. Appropriate correction is required.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Claim 1 recites a method that, under the broadest reasonable interpretation, claims limitations that cover performance of the limitations in the human mind with the assistance of physical aids (e.g., pen and paper), but for the recitation of generic or well-known or conventional computer components. That is, other than reciting “a model” nothing in these claim limitations precludes the steps from practically being performed in the mind. As a whole, claim 1 pertains to supplementing a response to a call, which is a mental process that a human can do. Individually, each of the limitations also pertains to a mental process and/or insignificant extra solution activity, for example:
receiving audio data from a call; (e.g., a data gathering step, a human listen to a incoming call.)
performing services to process the audio data to automatically generate a response, wherein the services include converting the audio data to input text, inputting the input text into a model to automatically generate a text response, and converting the text response to an audio response; (e.g., analysis/translation of voice data, the human providing a translation of the voice into text, the human can mentally process or write it down the text using pen and paper, and then providing a response.) [the model mention could be a language model, or a handbook with rules/scenarios printed that the human agent has access to]
selecting one or more supplemental words based on the input text; (e.g., analysis/evaluation, the human is provided with a list of words that can be used based on the call. The claim does not recite where the selection of supplemental words based on the input text came from, there are various possibilities, including from a language model, or another human.)
determining a type of service based on services performed to generate the audio response; (e.g., analysis/judgement of the voice/speech data, the human analyzing the call and determining the type of service that needs to be performed.)
determining a position in the response to insert the one or more supplemental words based on the type of service; (e.g., involves language understanding/grammar rules of the response, the human figure out where to best say or insert the supplemental words.)
and providing the one or more supplemental words for insertion in the call at the position to supplement the audio response. (e.g., involves language understanding/grammar rules of the response, the human makes the determination on when to use the supplemental words and provides the response to the caller.)
The judicial exception is not integrated into a practical application. In particular, the claims only recites generic computing components. Such generic computing components are recited at a high-level of generality (i.e., as a generic processor performing a generic computer function of receiving, determining, or outputting information) such that they amount to no more than mere instructions to apply the exception using generic computer components. Accordingly, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. The claim is directed to an abstract idea.
Claim 1 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 limitations of using generic computer components amount to no more than mere instructions to apply the exception using generic computer components. Mere instructions to apply an exception using generic computer components cannot provide an inventive concept. Claim 1 is not patent eligible.
The examiner further notes that the use of claimed generic computer components (“a model” to obtain, extract, and/or generate data) invokes such generic computer components “merely as a tool to perform an existing process”. MPEP 2106.05(f). MPEP 2106.05(f) further explains:
Use of a computer or other machinery in its ordinary capacity for economic or other tasks (e.g., to receive, store, or transmit data) or simply adding a general purpose computer or computer components after the fact to an abstract idea (e.g., a fundamental economic practice or mathematical equation) does not integrate a judicial exception into a practical application or provide significantly more. See Affinity Labs v. DirecTV, 838 F.3d 1253, 1262, 120 USPQ2d 1201, 1207 (Fed. Cir. 2016) (cellular telephone); TLI Communications LLC v. AV Auto, LLC, 823 F.3d 607, 613, 118 USPQ2d 1744, 1748 (Fed. Cir. 2016) (computer server and telephone unit). Similarly, "claiming the improved speed or efficiency inherent with applying the abstract idea on a computer" does not integrate a judicial exception into a practical application or provide an inventive concept. Intellectual Ventures I LLC v. Capital One Bank (USA), 792 F.3d 1363, 1367, 115 USPQ2d 1636, 1639 (Fed. Cir. 2015).
Claim 1 recites generic computer components (“a model”), with respect to performing tasks. MPEP 2106.05(d) and (f) further provides examples of court decisions where the courts found generic computing components to be mere instructions to apply a judicial exception, and further explains “increased speed” (e.g., using a computer to increase the speed of an otherwise mental process) does not provide an inventive concept. For example:
A commonplace business method or mathematical algorithm being applied on a general purpose computer, Alice Corp. Pty. Ltd. V. CLS Bank Int’l, 573 U.S. 208, 223, 110 USPQ2d 1976, 1983 (2014); Gottschalk v. Benson, 409 U.S. 63, 64, 175 USPQ 673, 674 (1972); Versata Dev. Group, Inc. v. SAP Am., Inc., 793 F.3d 1306, 1334, 115 USPQ2d 1681, 1701 (Fed. Cir. 2015).
A process for monitoring audit log data that is executed on a general-purpose computer where the increased speed in the process comes solely from the capabilities of the general-purpose computer, FairWarning IP, LLC v. Iatric Sys., 839 F.3d 1089, 1095, 120 USPQ2d 1293, 1296 (Fed. Cir. 2016) (emphasis added).
