Interrupt processing method, apparatus and electronic device

Abstract

The application discloses a method, device and electronic equipment for interrupt processing, and relates to the technical field of voice interaction. The method comprises the following steps: evaluating the content importance of a first voice currently broadcast by the electronic equipment, and obtaining importance distribution information of the first voice changing with a broadcast process; in the process of broadcasting the first voice by the electronic equipment, in response to collecting a second voice input by a user, evaluating an interrupt intention of the second voice, and obtaining an interrupt intention score; based on the importance distribution information and the interrupt intention score, determining a target interrupt time point; interrupting the broadcasting of the first voice at the target interrupt time point, and triggering a response to the second voice. Through the application, the problem that processing of an interrupt request may cause loss of key information is solved, and the technical effect of reducing the risk of losing key information in the interrupt processing process is achieved.

Description

Technical Field

[0001] This application belongs to the field of voice interaction technology, and in particular relates to an interruption handling method, device and electronic device. Background Technology

[0002] In a full-duplex voice interaction system, if an electronic device captures user input voice while playing voice, the electronic device will immediately stop playing the voice and process or respond to the user input voice.

[0003] However, in actual voice interaction, if electronic devices immediately interrupt the playback of voice every time they capture user input, it may lead to the loss of critical information. For example, in a navigation scenario, if the electronic device immediately interrupts the playback of the current navigation information upon capturing user input, the user may lose crucial navigation information, potentially causing them to drive on the wrong road or exit at the wrong time.

[0004] There is no effective solution yet to address the issue that handling interruption requests in related technologies may lead to the loss of critical information. Summary of the Invention

[0005] This application provides an interrupt handling method, apparatus, and electronic device to at least address the problem in related technologies that the processing of interrupt requests may lead to the loss of critical information.

[0006] In a first aspect, embodiments of this application provide an interruption handling method applied to an electronic device. The method includes: evaluating the importance of the content of a first voice currently being played by the electronic device to obtain importance distribution information of the first voice as the playback progresses; during the playback of the first voice by the electronic device, in response to the acquisition of a second voice input by a user, evaluating the interruption intention of the second voice to obtain an interruption intention score; determining a target interruption time point based on the importance distribution information and the interruption intention score; interrupting the playback of the first voice at the target interruption time point, and triggering a response to the second voice.

[0007] In some embodiments, determining the target interruption time point based on the importance distribution information and the interruption intent score includes: selecting multiple candidate interruption time points from the first speech based on the importance distribution information; the importance score of the broadcast content corresponding to each candidate interruption time point is less than a first threshold; determining the interruption cost of each candidate interruption time point based on the importance score corresponding to each candidate interruption time point, the time interval between each candidate interruption time point and the interruption occurrence time, and the interruption intent score; the interruption cost is positively correlated with the importance score, the time interval, and the interruption intent score; and selecting the candidate interruption time point with the lowest interruption cost among the multiple candidate interruption time points as the target interruption time point.

[0008] In some embodiments, selecting multiple candidate interruption time points from the first speech based on the importance distribution information includes: if the interruption intention score is less than a second threshold, and the importance distribution information indicates that the importance score of the broadcast content corresponding to the interruption time is less than a third threshold, then selecting multiple candidate interruption time points from the first speech based on the importance distribution information.

[0009] In some embodiments, determining the target interruption time point based on the importance distribution information and the interruption intent score further includes: if the interruption intent score is less than the second threshold, and the importance score of the broadcast content corresponding to the interruption time is greater than or equal to the third threshold, then determining a target broadcast segment from the first speech; the target broadcast segment is characterized as: in the first speech, a speech segment located after the interruption time, whose importance score is always greater than or equal to the third threshold; the end time of the broadcast of the target broadcast segment is taken as the target interruption time point.

[0010] In some embodiments, using the end time of the broadcast of the target broadcast segment as the target interruption time point includes: if the duration of the target broadcast segment is less than a fourth threshold, then using the end time of the broadcast of the target broadcast segment as the target interruption time point.

[0011] In some embodiments, after determining the target broadcast segment from the first speech, the method further includes: if the duration of the target broadcast segment is greater than or equal to the fourth threshold, dividing the target broadcast segment into multiple broadcast sub-segments according to semantics; selecting multiple candidate interruption time points from the connection points between each of the broadcast sub-segments; determining the interruption cost of each candidate interruption time point based on the importance score corresponding to each candidate interruption time point, the time interval between each candidate interruption time point and the interruption occurrence time, and the interruption intention score; and selecting the candidate interruption time point with the lowest interruption cost among the multiple candidate interruption time points as the target interruption time point.

[0012] In some embodiments, determining the target interruption time point based on the importance distribution information and the interruption intent score further includes: if the interruption intent score is greater than or equal to the second threshold, taking the target time point after the interruption occurrence time as the target interruption time point; the time interval between the interruption occurrence time and the target time point is less than the fourth threshold.

[0013] In some embodiments, the candidate interruption time point is located within a time period starting from the interruption time and lasting for a fourth threshold.

[0014] Secondly, embodiments of this application provide an interruption processing device applied to an electronic device. The device includes: a first evaluation module, configured to evaluate the importance of the content of a first voice currently being played by the electronic device, and obtain importance distribution information of the first voice as the playback progresses; a second evaluation module, configured to evaluate the interruption intention of the second voice in response to the acquisition of a second voice input by a user during the playback of the first voice by the electronic device, and obtain an interruption intention score; a determination module, configured to determine a target interruption time point based on the importance distribution information and the interruption intention score; and a response module, configured to interrupt the playback of the first voice at the target interruption time point and trigger a response to the second voice.

[0015] Thirdly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the interrupt handling method described in any one of the first aspects above.

[0016] Fourthly, embodiments of this application provide a computer program product, including a computer program, which, when run by a processor, causes the interrupt handling method described in any one of the first aspects to be executed.

[0017] Compared to related technologies, the interruption handling method, apparatus, and electronic device provided in this application can evaluate the importance of the content of the first voice currently being played by the electronic device, obtain the importance distribution information of the first voice as the playback progresses, and then, in response to the second voice input by the user during the playback of the first voice, evaluate the interruption intention of the second voice to obtain an interruption intention score. Furthermore, based on the importance distribution information and the interruption intention score, a target interruption time point can be determined. Finally, the playback of the first voice can be interrupted at the target interruption time point, and a response to the second voice can be triggered. In this way, by combining the importance distribution information of the device voice (i.e., the first voice) with the interruption intention score of the user voice (i.e., the second voice) to select the target interruption time, and interrupting the playback of the device voice and triggering a response to the user voice at the target interruption time point, compared to the solution in related technologies that immediately responds to interruption requests, the risk of loss of critical information can be reduced, thereby improving the user's interactive experience. This application solves the problem in related technologies where the processing of interruption requests may lead to the loss of critical information, achieving the technical effect of reducing the risk of loss of critical information during interruption handling.

