CN118155613A - Voice processing method, device, equipment and medium - Google Patents

Abstract

The embodiments of the present disclosure relate to a speech processing method, apparatus, device and medium, wherein the method comprises: in response to a received voice control instruction, determining a semantic service object corresponding to the voice control instruction and a first service type of the semantic service object; in the case where there are multiple first service types, determining a second service type of an application currently in an open state, and matching the second service type with the first service type to obtain a matching result; when the matching result is that there is a first target service type in the second service type that successfully matches the first service type, providing a speech service corresponding to the semantic service object through the application corresponding to the first target service type. In the embodiments of the present disclosure, in the case where there are multiple semantics, the correct semantics can be determined in the multiple semantics according to the service type of the open application, which reduces the interaction with the user for semantic clarification, reduces the time consumption of semantic clarification, and improves the speech processing efficiency.

Description

Voice processing method, device, equipment and medium

Technical Field

The present disclosure relates to the field of artificial intelligence technology, and in particular to a speech processing method, device, equipment and medium.

Background technique

Artificial Intelligence (AI) is a branch of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Research in this field includes robotics, language recognition, etc. With the development of artificial intelligence, speech recognition processing based on artificial intelligence has become a common way of interaction.

In the related art, when the user's voice control command has multiple semantics, clarification technology is often used. Through multiple interactions with the user, the user is allowed to select the correct semantics to complete the voice processing. For example, when the user's voice control command is "play Fleet of Time", the corresponding semantic service object may be a song type or a video type. Therefore, it is necessary to interact with the user again for semantic clarification. For example, the user is asked "Do you want to listen to the song or video of Fleet of Time?" When the user confirms "I want to listen to the song", the final semantic recognition result is determined to be "Play Fleet of Time song". The time-consuming process of interacting with the user to obtain the correct semantics leads to low efficiency of voice processing.

Summary of the invention

In order to solve the above technical problems or at least partially solve the above technical problems, the present disclosure provides a speech processing method, apparatus, device and medium. When there are multiple service types of the speech service object corresponding to the voice control instruction, the application corresponding to the voice control instruction is determined by matching the semantic service object corresponding to the voice control instruction with the service type of the opened application, so as to provide voice service based on the corresponding application, reduce the interaction with the user for semantic clarification, reduce the time consumption of semantic clarification, and improve the speech processing efficiency.

An embodiment of the present disclosure provides a voice processing method, the method comprising: in response to a received voice control instruction, determining a semantic service object corresponding to the voice control instruction and a first service type of the semantic service object; in the case where there are multiple first service types, determining a second service type of an application that is currently turned on, and matching the second service type with the first service type; in the case where there is a first target service type in the second service type that successfully matches the first service type, providing a voice service corresponding to the semantic service object through the application corresponding to the first target service type.

The present disclosure also provides a speech processing device, which includes: a determination module for determining, in response to a received voice control instruction, a semantic service object corresponding to the voice control instruction and a first service type of the semantic service object; a matching module for determining, when there are multiple first service types, a second service type of an application currently in an open state, and matching the second service type with the first service type; and a processing module for providing, when there is a first target service type in the second service type that successfully matches the first service type, a speech service corresponding to the semantic service object through an application corresponding to the first target service type.

An embodiment of the present disclosure also provides an electronic device, which includes: a processor; a memory for storing executable instructions of the processor; the processor is used to read the executable instructions from the memory and execute the instructions to implement the speech processing method provided in the embodiment of the present disclosure.

The embodiment of the present disclosure further provides a computer-readable storage medium, wherein the storage medium stores a computer program, and the computer program is used to execute the speech processing method provided by the embodiment of the present disclosure.

Compared with the prior art, the technical solution provided by the embodiments of the present disclosure has the following advantages:

The speech processing scheme provided by the embodiment of the present disclosure, in response to a received voice control instruction, determines the semantic service object corresponding to the voice control instruction and the first service type of the semantic service object, and in the case where there are multiple first service types, determines the second service type of the application currently in the open state, and matches the second service type with the first service type, and then, in the case where there is a first target service type in the second service type that successfully matches the first service type, provides the speech service corresponding to the semantic service object through the application corresponding to the first target service type. In the embodiment of the present disclosure, when there are multiple service types of the speech service object corresponding to the voice control instruction, the application corresponding to the voice control instruction is determined by matching the semantic service object corresponding to the voice control instruction with the service type of the open application, so as to provide the speech service based on the corresponding application, thereby reducing the interaction with the user for semantic clarification, reducing the time consumption of semantic clarification, and improving the speech processing efficiency.