Performing repetitive calculations. Bancorp Services v. Sun Life, 687 F.3d 1266, 1278, 103 USPQ2d 1425, 1433 (Fed. Cir. 2012) ("The computer required by some of Bancorp’s claims is employed only for its most basic function, the performance of repetitive calculations, and as such does not impose meaningful limits on the scope of those claims.")
Claim 18 recites a CRM claim that corresponds to the method of claim 1 and is therefore rejected under the same grounds as claim 1 above. While claim 18 further recites a “A non-transitory computer-readable storage medium having stored thereon computer executable instructions”, these are merely generic computer components recited at a high-level of generality such that they amount to no more than mere instructions to apply the exception using a generic computer component. Therefore, none of these limitations (a) integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea or (b) amount to significantly more than the judicial exception, because in either case the additional limitations merely utilize generic computer components that amounts to no more than mere instructions to apply the exception using generic computer function. Claim 18 is not patent eligible.
Claim 20 recites an apparatus claim that corresponds to the method of claim 1 and is therefore rejected under the same grounds as claim 1 above. While claim 20 further recites a “one or more processors, and a computer-readable storage medium comprising instructions”, these are merely generic computer components recited at a high-level of generality such that they amount to no more than mere instructions to apply the exception using a generic computer component. Therefore, none of these limitations (a) integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea or (b) amount to significantly more than the judicial exception, because in either case the additional limitations merely utilize generic computer components that amounts to no more than mere instructions to apply the exception using generic computer function. Claim 20 is not patent eligible.
Claims 2-17 and 19 depend from independent claims 1, and 18 respectively, do not remedy any of the deficiencies of claims 1, and 18, and therefore are rejected on the same grounds as claim 1, and 18 from above.
Claim 2 further recites: wherein receiving audio data comprises: receiving the audio data from a call system via a first connection between the call system and a first endpoint, wherein the call system is connected to a call endpoint via a second connection. (e.g., claim recites common telecommunication process, essentially describing system handle call and pass audio between different points in a network or system.)
Claim 3 further recite: wherein performing services comprises: converting the audio data to input text using a speech to text conversion; (e.g., the human performing speech to text conversion.)
inputting the input text into the model to generate the text response; (e.g., the human inputting the text into the model which generate a response.)
and converting the text response to the audio response using a text to speech conversion. (e.g., the human reading out the text from the model.)
Claim 4 further recite: wherein performing services comprises: inputting the input text into a first service to mask a portion of the input text to generate masked input text, wherein the masked input text is input into the model to generate a masked text response. (e.g., the human omitting a word from a text and use knowledge to determine the most probable word, or in this instance, the human can input a masked text into the model.)
Claim 5 further recite: wherein performing services comprises: inputting the masked text response into the first service to unmask a portion of the masked text response to generate an unmasked text response, wherein the unmasked text response is converted to the audio response. (e.g., the human omitting a word from a text and use knowledge to determine the most probable word, then read out the whole text.)
Claim 6 further recite: wherein performing services comprises: determining a first type of service being performed; (e.g., the human determining what the caller is requesting.)
determining a first position in the audio response to insert a first supplemental word based on determining the first type of service; (e.g., the human determining what order to insert the first supplemental word based on the understanding of the type of service the caller is seeking.)
determining a second type of service being performed; (e.g., the human determining what else the caller is requesting.)
and determining a second position in the audio response to insert a second supplemental word based on determining the second type of service. (e.g., the human determining what order to insert the second supplemental word based on the understanding of the second type of service the caller is seeking.)
Claim 7 further recite: wherein selecting one or more supplemental words based on the input text comprises: analyzing the input text to determine a supplemental word type from a plurality of supplemental word types. (e.g., the human analyzing the input to determine the supplemental word type.)
Claim 8 further recite: wherein analyzing the input text comprises: determining an intent of the input text; (e.g., the human analyzing the input to determine the intent.)
and using the intent to select the supplemental word type. (e.g., the human analyzes the intent to select the supplemental word type.)
Claim 9 further recite: wherein the intent is based on a question, a statement, or an emotion that is detected. (e.g., the human figures out the intent either from the question, a statement or from the emotion.)
Claim 10 further recite: wherein selecting one or more supplemental words based on the input text comprises: selecting from a group of supplemental words for the supplemental word type to select the one or more supplemental words. (e.g., the human figures out which words to used from a group of supplemental words based on context.)
Claim 11 further recite: wherein the selection is a random selection from the group of supplemental words. (e.g., the human can randomly pick words to use from a group of supplemental words.)
Claim 12 further recite: wherein selecting one or more supplemental words based on the input text comprises: analyzing the audio data to determine an intent that is used to determine a supplemental word type from a plurality of supplemental word types. (e.g., the human analyzing the audio data to figure out the intent, and based on the intent determine the supplemental word type.)