[0018] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of an interrupt handling method according to an embodiment of this application; Figure 2 This is a flowchart of an interrupt handling method according to another embodiment of this application; Figure 3 This is a schematic diagram of importance distribution information according to one embodiment of this application; Figure 4 This is a flowchart of an interrupt handling method according to another embodiment of this application. Figure 1 ; Figure 5 This is a flowchart of an interrupt handling method according to another embodiment of this application. Figure 2 ; Figure 6 This is a schematic diagram of the structure of an interrupt handling device according to an embodiment of this application; Figure 7 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0021] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0022] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0023] In a full-duplex voice interaction system, if an electronic device captures user input voice while playing voice, the electronic device will immediately stop playing the voice and process or respond to the user input voice.

[0024] However, in actual voice interaction, if electronic devices immediately interrupt the playback of voice every time they capture user input, it may lead to the loss of critical information. For example, in a navigation scenario, if the electronic device immediately interrupts the playback of the current navigation information upon capturing user input, the user may lose crucial navigation information, potentially causing them to drive on the wrong road or exit at the wrong time.

[0025] There is no effective solution yet to address the issue that handling interruption requests in related technologies may lead to the loss of critical information.

[0026] In view of this, embodiments of this application provide an interruption handling method. This method assesses the importance of the content of a first voice message currently being played by an electronic device, obtaining importance distribution information of the first voice message as the playback progresses. Then, during the playback of the first voice message, in response to the acquisition of a second voice message input by the user, the interruption intention of the second voice message is assessed, obtaining an interruption intention score. Furthermore, based on the importance distribution information and the interruption intention score, a target interruption time point can be determined. Finally, the playback of the first voice message can be interrupted at the target interruption time point, and a response to the second voice message can be triggered. In this way, by combining the importance distribution information of the device voice message (i.e., the first voice message) with the interruption intention score of the user voice message (i.e., the second voice message) to select the target interruption time, and interrupting the playback of the device voice message and triggering a response to the user voice message at the target interruption time point, compared to related technologies that immediately respond to interruption requests, the risk of loss of critical information can be reduced, thereby improving the user's interactive experience. This application solves the problem in related technologies where the processing of interruption requests may lead to the loss of critical information, achieving the technical effect of reducing the risk of loss of critical information during interruption handling.

[0027] The following will combine Figure 1 An interrupt handling method provided in one embodiment of this application will be described.

[0028] It should be noted that the execution subject in this embodiment can be a computing service device with data processing, network communication and program execution functions, such as electronic devices such as tablet computers, personal computers, and mobile phones.

[0029] Specifically, the electronic device may include a speaker and a microphone.

[0030] The microphone can be a microphone array, which includes multiple sub-microphones arranged in an array.

[0031] Please see Figure 1 , Figure 1 This is a flowchart of an interrupt handling method according to an embodiment of this application, such as... Figure 1 As shown, the method includes: Step S101: Evaluate the importance of the content of the first voice currently being broadcast by the electronic device to obtain the importance distribution information of the first voice as the broadcast progresses.

[0032] In this embodiment, the electronic device may be a mobile phone, smart TV, smartwatch, smart home control screen, smart speaker, tablet computer, in-vehicle smart terminal, and wearable smart device, etc., equipped with a microphone, processor and corresponding software logic, and capable of capturing user voice input, performing local or cloud processing and executing related operations.

[0033] In this embodiment, the user can interact with the electronic device via voice and perform voice broadcast tasks. During the broadcast, the user may have new or higher-priority needs that require interrupting the voice broadcast of the electronic device. Therefore, it is necessary to collect the user's audio as a basis for judgment. Specifically, the user's audio can be collected in real time using an audio acquisition device such as a microphone.

[0034] In one embodiment, the step S101 above, "evaluating the importance of the content of the first speech currently being broadcast by the electronic device and obtaining the importance distribution information of the first speech as the broadcast process progresses", includes: evaluating the importance of the content of the first speech based on multiple dimensions, and merging the evaluation results of each dimension into importance distribution information as the broadcast process progresses.

[0035] In this embodiment, the aforementioned multiple dimensions may include at least two of the following: information density dimension, criticality dimension, and task or security criticality dimension.

[0036] The information density dimension measures the amount of information carried per unit broadcast duration in the first speech; the criticality dimension measures the criticality of the broadcast content to completing the current task. For example, navigation instructions, security prompts, verification codes, and payment confirmations can be highly critical, while small talk, introductory statements, and other content can be less critical; the task or security criticality dimension measures the severity of the consequences if a user misses the broadcast content. Each dimension, its weight, and the fusion method can be pre-configured or dynamically adjusted.

[0037] Specifically, the evaluation results of each dimension can be mapped to dimension scores. Then, based on the dimension scores and weights corresponding to each dimension, the evaluation results of each dimension can be integrated into a comprehensive importance score, which serves as the importance score for the corresponding broadcast content.

[0038] In this embodiment, the importance of the first speech content at the text level can be aligned and mapped onto the broadcast timeline based on at least one of the timestamps of the synthesized words, phrases, or sentences, prosodic boundaries, audio playback progress, and semantic unit indexes, thereby determining the content importance of the broadcast content corresponding to the current broadcast position in real time during the broadcast process.

[0039] As an example, the first speech can be segmented into multiple semantic units at sentence boundaries, prosodic boundaries, semantic unit boundaries, or secondary information points. Each semantic unit corresponds to one or more speech segments, and the importance distribution information can characterize the mapping relationship between each speech segment and the importance of the content.

[0040] In one embodiment, the importance distribution information can be adjusted by incorporating at least one of the following contextual information: the current task state of the electronic device, scene type, user role, driving or safety status, and user historical preferences. This allows the importance distribution information to further adapt to the interaction context. For example, in an in-vehicle scenario where the user is driving, the importance of the broadcast content corresponding to navigation steering instructions and safety prompts can be increased.

[0041] In this way, by assessing the importance of the content as the electronic device broadcasts the first voice message as the broadcast progresses, subsequent interruptions can be made based on whether the current broadcast content is appropriate to be interrupted, rather than simply based on whether the user's voice has ended. This avoids the loss of critical information caused by the immediate interruption of highly important content during broadcast.