BRIEF DESCRIPTION OF THE DRAWINGS

The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale.

FIG1 is a flow chart of a speech processing method provided by an embodiment of the present disclosure;

FIG2 is a flow chart of another speech processing method provided by an embodiment of the present disclosure;

FIG3 is a flow chart of another speech processing method provided by an embodiment of the present disclosure;

FIG4 is a schematic diagram of a speech processing scenario provided by an embodiment of the present disclosure;

FIG5 is a schematic diagram of another speech processing scenario provided by an embodiment of the present disclosure;

FIG6 is a schematic diagram of another speech processing scenario provided by an embodiment of the present disclosure;

FIG7 is a schematic diagram of the structure of a speech processing device provided by an embodiment of the present disclosure;

FIG8 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present disclosure.

Detailed ways

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and/or in parallel. In addition, the method embodiments may include additional steps and/or omit the steps shown. The scope of the present disclosure is not limited in this respect.

The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

In order to solve the above problems, the embodiments of the present disclosure provide a voice processing method, in which, when an application is opened, the application corresponding to the voice control instruction is determined according to the semantic service object corresponding to the voice control instruction and the service type of the opened application, so as to provide voice service based on the corresponding application, reduce the interaction of semantic clarification with the user, and improve the efficiency of voice processing. The method is introduced below in conjunction with a specific embodiment.

FIG1 is a flow chart of a voice processing method provided by an embodiment of the present disclosure. The method can be executed by a voice processing device, wherein the device can be implemented by software and/or hardware, and can generally be integrated in an electronic device, and the electronic device can be integrated in a vehicle or other equipment, so that the voice processing method in this embodiment can also provide voice processing services in a vehicle. As shown in FIG1 , the method includes:

Step 101: In response to a received voice control instruction, determine a semantic service object corresponding to the voice control instruction and a first service type of the semantic service object.

Among them, the semantic service object can be understood as the control object corresponding to the voice control instruction. For example, when the voice control instruction is "play Fleet of Time", the corresponding semantic service object is "Fleet of Time" and so on.

It should be noted that, in some possible embodiments, the voice information contained in the voice control instruction can be converted into text information through voice recognition technology, and the part of speech of each word in the text information is recognized, and the semantic service object is determined according to the recognition result, for example, the noun after the part of speech of the verb is determined as the voice service object, etc.

In some possible embodiments, a deep learning model can also be trained based on a large amount of sample data, the voice information contained in the voice control instruction can be converted into text information through voice recognition technology, and the text information can be input into a pre-trained deep learning model to obtain the semantic service object output by the deep learning model.

Further, the first service type of the voice service object is determined, wherein the first service type is used to indicate the possible information type given to the semantic service type. For example, when the semantic service object is the above-mentioned "Fleeting Years", since the information type corresponding to "Fleeting Years" may be "song" or "video", the first service type corresponding to "Fleeting Years" includes "song" and "video".

It should be noted that, in some possible embodiments, semantic recognition is performed on the voice control instruction to determine the semantic service object corresponding to the voice control instruction object, wherein the method of semantic recognition can refer to the above-mentioned embodiment. Further, a query request carrying the semantic service object is sent to the preset server, and the first service type of the semantic service object fed back by the preset server is obtained. The preset server can determine the first service type corresponding to the voice service object in the online data or the pre-stored offline data based on the semantic service object.

In some possible embodiments, the search results corresponding to the semantic service object can also be retrieved on a preset search platform, and the number of each candidate service type is determined according to the search results, and the candidate service type whose number belongs to the preset number is determined as the first service type. In the actual execution process, the ratio of the number of candidate service types whose number belongs to the preset number to the total search results can be further determined. When the ratio is less than a preset ratio threshold, the candidate service type of the object is not taken as the first service type. When the ratios corresponding to all candidate service types whose number belongs to the preset number are not greater than the preset ratio threshold, the candidate service object with the highest number is taken as the first service type, thereby further improving the accuracy of the first service type.