Claim 13 further recite: wherein determining the position comprises: inserting the one or more supplemental words before the audio response is output. (e.g., the human determining where to place the supplemental word before responding to the caller.)
Claim 14 further recites: wherein determining the position comprises: inserting the one or more supplemental words during the audio response. (e.g., the human determining where to place the supplemental word while responding to the caller.)
Claim 15 further recites: wherein determining the position comprises: supplemental words in the one or more supplemental words are inserted at multiple positions during the audio response. (e.g., the human analyzing and inputting words in multiple position of the response.)
Claim 16 further recites: wherein determining the position comprises: determining the position in the multiple positions based on a type of service being performed. (e.g., the human analyzing and inputting words in various location of the response depending on the type of service being performed.)
Claim 17 further recites: wherein determining the position comprises: determining the position based on a limitation of the position is not at before a last word of a sentence or after an end of a sentence, and determining the position based on a guideline of the position is before the beginning of the sentence. (e.g., the human following a rule or guideline where the supplemental words have to be placed before the last word of the sentence or has to be within the sentence, the human can follow the guideline to put or insert supplemental words in beginning of the sentence.)
Claims 19 is CRM claim that corresponds to claim 6 and is rejected under similar rationale.
In sum, claims 2-17 and 19 depend from claim 1, and 18 respectively, and further recite mental processes as explained above. None of the additional limitations recited in claims 2-17 and 19 amount to anything more than the same or a similar abstract idea as recited in claims 1, and 19. Nor do any limitations in claims 2-17 and 19: (a) integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea or (b) amount to significantly more than the judicial exception because the additional limitations of using generic computer components amounts to no more than mere instructions to apply the exception using generic computer components. Claims 2-17 and 19 are not patent eligible.
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.
Claims 1, 3, 7-14, 17-18 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Mcquinn (US 20240395246), in view of Fujisawa (US 20150206532).
Regarding Claim 1, Mcquinn discloses: 1. A method comprising: receiving audio data from a call; ([0004] receiving audio data characterizing the utterance, the audio data captured by the user device,)
performing services to process the audio data to automatically generate a response, wherein the services include converting the audio data to input text, ([0021] FIG. 1 illustrates an example of a system 100 including a low-latency conversational system 105 for performing automatic speech recognition (ASR) of an utterance 101 spoken by a user 10 of a user device 110,) inputting the input text into a model to automatically generate a text response, ([0021] and providing a low-latency response 102 to the utterance 101 to the user 10 via the user device 110. [0030] In some examples, the first model 150 includes a language model or an LLM, albeit having fewer parameters than an LLM corresponding to the second model 160. In such examples, the first model 150 could be a first LLM associated with a scaled down parameter count version of a second LLM that corresponds to the second model 160. Alternatively, a language model or an LLM of the first model 150 may be trained separately and differently from the second model 160 on a task that only includes predicting the initial portion 102a of a response 102. Notably, separate training of the first model 150 and the second model 160 may better enable the second model 160 to recover from errors of the first model 150 in generating the first text segment 172a. Alternatively, the first model 150 may include an embedding model that projects the transcription 142 into an embedding space corresponding to a plurality of pre-determined first text segments. Alternatively, the first model 150 may include a classifier model configured to select, based on the transcription 142, the first text segment 172a from a plurality of pre-determined first text segments. Alternatively, the first model 150 may include a natural language processing/understanding (NLP/NLU) module. In some implementations, the ASR system 140 and the first model 150 are combined into and trained as a single system or model.) and converting the text response to an audio response; ([0021] In the example of FIG. 1, the response 102 is audibly output by the user device 110.)
selecting one or more supplemental words based on the input text; ([0027] During stage (C), a first model 150 of the low-latency conversational system 105 processes the transcription 142 to generate a first text segment 172, 172a that represents a predicted initial portion 102a of a response 102 to the transcription 142 or, more generally, the utterance 101. …the first text segment 172a represents a generic phrase, a filler phrase, or a prefix phrase.)
determining a position in the response to insert the one or more supplemental words ([0019] The utterance may correspond to a query directed toward the cloud-based LLM. The user device may then immediately start to audibly output a corresponding synthesized speech representation of the initial portion of the response to the query. Because the first model is small (i.e., in terms of computational and memory requirements compared to the cloud-based LLM) and executes on the user device, the audible output of the synthesized speech representation of the initial portion of the response can begin very shortly after the utterance ends, thus, enabling disclosed systems and methods to begin conversationally responding to the utterance with very low latency.)
and providing the one or more supplemental words for insertion in the call at the position to supplement the audio response. ([0020] While the user device generates and/or audibly outputs the synthesized speech representation of the initial portion of the response, a much larger second model (e.g., an LLM having hundreds of billions of parameters) executed by a remote computing system in the cloud may process the transcription, the first text segment and, in some examples, additional context to generate a second text segment that represents a remaining portion of the response to the utterance.)