[0042] In step S102, during the process of the electronic device playing the first voice, in response to the second voice input by the user, the interruption intention of the second voice is evaluated to obtain an interruption intention score.

[0043] In one embodiment, the "evaluating the interruption intention of the second speech and obtaining an interruption intention score" in step S102 above includes: performing echo cancellation on the acquired near-end speech to suppress the self-echo of the first speech being played by the electronic device, thereby obtaining the second speech; extracting semantic features and acoustic features from the second speech, and evaluating the interruption intention of the second speech based on the semantic features and acoustic features to obtain an interruption intention score.

[0044] In this embodiment, considering that the near-end speech may be mixed with the self-echo of the first speech while the electronic device is broadcasting the first speech, the near-end speech collected by the microphone may be mixed with the self-echo of the first speech. Therefore, the near-end speech can be echo-cancelled first, and then the interruption intent can be evaluated to reduce the probability of misjudging the self-echo as a user interruption request.

[0045] Specifically, semantic features may include emergency keywords (e.g., "stop", "wait", "no", etc.), semantic scores and confidence levels corresponding to emergency keywords, stability of emergency keywords, and partial recognition results of incremental speech recognition; acoustic features may include prosodic features of the second speech, such as volume, speech rate, intonation, fundamental frequency change rate, and sound energy change rate.

[0046] In this embodiment, the number of repetitions of the interruption request corresponding to the second voice and the time when the interruption occurs can also be included in the evaluation criteria.

[0047] As an example, if a user attempts to interrupt the first voice broadcast multiple times, and the interruption intent score for each interruption request is less than the second threshold, the number of repetitions can be included as a bonus in the current interruption intent score to represent the user's urgent desire to continue expressing themselves.

[0048] In this embodiment, the interruption intent of the second speech can be scored using the following mathematical expression: U=sigmoid(a×z) F0 +b×z E +c×z R +d×K+e×N); where sigmoid() is the activation function used to map any real number to the interval (0,1); z F0 z is the rate of change of the fundamental frequency of speech. E z is the rate of change of sound energy. R K represents the speech rate, K represents the semantic parameter corresponding to the urgent keyword (e.g., the product of the semantic score and confidence level of the keyword), N represents the number of times the interruption request is repeated, and a, b, c, d, and e are all preset calculation weights.

[0049] In one embodiment, if the confidence level of the interruption intent score is lower than a preset confidence threshold, the interruption request will not be considered a high-urgency request (i.e., an interruption request with an interruption intent score greater than or equal to a second threshold). Specifically, when the detection confidence of the second speech is low, or when some recognition results of incremental speech recognition are insufficient to reliably determine the interruption intent, the interruption request can be processed as a low-urgency request (i.e., an interruption request with an interruption intent score less than the second threshold) or the current broadcast can be maintained to avoid high-urgency misjudgments caused by echo remnants and low-confidence recognition.

[0050] In one embodiment, the method further includes: determining the confidence level of the interruption request corresponding to the second speech. The confidence level can be determined based on at least one of the following: the speech recognition result after echo cancellation, the partial recognition result of incremental speech recognition, and the continuity of the near-end speech. When the confidence level of the interruption is lower than a preset confidence threshold, it can be determined that the interruption request is likely a false trigger caused by noise, self-echo of the first speech, or speech from others. In this case, the process of determining the target interruption time point can be skipped; instead, continuous acquisition of the near-end speech can be maintained, and the current broadcast content can be neither interrupted nor delayed, thereby avoiding responding to a non-existent interruption request.

[0051] Step S103: Based on the importance distribution information and the interruption intention score, determine the target interruption time point.

[0052] In this embodiment, the interruption request corresponding to the second voice can be determined as a high-urgency request based on the interruption intent score. Furthermore, if the interruption request is not a high-urgency request, a target interruption time point can be selected from the broadcast timeline of the first voice based on importance distribution information. For example, the target interruption time point is located after the interruption occurs, and the importance score of the broadcast content corresponding to the target interruption time point is low (e.g., the importance score is less than a first threshold).

[0053] Figure 2 This is a flowchart of an interrupt handling method according to another embodiment of this application, such as... Figure 2 As shown, in one embodiment, step S103 includes: Step S201: Based on importance distribution information, select multiple candidate interruption time points from the first speech; the importance score of the broadcast content corresponding to each candidate interruption time point is less than the first threshold.

[0054] In this embodiment, starting from the time of interruption, the system can search backward along the broadcast time axis of the first voice to find troughs where the importance score is less than the first threshold, and use the broadcast position corresponding to the trough as a candidate interruption time point.

[0055] Among them, the candidate interruption time point can be the center position, the start position or the end position of the trough, or it can be the natural pause position of the broadcast, such as the sentence boundary, the prosodic boundary or the semantic unit boundary, so that the interruption occurs at the connection point of the broadcast and reduces the abruptness of the interruption.

[0056] In this embodiment, the first threshold can be pre-configured or dynamically adjusted based on the interaction scenario, the task status of the electronic device, or user preferences, and is used to define low-importance content that can be interrupted; the preset minimum time interval can be met between adjacent candidate interruption time points to avoid jitter in interruption timing caused by too dense candidate interruption time points.

[0057] Step S202: Based on the importance score corresponding to each candidate interruption time point, the time interval between each candidate interruption time point and the interruption occurrence time, and the interruption intent score, the interruption cost of each candidate interruption time point is determined respectively; the interruption cost is positively correlated with the importance score, the time interval, and the interruption intent score.

[0058] In this embodiment, the interruption cost can characterize the comprehensive loss of executing an interruption at a candidate interruption time point. Specifically, the higher the importance score of the broadcast content corresponding to the candidate interruption time point, the greater the loss of key information caused by the interruption at that candidate interruption time point; the larger the time interval between the candidate interruption time point and the interruption occurrence time, the longer the user waits for a response, and the greater the user's waiting loss; the higher the interruption intent score, the stronger the user's expectation of a prompt response, and the stronger the impact of the time interval on the interruption cost.

[0059] Therefore, the interruption cost is positively correlated with the importance score, time interval, and interruption intention score.

[0060] Specifically, the interruption cost of each candidate interruption time point can be determined using the following mathematical expression: C = w1 × S + w2 × U × Δt; where C is the interruption cost of the candidate interruption time point, S is the importance score of the broadcast content corresponding to the candidate interruption time point, U is the interruption intention score, Δt is the time interval between the candidate interruption time point and the interruption occurrence time, and w1 and w2 are preset calculation weights.