Step 102, when there are multiple first service types, determine the second service type of the application currently in the open state, and match the second service type with the first service type to obtain a matching result.

In one embodiment of the present disclosure, when there are multiple first service types, semantic clarification is not performed directly, but the second service type of the application that is currently in the open state is determined, for example, the second service type of the in-vehicle application that is currently running in the vehicle is determined, wherein the second service type indicates the application service type of the application, for example, for a song playing application, its corresponding second service type is "song", for example, for a video playing application, its object's second service type is "video", and so on.

In some possible embodiments, the program name of the application program currently in the open state may be obtained, and the pre-built corresponding relationship may be queried to obtain the second service type corresponding to the program name.

In this embodiment, after the second service type is determined, the service type is further matched with the first service type to determine whether there is currently an application that can directly provide a service corresponding to the voice control instruction.

Step 103: When the matching result is that there is a first target service type in the second service type that successfully matches the first service type, a voice service corresponding to the semantic service object is provided through an application corresponding to the first target service type.

In one embodiment of the present disclosure, when the matching result is that there is a first target service type in the second service type that successfully matches the first service type, a voice service corresponding to the semantic service object is provided through an application corresponding to the first target service type. That is, in the currently opened application, there is a first target service type that provides a second service type that matches the first service type of the semantic service object. At this time, the voice service is provided directly through the corresponding application. Not only can the application for the voice service be directly determined in the already opened applications, but voice processing can also be provided directly through the first target application. At this time, there is no need for semantic clarification, which reduces the number of interactions for semantic clarification and improves the efficiency of voice processing.

Among them, the voice service corresponding to the semantic service object is provided through the application corresponding to the first target service type mentioned in the embodiment of the present disclosure, which is to determine the multimedia resources corresponding to the semantic service object through the application corresponding to the first target service type, and play the multimedia resources according to the application corresponding to the first target service type. For example, when the semantic service object is "Fleeting Time", the corresponding multiple first service types are "songs" and "videos", and the second service type of the opened application is "video" and "map", then it is obvious that since the second service type "video" is consistent with the first service type "video", the playback service of "Fleeting Time" can be provided directly through the "video" application.

Of course, in the actual execution process, there may be multiple first target service types. In this case, the start time of the application corresponding to each first target service type can be determined, and the application corresponding to the first target service type with the most recent start time can be determined as the application providing voice service; or, the user's preference information can be obtained, and based on the user's preference information, the application corresponding to the first target service type that best meets the user's preference information among multiple first target service types can be determined as the application providing voice service.

In summary, the speech processing method of the embodiment of the present disclosure, in response to the received voice control instruction, determines the semantic service object corresponding to the voice control instruction and the first service type of the semantic service object, and in the case where there are multiple first service types, determines the second service type of the application currently in the open state, and matches the second service type with the first service type, and then, in the case where there is a first target service type in the second service type that successfully matches the first service type, the speech service corresponding to the semantic service object is provided by the application corresponding to the first target service type. In the embodiment of the present disclosure, when there are multiple service types of the speech service object corresponding to the voice control instruction, the application corresponding to the voice control instruction is determined by matching the semantic service object corresponding to the voice control instruction with the service type of the open application, so as to provide a speech service based on the corresponding application, that is, in the case of multiple semantics, the correct semantics can be determined in the multiple semantics according to the service type of the open application, which reduces the interaction with the user for semantic clarification, reduces the time consumption of semantic clarification, and improves the speech processing efficiency.

Based on the above embodiment, after the second service type is matched with the first service type, there may be other matching results. In the embodiment of the present disclosure, further semantic clarification processing is performed on the other matching results.

In one embodiment of the present disclosure, as shown in FIG2 , after matching the second service type with the first service type, the method further includes:

Step 201: When the matching result is that there is no first target service type in the second service type that successfully matches the first service type, semantic clarification prompt information is generated according to multiple first service types.

In an embodiment of the present disclosure, when the matching result is that there is no first target service type in the second service type that successfully matches the first service type, a semantic clarification prompt message is generated based on the first service type, wherein the semantic clarification prompt message generally includes a semantic service object and multiple first service types corresponding to the semantic service object.