Mcquinn does not disclose determining a type of service based on services performed to generate the audio response; determining a position in the response to insert the one or more supplemental words based on the type of service;
Fujisawa discloses: determining a type of service based on services performed to generate the audio response; ([0031] Based on the timing chart of FIG. 3, explanation will be given on the timing at which the speech recognition terminal device 10 says a filler word during the later-described speech command process. Now, suppose that the user speaks a speech command with the contents "How's the weather tomorrow?" The speech recognition terminal device 10, on detecting the end of the speech command, determines the contents of filler word of a temporal length, and says the filler word with the determined contents during the response delay time. In the example of FIG. 3, the keyword "weather" contained in the inputted speech command is used and the filler words with the content "weather, let's see " are vocally output from the speaker 18. [0046] In the above filler speaking process (see FIG. 4), a filler word(s) with a time length matching a response delay time that is predicted when the speech command is transmitted to the speech recognition server 20 may be selected. Specifically, multiple kinds of ready-made filler words with time lengths taken for speech may be pre-defined in a dictionary stored in, for example, a storage device 14. In accordance with the predicted response delay time, filler words with an appropriate time length may be selected. When filler words are created using the keyword acquired in the local speech recognition, a time length of filler words may be adjusted by connecting some of the templates. It is noted that the response delay time may be predicted based on, for example, a communication condition with the speech recognition server 20, a past communication history or the like.) [weather request is a type of service perform] Also see para 0024, 0032, 0038 and 00046.
determining a position in the response to insert the one or more supplemental words based on the type of service; ([0031] Based on the timing chart of FIG. 3, explanation will be given on the timing at which the speech recognition terminal device 10 says a filler word during the later-described speech command process. Now, suppose that the user speaks a speech command with the contents "How's the weather tomorrow?" The speech recognition terminal device 10, on detecting the end of the speech command, determines the contents of filler word of a temporal length, and says the filler word with the determined contents during the response delay time. In the example of FIG. 3, the keyword "weather" contained in the inputted speech command is used and the filler words with the content "weather, let's see " are vocally output from the speaker 18.) Also see para 0032.
Mcquinn and Fujisawa are considered analogous art. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Mcquinn to combine the teaching of Fujisawa, because the user can recognize that the speech command he or she has spoken has been received by the system. Because of this, even when there is something of a delay until obtaining the result of speech recognition for the speech command, the user's uneasiness due to not knowing whether or not the speech command has been received can be prevented, and the user's needlessly repeating the speech command can be prevented (Fujisawa, [0011]).
Regarding Claim 3, Mcquinn/Fujisawa discloses all of Claim 1,
Mcquinn further discloses: wherein performing services comprises: converting the audio data to input text using a speech to text conversion; ([0021] FIG. 1 illustrates an example of a system 100 including a low-latency conversational system 105 for performing automatic speech recognition (ASR) of an utterance 101 spoken by a user 10 of a user device 110,)
inputting the input text into the model to generate the text response; ([0021] and providing a low-latency response 102 to the utterance 101 to the user 10 via the user device 110. [0030] In some examples, the first model 150 includes a language model or an LLM, albeit having fewer parameters than an LLM corresponding to the second model 160. In such examples, the first model 150 could be a first LLM associated with a scaled down parameter count version of a second LLM that corresponds to the second model 160. Alternatively, a language model or an LLM of the first model 150 may be trained separately and differently from the second model 160 on a task that only includes predicting the initial portion 102a of a response 102. Notably, separate training of the first model 150 and the second model 160 may better enable the second model 160 to recover from errors of the first model 150 in generating the first text segment 172a. Alternatively, the first model 150 may include an embedding model that projects the transcription 142 into an embedding space corresponding to a plurality of pre-determined first text segments. Alternatively, the first model 150 may include a classifier model configured to select, based on the transcription 142, the first text segment 172a from a plurality of pre-determined first text segments. Alternatively, the first model 150 may include a natural language processing/understanding (NLP/NLU) module. In some implementations, the ASR system 140 and the first model 150 are combined into and trained as a single system or model.)
and converting the text response to the audio response using a text to speech conversion. ([0008] processing, using a text-to-speech system, the first text segment to generate a first synthesized speech representation of the initial portion of the response to the utterance,)
Regarding Claim 7, Mcquinn/Fujisawa discloses all of Claim 1,
Fujisawa further discloses: wherein selecting one or more supplemental words based on the input text comprises: analyzing the input text to determine a supplemental word type from a plurality of supplemental word types. ([0038] In S110, the control device 15 creates a filler word by applying the keyword extracted in S108 to the templates of filler word acquired from the template dictionary stored in the storage device 14. For example, the keyword "weather" may be applied to a black portion of the template ".smallcircle. .smallcircle. (blank), let's see", so that the filler words "weather, let's see" are created. Multiple templates may be prepared on an attribute-by-attribute basis, and an optimum template may be selected according to the attribute of the extracted keyword.) Also see para 0024 and 00046.