[0061] In some other embodiments, S may also be the average importance score of the broadcast content corresponding to the part to be broadcast after the candidate interruption time point (e.g., the broadcast content within a preset time period after the candidate interruption time point (e.g., 0.5 seconds, 1 second, etc.)) to characterize the amount of key information that was not broadcast when the interruption occurred at the candidate interruption time point.

[0062] In other embodiments, the interruption cost may also include a recovery cost term after the interruption, for example, C=w1×S+w2×U×Δt+w3×R; where R is the cost required to resume broadcasting or initiate a callback after the interruption, and w3 is a preset calculation weight.

[0063] In this embodiment, w1 and w2 (or w3) can be dynamically adjusted according to the interaction scenario. As an example, in security-sensitive scenarios such as navigation broadcasts and payments, w1 can be increased to make the loss of critical information account for a larger proportion of the interruption cost; in scenarios where users show a strong desire for help (for example, the user interrupts multiple times during the first voice broadcast but is not dealt with in a timely manner), w2 can be increased to make the user's waiting loss account for a larger proportion of the interruption cost.

[0064] Step S203: Select the candidate interrupt time point with the lowest interrupt cost among multiple candidate interrupt time points as the target interrupt time point.

[0065] In this embodiment, if the interruption cost of multiple candidate interruption time points is the same, the candidate interruption time point with the smallest time interval between the interruption time point and the interruption time point can be selected as the target interruption time point to respond to the user's interruption request as soon as possible.

[0066] In this way, by selecting the candidate interruption time point with the lowest interruption cost from multiple candidate interruption time points, the target interruption time point takes into account both the loss of critical information and the user's waiting time, rather than indiscriminately interrupting or delaying until the end of the broadcast. This can reduce the risk of loss of critical information while taking into account the user's interactive experience.

[0067] Figure 3 This is a schematic diagram of importance distribution information according to one embodiment of this application, such as... Figure 3 As shown, sentence boundaries and secondary information areas can form troughs with lower importance scores, while navigation instructions, security tips, CAPTCHAs, and other content form peaks with higher importance scores.

[0068] like Figure 3 As shown, if user input of the second voice is detected at t0, then t0 can be taken as the time when the interrupt occurs.

[0069] Then, multiple candidate interruption time points can be selected from the interruption time t0, for example, candidate interruption time points t1 and t2; wherein the importance scores of the broadcast content corresponding to candidate interruption time points t1 and t2 are both less than the first threshold.

[0070] Subsequently, the interruption cost for each candidate interruption time point can be determined based on the importance score of the broadcast content corresponding to the candidate interruption time point, the time interval between the candidate interruption time point (t1 or t2) and the interruption time t0, and the interruption intention score.

[0071] Finally, the candidate interrupt cost with the minimum interrupt cost among the candidate interrupt time points t1 and t2 can be taken as the target interrupt time point.

[0072] As an example, assume the interruption intention score U is 0.4, the importance score S1 corresponding to the candidate interruption time point t1 is 0.2 and the time interval Δt1 is 0.5 seconds, the importance score S2 corresponding to the candidate interruption time point t2 is 0.1 and the time interval Δt2 is 1.5 seconds, and if w1 and w2 are both 1, then the interruption cost C of the candidate interruption time point t1 is... 1 =1×0.2+1×0.4×0.5=0.4, the interrupt cost C2 of the candidate interrupt time point t2 is C2=1×0.1+1×0.4×1.5=0.7.

[0073] Since C1 is less than C2, the candidate interruption time point t1 can be used as the target interruption time point. It should be understood that the values ​​of the above parameters are merely examples for ease of understanding and do not constitute a limitation on the scope of protection of this application.

[0074] In one embodiment, the step S201 above, "selecting multiple candidate interruption time points from the first speech based on importance distribution information," includes: If the interruption intent score is less than the second threshold, and the importance distribution information indicates that the importance score of the broadcast content corresponding to the interruption time is less than the third threshold, then multiple candidate interruption time points are selected from the first speech based on the importance distribution information.

[0075] In this embodiment, the second threshold is used to distinguish between high-urgency requests and non-high-urgency requests, and the third threshold is used to determine whether the broadcast content is critical information. When the interruption intent score is less than the second threshold, and the importance score of the broadcast content corresponding to the interruption time is less than the third threshold, it indicates that the user's interruption request is a non-high-urgency request, and the current broadcast content can be interrupted at an opportune time. At this time, the target interruption time point determination process described in steps S201 to S203 above can be executed.

[0076] Figure 4 This is a flowchart of an interrupt handling method according to another embodiment of this application. Figure 1 ,like Figure 4 As shown, in one embodiment, step S103 includes: Step S401: Determine whether the interruption intention score is less than the second threshold; if yes, proceed to step S402; otherwise, proceed to step S408.

[0077] In this embodiment, the second threshold can be pre-configured or dynamically adjusted based on interaction scenarios, historical statistics, or user preferences.

[0078] If the interruption intent score is greater than or equal to the second threshold, it indicates that the user's interruption request is highly urgent (e.g., a request for security alerts, error correction, or cessation) and should be responded to as soon as possible. In this case, the target interruption time point determination process described in step S408 can be entered. If the interruption intent score is less than the second threshold, it indicates that the user's interruption request is less urgent. In this case, further decisions can be made based on the importance score of the current broadcast content, i.e., step S402 can be entered.

[0079] Step S402: Determine whether the importance score of the broadcast content corresponding to the time of the interruption is less than the third threshold; if yes, proceed to step S403; otherwise, proceed to step S406.

[0080] In this embodiment, the third threshold can be pre-configured or dynamically adjusted based on the type of broadcast content, interaction scenario, or user preferences.

[0081] If the importance score of the broadcast content corresponding to the time of interruption is less than the third threshold, it indicates that the currently broadcast content is not critical information. In this case, the process of determining the target interruption time point described in steps S403 to S405 can be entered. If the importance score is greater than or equal to the third threshold, it indicates that the currently broadcast or soon-to-be-broadcast content is critical information and should not be interrupted. In this case, the process of determining the target interruption time point described in steps S406 to S407 can be entered.

[0082] Step S403: Based on the importance distribution information, select multiple candidate interruption time points from the first speech.

[0083] In this embodiment, the specific method for selecting candidate interruption time points can be found in the above description. Figure 2 The relevant descriptions of step S201 in the illustrated embodiment will not be repeated here.