For instance, continuing to use the example of the semantic service object "Fleeting Time", when there is no first target service type in the second service type that successfully matches the first service type, the voice clarification prompt message generated based on multiple first service types "songs" and "videos" can be "Do you want to listen to the song "Fleeting Time" or watch the video "Fleeting Time".

Step 202: Perform semantic clarification prompt processing according to the semantic clarification prompt information.

It should be noted that in different application scenarios, the methods of processing semantic clarification prompts are different according to the semantic clarification prompt information. For example, semantic playback of semantic clarification prompt information can be used; for example, it can be detected whether the current user is using the screen. When the user is using the screen, the semantic clarification prompt information can be displayed in pop-up text to perform semantic clarification prompt processing, etc.

Furthermore, in response to receiving a second target service type input according to the semantic clarification prompt information within a preset time period, an application corresponding to the second target service type is determined, wherein the second target service type may be input by voice input or by triggering an input such as a label of the first service type contained in the corresponding pop-up text.

Furthermore, in order to prevent the opened application corresponding to the second target service type from affecting driving safety, in an embodiment of the present disclosure, the current driving parameter information of the vehicle is also identified, wherein the driving parameter information includes driving speed information, driving road type information, etc., and the current driving safety level is determined according to the current driving parameter information. For example, the current driving parameter information is input into a pre-trained deep learning model to obtain the current driving safety level obtained by the deep learning model. Furthermore, the program security level of the application corresponding to the second target service type can be obtained by querying a preset corresponding relationship, etc. When the application security level matches the current driving safety level, for example, when the program security level is greater than or equal to the current driving safety level, it is determined that the currently running application corresponding to the second target service type will not bring safety hazards to driving safety. In this case, the voice service corresponding to the semantic service object is provided through the application corresponding to the second target service type, that is, the application corresponding to the second target service type is opened, the multimedia resource corresponding to the semantic service object is obtained through the application corresponding to the second target service type, and the application corresponding to the second target service type is opened to play the multimedia resource. Therefore, in this embodiment, semantic clarification processing is performed only when there is no first target service type in the second service type that successfully matches the first service type, and when there is a first target service type that matches the first service type in the second service type, semantic clarification processing is not performed, thereby reducing the number of interactions during semantic clarification.

In one embodiment of the present disclosure, it is possible that the second target service type input according to the semantic clarification prompt information is not received within a preset time period. In order to improve the intelligence of voice processing, the popularity information of each first service type can also be determined, wherein the popularity information can be determined based on the number of historical searches corresponding to each first service type of the semantic service object by the user, wherein the number of historical searches is proportional to the popularity information. Alternatively, the number of playbacks of the multimedia resources corresponding to each first service type of the semantic service object can also be determined, and the popularity information of each first service type can be determined based on the playback coefficient, wherein the popularity information is proportional to the number of playbacks.

Furthermore, a third target service type is determined among multiple first service types based on the popularity information, and an application corresponding to the third target service type is determined, wherein the third target service type is usually the service type with the highest popularity information. Then, the application corresponding to the third target service type is started to provide voice service corresponding to the semantic service object through the application corresponding to the third target service type.

In one embodiment of the present disclosure, when the first service type is single, a fourth target service type corresponding to the first service type is determined, and an application corresponding to the fourth target service type is determined, and a voice service corresponding to the semantic service object is provided through the application corresponding to the fourth target service type, wherein the fourth target service type may be an application that is currently opened, or may be an application that is not opened, and no limitation is made here, that is, when there is only one first service type, the corresponding application is directly opened to provide voice service without matching the service type of the opened application with the first service type, thereby further improving the voice processing efficiency.

In summary, the speech processing method of the embodiment of the present disclosure, when there is no first target service type in the second service type that successfully matches the first service type, generates semantic clarification prompt information according to multiple first service types, and performs semantic clarification prompt processing according to the semantic clarification prompt information, taking into account both speech processing efficiency and speech processing reliability.

Based on the above embodiments, it is easy to understand that when determining the semantic service object and the first service type corresponding to the voice control instruction, in addition to considering the semantic information of the semantic recognition result corresponding to the semantic control instruction itself, the relationship between the semantic word segmentations in the semantic recognition result can also be further explored, and the semantic service object and the first service type corresponding to the voice control instruction are determined based on the relationship between the semantic word segmentations, thereby improving the recognition accuracy of the semantic service object and the first service type corresponding to the voice control instruction.