The rationale for the combination would be similar to the one already provided.
Regarding Claim 8, Mcquinn/Fujisawa discloses all of Claim 7,
Mcquinn further discloses: wherein analyzing the input text comprises: determining an intent of the input text; ([0018] Automatic speech recognition (ASR) systems and large language models (LLMs) are increasingly used to provide conversational experiences between users and user devices. In general, an ASR system attempts to determine an accurate transcription of what a user utters to a user device, and an LLM generates, based on the transcription and/or unspoken or spoken device context represented as text or embeddings such as a user's location, prior conversation history, address book, on screen summary, etc., a response to the utterance.)
Fujisawa further discloses: wherein analyzing the input text comprises: determining an intent of the input text; and using the intent to select the supplemental word type. ([0031] Based on the timing chart of FIG. 3, explanation will be given on the timing at which the speech recognition terminal device 10 says a filler word during the later-described speech command process. Now, suppose that the user speaks a speech command with the contents "How's the weather tomorrow?" The speech recognition terminal device 10, on detecting the end of the speech command, determines the contents of filler word of a temporal length, and says the filler word with the determined contents during the response delay time. In the example of FIG. 3, the keyword "weather" contained in the inputted speech command is used and the filler words with the content "weather, let's see " are vocally output from the speaker 18.) Also see para 0024, 0038 and 00046.
Wherein the rationale for the combination would be similar to the one already provided.
Regarding Claim 9, Mcquinn/Fujisawa discloses all of Claim 8,
Mcquinn further discloses: wherein the intent is based on a question, a statement, or ([0018] Automatic speech recognition (ASR) systems and large language models (LLMs) are increasingly used to provide conversational experiences between users and user devices. In general, an ASR system attempts to determine an accurate transcription of what a user utters to a user device, and an LLM generates, based on the transcription and/or unspoken or spoken device context represented as text or embeddings such as a user's location, prior conversation history, address book, on screen summary, etc., a response to the utterance.)
Regarding Claim 10, Mcquinn/Fujisawa discloses all of Claim 7,
Fujisawa further discloses: wherein selecting one or more supplemental words based on the input text comprises: selecting from a group of supplemental words for the supplemental word type to select the one or more supplemental words. ([0038] In S110, the control device 15 creates a filler word by applying the keyword extracted in S108 to the templates of filler word acquired from the template dictionary stored in the storage device 14. For example, the keyword "weather" may be applied to a black portion of the template ".smallcircle. .smallcircle. (blank), let's see", so that the filler words "weather, let's see" are created. Multiple templates may be prepared on an attribute-by-attribute basis, and an optimum template may be selected according to the attribute of the extracted keyword.) Also see para 0024 and 00046.
The rationale for the combination would be similar to the one already provided.
Regarding Claim 11, Mcquinn/Fujisawa discloses all of Claim 10,
Fujisawa further discloses: wherein the selection is a random selection from the group of supplemental words. ([0039] As an example, the configuration may be such that the control device 15 randomly selects a filler word from among a plurality of kinds of filler word recorded in advance in the dictionary, those being "Let's see", "Just a moment", and the like. At this time, the configuration may be such that a filler word differing from the filler words used on the previous occasion is selected, so that the filler words the same as the filler words used on the previous occasion are not used on consecutive occasions.)
Where the rationale for the combination would be similar to the one already provided.
Regarding Claim 12, Mcquinn/Fujisawa discloses all of Claim 1,
Mcquinn further discloses: wherein selecting one or more supplemental words based on the input text comprises: analyzing the audio data to determine an intent that is used to ([0018] Automatic speech recognition (ASR) systems and large language models (LLMs) are increasingly used to provide conversational experiences between users and user devices. In general, an ASR system attempts to determine an accurate transcription of what a user utters to a user device, and an LLM generates, based on the transcription and/or unspoken or spoken device context represented as text or embeddings such as a user's location, prior conversation history, address book, on screen summary, etc., a response to the utterance.)