[0084] Step S404: Based on the importance score corresponding to each candidate interruption time point, the time interval between each candidate interruption time point and the interruption occurrence time, and the interruption intent score, determine the interruption cost for each candidate interruption time point.

[0085] In this embodiment, the method for calculating the interruption cost at candidate interruption time points can be found in the above description. Figure 2 The relevant descriptions of step S202 in the illustrated embodiment will not be repeated here.

[0086] Step S405: Select the candidate interrupt time point with the lowest interrupt cost among multiple candidate interrupt time points as the target interrupt time point.

[0087] Step S406: Determine the target broadcast segment from the first speech; the target broadcast segment is characterized as: in the first speech, a speech segment located after the interruption time, whose importance score is always greater than or equal to the third threshold.

[0088] In this embodiment, starting from the moment of interruption, the position where the importance score first drops below the third threshold can be found along the broadcast time axis of the first speech, and the speech segment between the moment of interruption and that position can be used as the target broadcast segment.

[0089] The target broadcast segment can be, for example, a voice segment corresponding to key content such as navigation turning instructions, safety prompts, verification codes, or amount confirmations. In other words, the target broadcast segment can be understood as an uninterruptible interval. During the broadcast of the target broadcast segment, the electronic device can temporarily postpone the interruption and continue to broadcast the target broadcast segment to ensure that the key information is not interrupted or lost.

[0090] In this embodiment, after determining the target broadcast segment from the first speech, the electronic device can enter a delayed interruption window, which can end when the target broadcast segment is finished being broadcast.

[0091] Step S407: The end time of the broadcast of the target broadcast segment is taken as the target interruption time point.

[0092] In this embodiment, the end time of the broadcast of the target broadcast segment is taken as the target interruption time point. The interruption can be executed immediately after the key information is broadcast and the user is responded to. This ensures the complete broadcast of key information and avoids indefinite delay of interruption request.

[0093] Step S408: The target time after the interruption occurs is taken as the target interruption time point; the time interval between the interruption occurrence time and the target time is less than the fourth threshold.

[0094] In this embodiment, when the interruption intent score is greater than or equal to the second threshold, it indicates that the user's interruption request is highly urgent and requires a prompt response. Considering the inherent time consumption of processing steps such as speech recognition, semantic understanding, and interrupt command issuance, the target time after the interruption occurs can be used as the target interruption time point, that is, the interruption should be executed as soon as possible after the interruption occurs.

[0095] Specifically, the time interval between the interruption occurrence and the target time must be less than the fourth threshold. In other words, the fourth threshold also limits the maximum response delay for high-urgency requests. The fourth threshold can be pre-configured or dynamically adjusted based on the interaction scenario, historical statistics, or user preferences. For example, the fourth threshold can be any value between 0.5 seconds and 2 seconds.

[0096] In one embodiment, step S407 includes: if the duration of the target broadcast segment is less than the fourth threshold, then the end time of the broadcast of the target broadcast segment is taken as the target interruption time point.

[0097] In this embodiment, within the delayed interruption window, the electronic device can continuously collect and analyze the user's subsequent voice messages to continuously evaluate the user's interruption intent. When the interruption intent score of the interruption request rises to a level greater than or equal to a second threshold, the electronic device can immediately interrupt the playback of the first voice message and respond to the user, serving as a safety net for high-urgent requests to reduce the probability of delays in high-urgent requests such as safety, error correction, and cessation.

[0098] In this embodiment, considering that the target broadcast segment may be quite long, if no interruption is performed throughout the entire target broadcast segment, the user's interruption request may be suppressed for an extended period. Therefore, the following can be used... Figure 5 In the embodiment shown, when the duration of the target broadcast segment is greater than or equal to the fourth threshold, the target broadcast segment is divided into broadcast sub-segments according to semantics, and the target interruption time point is selected from the connection points between each broadcast sub-segment to limit the continuous delay duration of the interruption request.

[0099] Figure 5 This is a flowchart of an interrupt handling method according to another embodiment of this application. Figure 2 ,like Figure 5 As shown, in one embodiment, after step S406 described above, the method includes: Step S501: Determine whether the duration of the target broadcast segment is less than the fourth threshold; if yes, proceed to step S506; otherwise, proceed to step S502.

[0100] In this embodiment, the fourth threshold also limits the maximum duration of the target broadcast segment as an uninterruptible interval. When the duration of the target broadcast segment is less than the fourth threshold, the broadcast duration of the key information is short, and the delay in waiting for the target broadcast segment to finish broadcasting is limited. Therefore, it is not necessary to divide the broadcast segment into sub-segments, and the end time of the broadcast can be directly used as the target interruption time point. When the duration of the target broadcast segment is greater than or equal to the fourth threshold, in order to avoid suppressing the user's interruption request for a long time, the target broadcast segment can be divided into multiple broadcast sub-segments according to semantics, providing the user's interruption request with an opportunity to execute midway.

[0101] Step S502: Divide the target broadcast segment into multiple broadcast sub-segments according to semantics.

[0102] In this embodiment, the target broadcast segment can be divided into multiple broadcast sub-segments according to key semantic units. As an example, in a navigation broadcast scenario, "turn right at the intersection ahead to enter the main road" can be divided into two broadcast sub-segments: "intersection ahead" and "turn right to enter the main road".

[0103] The granularity of the division allows the key semantic units carried by each broadcast sub-segment to be relatively independent, thereby ensuring the integrity of the key semantic units as much as possible when the execution is interrupted at the connection point between each broadcast sub-segment.

[0104] Step S503: Select multiple candidate interruption time points from the connection points between each broadcast sub-segment.

[0105] In this embodiment, the connection point is the boundary position between two adjacent broadcast sub-segments. Selecting candidate interruption time points from the connection points between each broadcast sub-segment can avoid abruptly cutting off the broadcast content in the middle of the same broadcast sub-segment, thus making the interruption occur at the natural connection point of the broadcast.

[0106] Step S504: Based on the importance score corresponding to each candidate interruption time point, the time interval between each candidate interruption time point and the interruption occurrence time, and the interruption intent score, determine the interruption cost of each candidate interruption time point.

[0107] Step S505: Select the candidate interrupt time point with the lowest interrupt cost among multiple candidate interrupt time points as the target interrupt time point.

[0108] Step S506: The end time of the broadcast of the target broadcast segment is taken as the target interruption time point.