In one embodiment of the present disclosure, as shown in FIG3 , determining a semantic service object corresponding to the voice control instruction and a first service type of the semantic service object includes:

Step 301, performing semantic recognition according to the voice control instruction to obtain a target semantic recognition result.

In one embodiment of the present disclosure, semantic recognition is performed according to the voice control instruction to obtain a target semantic recognition result, and the format of the semantic recognition result is a semantic recognition text.

Step 302: Identify multiple semantic segmentations included in the target semantic recognition result.

In this embodiment, multiple semantic segmentations included in the target semantic recognition result can be identified based on segmentation part-of-speech recognition and other methods. For example, when the target semantic recognition result is "I, today, think, play, those years passed in a hurry".

Step 303, construct a word segmentation label unit set based on multiple semantic word segmentations, wherein the word segmentation label unit set is a unit set consisting of row word segmentations and column word segmentations, wherein the row word segmentations and column word segmentations are set in the same manner according to the word segmentation order of the multiple semantic word segmentations in the target semantic recognition result.

In an embodiment of the present disclosure, in order to mine the relationship between semantic participles, a participle label unit set is constructed according to multiple semantic participles, wherein the participle label unit set is a unit set consisting of row participles and column participles, that is, if multiple phonetic participles correspond to n participles, a square table consisting of n participles can be constructed, the table includes n*n cells, and the n*n cells serve as sub-units to form a corresponding participle label unit set, wherein the row participles and the column participles are obtained by setting the same order of the participles in the target semantic recognition result of the multiple semantic participles. The same setting here means that they are all set according to the order of the participles in the corresponding target semantic recognition result. The forward and backward order setting is either set according to the order from back to front in the corresponding target semantic recognition result, or set according to the scattered order in the corresponding sample sentence, that is, it is necessary to ensure that the order of the corresponding characters in the row and column settings is the same, and ensure that the characters corresponding to the rows and columns with the same numbers are the same semantic word segmentations in the corresponding target semantic recognition result. For example, if multiple semantic word segmentations are "I, today, want, play, those years in a hurry", then as shown in Figure 4, a 5*5 table is constructed, and the rows and columns of the table correspond to "I, today, want, play, those years in a hurry", and the 25 table cells corresponding to the 5*5 table constitute a word segmentation label unit set.

Step 304: label each sub-unit consisting of row segmentation and column segmentation in the segmentation label unit set with a corresponding information category label.

Among them, each sub-unit composed of row segmentation and column segmentation in the segmentation label unit set is marked with a corresponding information category label. Therefore, each segmentation label unit set contains two dimensions of information. One dimension is the information category of the semantic segmentation (the semantic segmentation in the table is all the semantic segmentation, and there is no need for preliminary extraction of semantic segmentation. Whether the semantic segmentation has a corresponding information category is reflected in the table), and the other dimension is the information category between the semantic segmentations.

In some possible embodiments, the first segmentation attribute of the row segmentation and the second segmentation attribute of the column segmentation corresponding to each sub-unit in the segmentation label unit set can be determined, and the first segmentation attribute and the second segmentation attribute are input into a preset segmentation relationship extraction model to obtain the information category label of the corresponding sub-unit, wherein the segmentation relationship extraction model is pre-trained to extract the relationship between some segmentation attributes with semantic execution meaning (related to service objects and service types). For example, in some scenarios, segmentation relationship extraction is not performed for some segmentation attributes that have no semantic execution meaning. For example, for the segmentation relationship between segmentation attributes such as "想" and "想", the output result is "no information category".

For example, continuing with the scenario shown in Figure 4, as shown in Figure 5, the corresponding information category labels are marked in the cells corresponding to the characters in the corresponding rows and columns, and the cells in the table that do not have corresponding information categories are marked with the information category label corresponding to "no information category" (in the disclosed embodiment, ┴ can represent "no information category", and the "no information category" identifier corresponding to the row semantic segmentation and the column semantic segmentation has a small relationship in identifying the semantic service object and the semantic service type). The information category labels marked in different cells can be the same or different. For example, the information category marked in the cells corresponding to "I" and "me" is "no information category", the information category marked in the cells corresponding to "today and today" is "no information category", the information category marked in the cells corresponding to "I" and "play" is "subject-verb", the information category corresponding to "I" and "those years" is "subject-noun", etc.