Fujisawa further discloses: determine an intent that is used to determine a supplemental word type from a plurality of supplemental word types. ([0031] Based on the timing chart of FIG. 3, explanation will be given on the timing at which the speech recognition terminal device 10 says a filler word during the later-described speech command process. Now, suppose that the user speaks a speech command with the contents "How's the weather tomorrow?" The speech recognition terminal device 10, on detecting the end of the speech command, determines the contents of filler word of a temporal length, and says the filler word with the determined contents during the response delay time. In the example of FIG. 3, the keyword "weather" contained in the inputted speech command is used and the filler words with the content "weather, let's see " are vocally output from the speaker 18.)
Where the rationale for the combination would be similar to the one already provided.
Regarding Claim 13, Mcquinn/Fujisawa discloses all of Claim 1,
Mcquinn further discloses: wherein determining the position comprises: inserting the one or more supplemental words before the audio response is output. ([0027] In some implementations, the first text segment 172a represents a generic phrase, a filler phrase, or a prefix phrase. For example, the first model 150 may be trained to mirror language from the utterance 101 as the first text segment 172a. For instance, in the example of FIG. 1, the utterance 101 starts with “tell me a story about . . . ” and the first model 150 may be trained to mirror back the phrase “once upon a time . . . ”, such that the first text segment 172a acknowledges the request for a story but is not specific to any particular requested story.)
Regarding Claim 14, Mcquinn/Fujisawa discloses all of Claim 1,
Mcquinn further discloses: wherein determining the position comprises: inserting the one or more supplemental words during the audio response. ([0027] In some implementations, the first text segment 172a represents a generic phrase, a filler phrase, or a prefix phrase. For example, the first model 150 may be trained to mirror language from the utterance 101 as the first text segment 172a. For instance, in the example of FIG. 1, the utterance 101 starts with “tell me a story about . . . ” and the first model 150 may be trained to mirror back the phrase “once upon a time . . . ”, such that the first text segment 172a acknowledges the request for a story but is not specific to any particular requested story.)
Regarding Claim 17, Mcquinn/Fujisawa discloses all of Claim 14,
Mcquinn further discloses: wherein determining the position comprises: determining the position based on a limitation of the position is not at before a last word of a sentence or after an end of a sentence and determining the position based on a guideline of the position is before the beginning of the sentence. ([0005] In some implementations, the first model is trained to generate the first text segment of one or more initial words in the response to the transcription such that the first synthesized speech representation generated from the one or more initial words includes a duration sufficient to mask a latency time period incurred while the second model processes the transcription and the first text segment to generate the second text segment. The one or more initial words may represent at least one of a generic phrase, a filler phrase, or a prefix phrase.) [The initial words may be filler phrase, the second text segment would follow that initial words which are the first text segment, there are no mentioning of text generated right before the end of the second text or after. The first text and second text are both part of the system generated response to the user query]
Regarding Claim 18, Mcquinn discloses: 18. A non-transitory computer-readable storage medium having stored thereon computer executable instructions, which when executed by a computing device, cause the computing device to be operable for: ([0053] These computer programs (also known as programs, software, software applications, or code) include machine instructions for a programmable processor, and can be implemented in a high-level procedural and/or object-oriented programming language, and/or in assembly/machine language. As used herein, the terms “machine-readable medium” and “computer-readable medium” refer to any computer program product, non-transitory computer readable medium, apparatus and/or device (e.g., magnetic discs, optical disks, memory, Programmable Logic Devices (PLDs)) used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal.)
As for the rest of the claim, they claim the elements of Claim 1, therefore the rejection applied to the rejection of Claim 1 is also applicable.
Regarding Claim 20, Mcquinn discloses: 20. An apparatus comprising: one or more computer processors; and a computer-readable storage medium comprising instructions for controlling the one or more computer processors to be operable for: ([0053] These computer programs (also known as programs, software, software applications, or code) include machine instructions for a programmable processor, and can be implemented in a high-level procedural and/or object-oriented programming language, and/or in assembly/machine language. As used herein, the terms “machine-readable medium” and “computer-readable medium” refer to any computer program product, non-transitory computer readable medium, apparatus and/or device (e.g., magnetic discs, optical disks, memory, Programmable Logic Devices (PLDs)) used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term “machine-readable signal” refers to any signal used to provide machine instructions and/or data to a programmable processor.) Also see fig. 3 for apparatus or system configuration.
As for the rest of the claim, they claim the elements of Claim 1, therefore the rejection applied to the rejection of Claim 1 is also applicable.
Claim 2 is rejected under 35 U.S.C. 103 as being unpatentable over Mcquinn, in view of Fujisawa, and further in view of Liu (US 11843719).
Regarding Claim 2, Mcquinn/Fujisawa discloses all of Claim 1,
Mcquinn/Fujisawa does not disclose wherein receiving audio data comprises: receiving the audio data from a call system via a first connection between the call system and a first endpoint, wherein the call system is connected to a call endpoint via a second connection.