[0109] In this embodiment, when the duration of the target broadcast segment is less than the fourth threshold, the end time of the broadcast of the target broadcast segment can be used as the target interruption time point, and the electronic device can execute the interruption after the key information has been broadcast completely.

[0110] like Figure 3 As shown, if user input of the second voice is detected at t3, then t3 can be taken as the interruption time. If the importance score of the broadcast content corresponding to the interruption time t3 is greater than or equal to the third threshold, then the voice segment located after the interruption time t3 and whose importance score is always greater than or equal to the third threshold is extracted from the first voice, for example, the voice segment corresponding to [t3, t6], and it is taken as the target broadcast segment.

[0111] Then, it can be determined whether the duration of the target broadcast segment (t6-t3) is less than the fourth threshold; if so, the broadcast end time t6 of the target broadcast segment can be directly used as the target interruption time point.

[0112] If not, the target broadcast segment can be divided into multiple broadcast sub-segments according to semantics, such as broadcast sub-segments [t3, t4], broadcast sub-segments [t4, t5], and broadcast sub-segments [t5, t6].

[0113] Multiple candidate interruption points can be selected from the connection points between each broadcast sub-segment, for example, candidate interruption points t4 and t5. Then, the interruption cost of each candidate interruption time point can be calculated separately, and the candidate interruption time point with the lowest interruption cost among multiple candidate interruption time points can be taken as the target interruption time point.

[0114] In this way, when the target broadcast segment is long, by dividing the target broadcast segment into multiple broadcast sub-segments according to semantics, and selecting the candidate interruption time point with the lowest interruption cost from the connection points between each broadcast sub-segment, it is possible to ensure the complete broadcast of key information as much as possible, while limiting the duration of continuous delay of the user's interruption request to within the fourth threshold, thus taking into account both the integrity of key information and the agility of interruption response.

[0115] In one embodiment, the candidate interruption time point is located within a time period starting from the interruption time and lasting for a fourth threshold.

[0116] In this embodiment, since the content that has not yet been broadcast in the first voice gradually decreases as the broadcast progresses, and the amount of key information that has not been broadcast decreases monotonically as the broadcast progresses, if the range of candidate interruption time points is not limited, the minimum interruption cost may degenerate into the timing when the first voice broadcast ends, causing the user's interruption request to be delayed indiscriminately until the broadcast ends.

[0117] Therefore, the candidate interruption time point can be limited to a period of time starting from the interruption time and lasting for a duration of the fourth threshold, so that the target interruption time point is the time within this period when the loss of key information is balanced with the user's waiting time, rather than the time when the first voice broadcast ends.

[0118] In one embodiment, during the selection of candidate interruption time points based on importance distribution information, the waiting time from the interruption occurrence to the selected candidate interruption time point can be limited to not exceeding a preset low-valley waiting time. If the low-valley waiting time is reached but the low point (i.e., the broadcast position where the importance score is less than the first threshold) has not yet been reached, an interruption can be executed immediately to ensure the agility of interruption response under low-importance content and avoid sluggish interaction.

[0119] Step S104: Interrupt the broadcast of the first voice message at the target interruption time point and trigger a response to the second voice message.

[0120] In this embodiment, interrupting the playback of the first voice at the target interruption time point may include: stopping the speech synthesis or audio output of the first voice at the target interruption time point. For example, a stop command may be sent to the speech synthesis subsystem or the audio playback subsystem. Alternatively, the playing audio may be quickly faded out to reduce the abruptness of the interruption.

[0121] In this embodiment, triggering a response to the second voice may include: performing semantic understanding on the second voice (or a portion of the cached recognition result of the second voice), executing an operation corresponding to the second voice, or generating a response voice. For example, in a navigation broadcast scenario, in response to the user's input of the second voice "Don't take this road," the navigation route is replanned; another example is generating and broadcasting a response voice to a user's question.

[0122] In one embodiment, after step S104 above, the method further includes: if the target interruption time point is determined based on the target broadcast segment, after the target broadcast segment has finished broadcasting, initiating a callback to the user to solicit the user's previously delayed interruption request.

[0123] In this embodiment, the callback can be initiated in the form of a prompt, such as prompting the user to repeat their delayed request with phrases like "Regarding what you just mentioned..." or "Do you have anything to say?" After the user responds to the callback, the electronic device can perform semantic understanding and response to the user's repeated speech, or it can directly respond to the interruption request corresponding to the cached second speech.

[0124] In this embodiment, the callback can be initiated when the callback validity conditions are met. As an example, the callback validity conditions may include: the delay duration of the interruption request does not exceed the preset callback validity time limit, the user's second voice or part of its recognition result is still valid, and the user has not canceled the interruption request or left the current turn.

[0125] When the validity conditions for a pullback are not met, a pullback can be skipped. Alternatively, a lightweight confirmation can be initiated instead, such as a brief restatement confirmation of the cached user request, to avoid disturbing the user with a mechanical pullback.

[0126] In one embodiment, the recall or lightweight confirmation can be initiated based on locally cached second speech or partial recognition results. For example, the electronic device can briefly cache the audio of the second speech or partial recognition results locally during the delay of the interruption request, and directly respond to or restate the user's request based on the cache after the recall is completed. To protect user privacy, the cache is only stored locally for a short time and can be deleted when any of the following occurs: the recall is completed, the recall validity period expires, or the user cancels the interruption request; sensitive content in the cache can be stored after de-identification processing.

[0127] In one embodiment, the method further includes: performing online personalized calibration of at least one of the first threshold, the second threshold, the third threshold, and the fourth threshold based on user feedback on interruption handling.

[0128] As an example, the feedback mentioned above could include: whether the user interrupts the request a second time after it has been delayed, the degree to which the user accepts the feedback, and the emotions displayed by the user during the interaction. When the user frequently interrupts the request a second time after it has been delayed, accepts the feedback more readily, or shows impatience, the parameters can be adjusted by lowering the third and fourth thresholds to make the user's interruption requests more likely to be responded to immediately; conversely, when the user prefers to hear the key information, the parameters can be adjusted in the opposite direction to make the electronic device more inclined to broadcast the key information in its entirety.

[0129] In this embodiment, the learning signal, update step size, and parameter range of online personalized calibration can all be configured, and online personalized calibration can be enabled or disabled independently without changing the basic process of the above-mentioned interruption handling method.

[0130] In one embodiment, the above-described interrupt handling method can be applied to various full-duplex voice interaction scenarios.