Therefore, in addition to the semantic participle marking corresponding to cells with specific information categories, the semantic participles corresponding to cells without information categories can also be reflected. Therefore, there is no need to extract the corresponding semantic participles with information categories in advance. Based on the marking of the information category labels in the label table, each semantic participle and all semantic participles in related sentences are traversed, thereby ensuring the accuracy of information category extraction.

Step 305: Determine the semantic service object and the first service type of the semantic service object according to the information category label in the word segmentation label unit set.

In an embodiment of the present disclosure, when determining the word segmentation label unit set, the semantic service object and the first service type of the semantic service object are determined according to the information category label in the word segmentation label unit set.

It should be noted that, since the information label categories in the above-mentioned segmentation label unit set reflect the relationship between semantic segmentations, some segmentations that are meaningful to the execution of semantic commands can be quickly obtained by combining the relationship between semantic segmentations. For example, the semantic segmentations corresponding to information categories such as "noun-noun" can be determined to determine the corresponding semantic service objects, and the corresponding first service type can be determined according to the semantic segmentations corresponding to information categories such as "noun-verb" and "noun-noun".

In some possible embodiments, it can be understood that since the table construction corresponding to the word segmentation label unit set has some table attribute information, for example, when the row segmentation and column segmentation components are the same, the component segmentations are usually on the diagonal of the table. For example, the information categories related to the semantic service object are usually symmetrically distributed along the diagonal of the table. Therefore, the attribute information of this table is combined to further determine the information categories related to the identification of the semantic service object and the service type. Among them, in one embodiment of the present disclosure, the corresponding table attribute information can be reflected by a table feature vector. In this embodiment, the table feature vector corresponding to the table attribute information is determined according to the word segmentation label unit set. For example, the vector of the semantic segmentation pair composed of the row segmentation and the column segmentation corresponding to each information category corresponding to the word segmentation label unit set can be extracted. According to the vector of the semantic segmentation pair, the head vector of the semantic segmentation pair can be extracted using a multi-layer perceptron or the like. and tail vector/> In this embodiment, the head vector and tail vector/> The extraction formula is shown in formula (1), where, in formula (1), /> represents the real number field, d is the dimension of the eigenvector,

Furthermore, the head vector and the tail vector of the semantic word segmentation pair are combined into a combined vector, and the combined vectors of all semantic word segmentations are used as the corresponding table feature vector. Since the table feature vector is related to the information category label, the information category label that can be mined to correspond to the first service type or the semantic service object is relatively certain. For example, the corresponding semantic service object can be determined by the "noun-noun" and "verb-noun" information categories, and the table feature vectors obtained by the semantic word segmentation corresponding to these information categories are relatively consistent. Therefore, the table segmentation unit where the corresponding information category is located can be screened out through the table feature vector. The information label corresponding to each table segmentation unit belongs to a large category, and can cooperate with each other to determine the corresponding semantic service object or the first service type, etc. Therefore, the table segmentation position of the word segmentation label unit set is determined according to the table feature vector, and the word segmentation label is segmented according to the table segmentation position to obtain multiple table segmentation units, wherein the information category label corresponding to each table segmentation unit can better determine whether the corresponding semantic word segmentation can determine the corresponding semantic service object or the first service type.

For example, as shown in Figure 6, continuing with the scenario shown in Figure 5, the table can be divided into 6 table segmentation units according to the table feature vector, and the information category in each table segmentation unit belongs to a large category, where, for example, the information category label under the first table segmentation unit belongs to a large category, which is irrelevant to determining the service type and semantic service object.

Furthermore, the first prediction score belonging to the semantic service object in each table segmentation unit and the second prediction score belonging to the first service type can be determined according to the information category label contained in each table segmentation unit, so as to determine the semantic service object and the first service type according to the first prediction score and the second prediction score. Since the determination of the semantic service object and the first service type in this embodiment is not only determined at the granularity of the table segmentation unit, the degree of refinement of the determination is improved, wherein each table segmentation unit contains information labels that belong to a large category, and each table segmentation unit reflects the information category between the corresponding semantic word segmentations, which further improves the degree of refinement of the determination of the semantic service object and the first service type.