Liu discloses: wherein receiving audio data comprises: receiving the audio data from a call system via a first connection between the call system and a first endpoint, wherein the call system is connected to a call endpoint via a second connection. ([col. 11, lines 58-67 – col. 12, lines 1-13] FIG. 2 illustrates an example data-communications system that analyzes digital voice data for tone or sentiment in accordance with various embodiments. In connection with the specifically-illustrated example, endpoint devices 239, 241, 243, 245 connected in a data network 231 are configured to place and receive VoIP telephone calls between other VoIP endpoint devices, and/or between non-VoIP endpoint devices, although embodiments are not limited to VoIP communications systems. Non-VoIP endpoint devices can include, for example, plain old telephone service (POTS) telephones and cellular-capable devices, which might also be VoIP capable (e.g., smart phones with appropriate VoIP software applications). The various endpoint devices 239, 241, 243, 245 are associated with an account 238 of a client, e.g., Client A, and include circuitry that is specially configured to provide calling functions that include interfacing with the appropriate circuitry of the call service provider used by the corresponding endpoint device. In many contexts, a VoIP endpoint device is a VoIP-capable telephone commonly referred to as IP phones. The VoIP endpoint devices 239, 241, 243, 245 can include, but are not limited to, desktop computers, mobile (smart) phones, laptop computers, and tablets, such as illustrated by 240, 242, 244.)
Mcquinn/Fujisawa/Liu are considered analogous art. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Mcquinn/Fujisawa to combine the teaching of Liu, because the system can centralize the analysis of digital voice data across a variety of endpoint devices (Liu, [col. 11-col.12]).
Claims 4-5 are rejected under 35 U.S.C. 103 as being unpatentable over Mcquinn, in view of Fujisawa, and further in view of Petrov (US 20240378427).
Regarding Claim 4, Mcquinn/Fujisawa discloses all of Claim 1,
Mcquinn/Fujisawa does not disclose wherein performing services comprises: inputting the input text into a first service to mask a portion of the input text to generate masked input text, wherein the masked input text is input into the model to generate a masked text response.
Petrov discloses: wherein performing services comprises: inputting the input text into a first service to mask a portion of the input text to generate masked input text, wherein the masked input text is input into the model to generate a masked text response. ([0212] For each span masked input sequence, the system processes, using the neural network, the span masked input sequence to generate a prediction of the one or more tokens that should occupy respective positions of the one or more mask tokens in the span masked input sequence (step 1104). For example, for any position that is occupied by a mask token, the neural network can generate a score distribution over the tokens in the vocabulary from which a predicted token to occupy the position, i.e., to replace the mask token, can be sampled.)
Mcquinn/Fujisawa/Petrov are considered analogous art. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Mcquinn/Fujisawa to combine the teaching of Petrov, because the training system can improve the overall quality of the outputs generated by the neural network after training, i.e., at inference time (Petrov, [0085]).
Regarding Claim 5, Mcquinn/Fujisawa/Petrov discloses all of Claim 4,
Mcquinn already discloses wherein the unmasked text response is converted to the audio response. ([0008] processing, using a text-to-speech system, the first text segment to generate a first synthesized speech representation of the initial portion of the response to the utterance,)
Petrov further discloses: wherein performing services comprises: inputting the masked text response into the first service to unmask a portion of the masked text response to generate an unmasked text response, ([0212] For each span masked input sequence, the system processes, using the neural network, the span masked input sequence to generate a prediction of the one or more tokens that should occupy respective positions of the one or more mask tokens in the span masked input sequence (step 1104). For example, for any position that is occupied by a mask token, the neural network can generate a score distribution over the tokens in the vocabulary from which a predicted token to occupy the position, i.e., to replace the mask token, can be sampled.)
wherein the unmasked text response is converted to the audio response. ([0040] the task can be a text to speech task, where the input is text in a natural language or features of text in a natural language and the network output is a spectrogram or other data defining audio of the text being spoken in the natural language.)
Where the rationale for the combination would be similar to the one already provided.
Claims 15-16 are rejected under 35 U.S.C. 103 as being unpatentable over Mcquinn, in view of Fujisawa, and further in view of Nizar (US 20220245362).
Regarding Claim 15, Mcquinn/Fujisawa discloses all of Claim 14,
Mcquinn/Fujisawa discloses insertion supplemental words during audio response, but they do not disclose inserted at multiple positions.
Nizar discloses: wherein determining the position comprises: supplemental words in the one or more supplemental words are inserted at multiple positions during the ([0066] As an example, a sufficiency criterion may require that the added expanded texts have insertion words at different positions within the base text.) [base text would read on model generated response, and expanded text would read on supplemental words]
Mcquinn/Fujisawa/Nizar are considered analogous art. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Mcquinn/Fujisawa to combine the teaching of Nizar, because the sufficiency criteria act as quality control mechanisms to guarantee the augmented data is sufficiently comprehensive and diversed, leading to more robust and effective systems (Nizar, [0066]).