[0131] For example, in a car navigation scenario, when an electronic device (e.g., an in-vehicle smart terminal) announces the navigation turning instruction "Turn right at the next intersection onto the main road," and the user interrupts with a question, the electronic device determines that the importance score of the content corresponding to the turning instruction is greater than or equal to the third threshold. If so, the electronic device can designate the turning instruction as the target broadcast segment and continue broadcasting the turning instruction to ensure that the user does not miss the turning opportunity. After the turning instruction is finished, the electronic device can initiate a callback to ask the user if they had anything to say. After the user repeats their question, the electronic device responds, thus handling the user's interruption request in a timely manner without losing key information.

[0132] For example, in a casual conversation scenario, if a user interrupts while the electronic device is broadcasting a preparatory statement with a low importance score, the device can immediately interrupt and respond to the user at the current or nearby candidate interruption point, ensuring a smooth and responsive interaction. Both scenarios share the same interruption handling mechanism; the difference lies only in the varying importance distribution information of the first speech and the interruption intent score of the second speech.

[0133] In other embodiments, the above-described interruption handling method can also be applied to full-duplex voice interaction scenarios that require handling user interruption requests during continuous broadcasting, such as smart speakers, wearable personal assistants, customer service or outbound call dialogues, and voice interaction with augmented reality glasses. This application embodiment does not limit this.

[0134] In one embodiment, the interruption handling method described above can be implemented independently or in conjunction with an existing voice interaction subsystem of the electronic device. For example, when the electronic device has at least one of a voice endpoint detection or termination subsystem, an interruption detection subsystem, or a speech synthesis subsystem, the interruption handling method can work in conjunction with these subsystems as an upper-level interruption timing decision logic. For instance, it can obtain timestamps of characters, words, or sentences from the speech synthesis subsystem for importance alignment, and obtain detection events of interruption requests from the interruption detection subsystem to trigger interruption intent assessment, without relying on the specific implementation of these subsystems. That is, the content importance assessment, interruption intent assessment, and interruption timing determination in the above embodiments can be tailored and combined according to device capabilities and business scenarios.

[0135] It should be noted that the information collection process involved in this application (e.g., voice collection process, feedback behavior collection process, etc.) is carried out with the user's knowledge and permission. That is, the information collection process complies with the requirements of relevant standards and does not constitute an act that harms the public interest.

[0136] Through the aforementioned steps S101 to S104, the importance of the first voice message currently being played by the electronic device can be evaluated, obtaining the importance distribution information of the first voice message as the playback progresses. Then, during the playback of the first voice message, in response to the second voice message input by the user, the interruption intention of the second voice message can be evaluated, obtaining an interruption intention score. Furthermore, based on the importance distribution information and the interruption intention score, a target interruption time point can be determined. Finally, the playback of the first voice message can be interrupted at the target interruption time point, and a response to the second voice message can be triggered. In this way, by combining the importance distribution information of the device voice message (i.e., the first voice message) with the interruption intention score of the user voice message (i.e., the second voice message) to select the target interruption time, and interrupting the playback of the device voice message and triggering a response to the user voice message at the target interruption time point, compared to the related technologies that respond immediately to interruption requests, the risk of loss of key information can be reduced, thereby improving the user's interactive experience. This application solves the problem that the processing of interruption requests in related technologies may lead to the loss of key information, achieving the technical effect of reducing the risk of loss of key information during interruption processing.

[0137] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution of the target. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0138] Corresponding to the interrupt handling method described in the above embodiments, Figure 6A schematic diagram of an interrupt handling apparatus according to an embodiment of this application is shown. For ease of explanation, only the parts related to the embodiment of this application are shown.

[0139] Please see Figure 6 The interruption handling device 6 includes: a first evaluation module 60, used to evaluate the importance of the content of the first voice currently being played by the electronic device, and obtain the importance distribution information of the first voice as the playback progresses; a second evaluation module 61, used to evaluate the interruption intention of the second voice in response to the second voice input by the user during the playback of the first voice by the electronic device, and obtain an interruption intention score; a determination module 62, used to determine the target interruption time point based on the importance distribution information and the interruption intention score; and a response module 63, used to interrupt the playback of the first voice at the target interruption time point and trigger a response to the second voice.

[0140] In one embodiment, the determining module 62 is further configured to select multiple candidate interruption time points from the first speech based on importance distribution information; the importance score of the broadcast content corresponding to each candidate interruption time point is less than a first threshold; based on the importance score corresponding to each candidate interruption time point, the time interval between each candidate interruption time point and the interruption occurrence time, and the interruption intention score, the interruption cost of each candidate interruption time point is determined respectively; the interruption cost is positively correlated with the importance score, the time interval, and the interruption intention score; the candidate interruption time point with the smallest interruption cost among the multiple candidate interruption time points is taken as the target interruption time point.

[0141] In one embodiment, the determining module 62 is further configured to select multiple candidate interruption time points from the first speech based on the importance distribution information if the interruption intention score is less than a second threshold and the importance distribution information indicates that the importance score of the broadcast content corresponding to the interruption time is less than a third threshold.

[0142] In one embodiment, the determining module 62 is further configured to determine a target broadcast segment from the first speech if the interruption intention score is less than a second threshold and the importance score of the broadcast content corresponding to the interruption time is greater than or equal to a third threshold; the target broadcast segment is characterized as: in the first speech, a speech segment located after the interruption time, whose importance score is always greater than or equal to the third threshold; the end time of the broadcast of the target broadcast segment is taken as the target interruption time point.

[0143] In one embodiment, the determining module 62 is further configured to use the end time of the broadcast of the target broadcast segment as the target interruption time point if the duration of the target broadcast segment is less than the fourth threshold.

[0144] In one embodiment, the determining module 62 is further configured to: divide the target broadcast segment into multiple broadcast sub-segments according to semantics if the duration of the target broadcast segment is greater than or equal to a fourth threshold; select multiple candidate interruption time points from the connection points between each broadcast sub-segment; determine the interruption cost of each candidate interruption time point based on the importance score corresponding to each candidate interruption time point, the time interval between each candidate interruption time point and the interruption occurrence time, and the interruption intent score; and select the candidate interruption time point with the smallest interruption cost among the multiple candidate interruption time points as the target interruption time point.

[0145] In one embodiment, the determining module 62 is further configured to, if the interruption intent score is greater than or equal to the second threshold, take the target time after the interruption occurrence time as the target interruption time point; the time interval between the interruption occurrence time and the target time point is less than the fourth threshold.