Regarding Claim 16, Mcquinn/Fujisawa/Nizar discloses all of Claim 15,
Nizar further discloses: wherein determining the position comprises: determining the position in the multiple positions based on a type of service being performed. ([0066] One or more embodiments include determining whether the augmented target set of texts satisfies one or more sufficiency criteria (Operation 216). The target set augmentation system identifies one or more sufficiency criteria. One sufficiency criterion may require that the number of expanded texts being added to the target set be above a threshold number and/or threshold percentage increase. Another sufficiency criterion may require that the types of expanded texts being added to the target set be associated with a certain level of diversity. As an example, a sufficiency criterion may require that the added expanded texts have insertion words at different positions within the base text.) [The cited text describes a process involving analyzing the diversity of positions of added (expanded) texts to satisfy a criterion for the text augmentation service. Sufficiency criterion reads on type of service performed within the context of text augmentation system described. This aligns with the claim's requirement of using position information related to a specific type of service.]
Where the rationale for the combination would be similar to the one already provided.
Potentially Allowable Subject Matter
Claims 6 and 19 are contains potentially allowable subject matter if 101 rejection can be overcome.
With regards to claim 6, Although Mcquinn/Fujisawa disclose supplemental words/phrases into an audio response, however it does not explicitly disclose determining a position for such words based on difference services types. Fujisawa for example discloses inserted supplemental words depending on the keyword extracted which reads on intent and service requested, and also disclose inserting filler words to mask the expected system delay in response, however it does not describe dealing with multiple different type of service and different positioning of multiple separate supplemental response as the claim requires. Other close reference – Hausmann US 20220286732 – disclose a system for dynamically inserting supplemental audio content (potentially a longer ad, song, etc., not just words) into a pre-defined content spot or timespan that has been specified by the content provider within the primary audio content. The selection of the content is based on parameters like user identifiers (device ID, cookie, etc.), not explicitly the type of service being performed as described in the claim. See para 0017-0019, and figs. 3-5 for details. Chen US 20240396856 – disclose insertion of emojis into text based on statistical analysis of historical interactions and emoji probabilities. However, it does not disclose insertion based on types of service. Further, none of them do so in the manner as specifically claimed.
Claim 19 recites elements that are similar to claim 6, and the rationale applied in claim 6 is also similar applicable.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure: Koukoumidis US 20220092131 -discloses generating multi-perspective response using multi agents to augment the response to create a enhance response by inserting supplement information that user may find helpful. See para 0075 and fig 5C for additional details.
Fujisawa (US 20150206531) – discloses determining a filler word with time length to match the system delay in processing and generating system response. See Abstract and figs. 2-4 for additional details.
Labsky US 20140058732 – discloses “Conventionally a client device simply waits for the response. Such a wait, however, can be perceived as a long delay. With techniques herein, however, the client device 112 can initiate a response to spoken utterance prior to having any specific results. For example, client device can analyze the spoken utterance 107 and identify that the user is searching for something. In response, the client device can initiate a response via a user interface, such as with a text-to-speech system. In this non-limiting example, the local recognizer initiates producing or speaking word 151, "Searching the Internet For." These introductory or filler words are then modified by adding words 152, "Apple Pie Recipe," which are presented after words 151. With such a technique, a response to user input is initiated via a user interface prior to having complete results, and then the UI response is modified (in this example the UI response is added-to) to convey results corresponding to the spoken query, such as search results.” See para 0027 for additional details.
Kim US 20150112665 -discloses “a controller for determining a type of the lexically analyzed at least one meaningful word from among a plurality of types, for identifying a plurality of supplementary services corresponding to the type of the lexically analyzed at least one meaningful word, for controlling the user interface displaying of a confirmation message for selecting one of the plurality of supplementary services corresponding to the type of the lexically analyzed at least one meaningful word, and for receiving a confirmation input to select one of the plurality of supplementary services through the user interface; and a dispatcher for driving an application in the mobile communication terminal the corresponding selected supplementary service,” see para 0022 and claim 9 for detail.
Garg, S., & Ramakrishnan, G. (2020). BAE: BERT-based adversarial examples for text classification. arXiv preprint arXiv:2004.01970. – discloses using BERT-MLM to predict mased token in the text for generating adversarial examples. See Abstract, fig. 1 for additional details.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Philip H Lam whose telephone number is (571)272-1721. The examiner can normally be reached 9 AM-3 PM Pacific time.
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, Bhavesh Mehta can be reached on 571-272-7453. 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.
/PHILIP H LAM/ Examiner, Art Unit 2656