[0146] In one embodiment, the candidate interruption time point is located within a time period starting from the interruption time and lasting for a fourth threshold.

[0147] It should be noted that the aforementioned interrupt handling device 6 can be or be applied to a computing service device with data processing, network communication and program execution functions, such as tablet computers, personal computers, mobile phones, servers and other electronic devices.

[0148] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0149] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0150] Figure 7 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Figure 7As shown, the electronic device 7 includes: at least one processor 70 ( Figure 7 (Only one is shown in the diagram), memory 71, and computer program 72 stored in memory 71 and executable on at least one processor 70, wherein processor 70 executes computer program 72 to implement the steps in any of the above-described interrupt handling method embodiments.

[0151] Electronic device 7 can be a computing service device with data processing, network communication and program execution functions, such as tablet computers, personal computers, mobile phones, servers and other electronic devices.

[0152] The electronic device 7 may include, but is not limited to, a processor 70 and a memory 71. Those skilled in the art will understand that... Figure 7 The example of electronic device 7 is merely an example and does not constitute a limitation on electronic device 7. It may include more or fewer components than shown, or combine certain components, or different components. For example, it may also include input / output devices, network access devices, environmental sensing units (e.g., inertial measurement units, cameras, audio sensors, etc.).

[0153] The processor 70 can be a Central Processing Unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0154] In some embodiments, memory 71 may be an internal storage unit of electronic device 7, such as a hard disk or memory of electronic device 7. In other embodiments, memory 71 may be an external storage device of electronic device 7, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., provided on electronic device 7. In other embodiments, memory 71 may include both internal and external storage units of electronic device 7. Memory 71 is used to store operating system, application programs, bootloader, data, and other programs, such as the program code of computer program 72. Memory 71 may also be used to temporarily store data that has been output or will be output.

[0155] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps in the above-described interrupt handling method embodiments.

[0156] This application provides a computer program product that, when run on an electronic device, enables the electronic device to implement the steps described in the various interrupt handling method embodiments.

[0157] This application implements all or part of the processes in the methods of the above embodiments, which can be accomplished by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate form. The computer-readable medium can include at least: any entity or device capable of carrying the computer program code to an interrupt handling device or electronic device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include a USB flash drive, a portable hard drive, a magnetic disk, or an optical disk.

[0158] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0159] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0160] In the embodiments provided in this application, it should be understood that the disclosed apparatus / network devices and methods can be implemented in other ways. For example, the apparatus / network device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0161] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0162] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. An interrupt handling method, characterized in that, Applied to electronic devices, the method includes: The importance of the content of the first voice currently being broadcast by the electronic device is evaluated to obtain the importance distribution information of the first voice as the broadcast progresses; During the process of the electronic device playing the first voice, in response to the second voice input by the user, the interruption intention of the second voice is evaluated to obtain an interruption intention score; Based on the importance distribution information and the interruption intent score, the target interruption time point is determined; The first voice message is interrupted at the target interruption time point, and a response to the second voice message is triggered.

2. The method according to claim 1, characterized in that, The determination of the target interruption time point based on the importance distribution information and the interruption intent score includes: Based on the importance distribution information, multiple candidate interruption time points are selected from the first speech; the importance score of the broadcast content corresponding to each candidate interruption time point is less than a first threshold. Based on the importance score corresponding to each candidate interruption time point, the time interval between each candidate interruption time point and the interruption occurrence time, and the interruption intent score, the interruption cost of each candidate interruption time point is determined; the interruption cost is positively correlated with the importance score, the time interval, and the interruption intent score. The candidate interrupt time point with the lowest interruption cost among the multiple candidate interrupt time points is taken as the target interrupt time point.

3. The method according to claim 2, characterized in that, The step of selecting multiple candidate interruption time points from the first speech based on the importance distribution information includes: If the interruption intent score is less than the second threshold, and the importance distribution information indicates that the importance score of the broadcast content corresponding to the interruption time is less than the third threshold, then based on the importance distribution information, multiple candidate interruption time points are selected from the first speech.

4. The method according to claim 2, characterized in that, The step of determining the target interruption time point based on the importance distribution information and the interruption intent score also includes: If the interruption intention score is less than the second threshold, and the importance score of the broadcast content corresponding to the interruption time is greater than or equal to the third threshold, then a target broadcast segment is determined from the first speech; the target broadcast segment is characterized as: in the first speech, a speech segment located after the interruption time, whose importance score is always greater than or equal to the third threshold; The end time of the broadcast of the target segment is taken as the target interruption time point.

5. The method according to claim 4, characterized in that, The step of using the end time of the broadcast of the target broadcast segment as the target interruption time point includes: If the duration of the target broadcast segment is less than the fourth threshold, then the end time of the broadcast of the target broadcast segment is taken as the target interruption time point.

6. The method according to claim 4, characterized in that, After determining the target broadcast segment from the first speech, the method further includes: If the duration of the target broadcast segment is greater than or equal to the fourth threshold, the target broadcast segment is divided into multiple broadcast sub-segments according to semantics. Multiple candidate interruption time points are selected from the connection points between the broadcast sub-segments; Based on the importance score corresponding to each candidate interruption time point, the time interval between each candidate interruption time point and the interruption occurrence time, and the interruption intent score, the interruption cost of each candidate interruption time point is determined respectively. The candidate interrupt time point with the lowest interruption cost among the multiple candidate interrupt time points is taken as the target interrupt time point.

7. The method according to claim 1 or 2, characterized in that, The step of determining the target interruption time point based on the importance distribution information and the interruption intent score also includes: If the interruption intent score is greater than or equal to the second threshold, the target time after the interruption occurs is taken as the target interruption time point; the time interval between the interruption occurrence time and the target time is less than the fourth threshold.

8. The method according to any one of claims 2 to 6, characterized in that, The candidate interruption time point is located within a time period starting from the interruption time and lasting for a fourth threshold.

9. An interrupt handling device, characterized in that, Applied to electronic devices, the device includes: The first evaluation module is used to evaluate the importance of the content of the first voice currently being broadcast by the electronic device, and to obtain the importance distribution information of the first voice as the broadcast process progresses. The second evaluation module is used to evaluate the interruption intention of the second voice in response to the second voice input by the user during the process of the electronic device playing the first voice, and to obtain an interruption intention score. The determination module is used to determine the target interruption time point based on the importance distribution information and the interruption intent score; The response module is used to interrupt the playback of the first voice at the target interruption time point and to trigger a response to the second voice.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the interrupt handling method as described in any one of claims 1 to 8.

Images