Among them, in some possible embodiments, a first number of first preset information categories belonging to the semantic service object in the information category label contained in each of the table segmentation units can be determined, and a first prediction score can be determined according to the first number, and a second number of second preset information categories belonging to the service type in the information category label contained in each table segmentation unit can be determined, and a second prediction score can be determined according to the second number; in some possible embodiments, the corresponding table feature vector in each table segmentation unit can be input into a preset convolutional neural network to obtain the information category label contained in each table segmentation unit to determine the first prediction score belonging to the semantic service object in each table segmentation unit, and the second prediction score belonging to the first service type, and then, the semantic service object and the first service type are determined according to the first prediction score and the second prediction branch. Type, for example, if the first prediction score is greater than the first preset score, the noun in the semantic segmentation corresponding to the table segmentation unit can be used as the first service object, and if the second prediction score is greater than the second preset score, the first candidate service type corresponding to the verb in the corresponding semantic segmentation in the corresponding table segmentation unit and the second candidate service type corresponding to the noun in the corresponding semantic segmentation are identified, and the intersection of the first candidate service type and the second candidate service type is used as the first service type. For example, when the first candidate service type determined according to the verb "play" in the corresponding semantic segmentation in the table segmentation unit is "song" and "video", and the second candidate service type corresponding to the noun "fleeting years" in the corresponding semantic segmentation is "song" and "video", the first service type is determined to be "song" and "video".

In summary, the speech processing method of the embodiment of the present disclosure performs semantic recognition according to the speech control instruction to obtain the target semantic recognition result, identifies multiple semantic participles contained in the target semantic recognition result, and determines the semantic service object and the first service type of the semantic service object based on the information category between the multiple semantic participles, thereby improving the refinement of the determination of the semantic service object and the first service type of the semantic service object, and ensuring the accuracy of the determination of the semantic service object and the first service type of the semantic service object. In order to implement the above embodiment, the present disclosure also proposes a speech processing device.

FIG7 is a schematic diagram of the structure of a speech processing device provided by an embodiment of the present disclosure. The device can be implemented by software and/or hardware and can generally be integrated into an electronic device. As shown in FIG7 , the device includes: a determination module 710, a matching module 720 and a processing module 730, wherein:

A determination module 710, configured to determine, in response to a received voice control instruction, a semantic service object corresponding to the voice control instruction and a first service type of the semantic service object;

A matching module 720, configured to determine a second service type of an application currently in an open state when there are multiple first service types, and match the second service type with the first service type to obtain a matching result;

The processing module 730 is used to provide a voice service corresponding to the semantic service object through an application corresponding to the first target service type when the matching result is that there is a first target service type in the second service type that successfully matches the first service type.

In one embodiment of the present disclosure, the determination module 710 is specifically configured to:

Performing semantic recognition on the voice control instruction to determine a semantic service object corresponding to the voice control instruction object;

A query request carrying the semantic service object is sent to a preset server, and a first service type of the semantic service object fed back by the preset server is acquired.

The speech processing device provided in the embodiments of the present disclosure can execute the speech processing method provided in any embodiment of the present disclosure, and has the corresponding functional modules and beneficial effects of the execution method.

In order to implement the above embodiments, the present disclosure further proposes a computer program product, including a computer program/instruction, which implements the speech processing method in the above embodiments when executed by a processor.

FIG8 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure.

8, which shows a schematic diagram of the structure of an electronic device 800 suitable for implementing the embodiment of the present disclosure. The electronic device 800 in the embodiment of the present disclosure may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. The electronic device shown in FIG8 is only an example and should not bring any limitation to the functions and scope of use of the embodiment of the present disclosure.

As shown in Figure 8, the electronic device 800 may include a processor (e.g., a central processing unit, a graphics processing unit, etc.) 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a memory 808 into a random access memory (RAM) 803. In the RAM 803, various programs and data required for the operation of the electronic device 800 are also stored. The processor 801, the ROM 802, and the RAM 803 are connected to each other via a bus 804. An input/output (I/O) interface 805 is also connected to the bus 804.

Typically, the following devices may be connected to the I/O interface 805: input devices 806 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 807 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 808 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 809. The communication device 809 may allow the electronic device 800 to communicate wirelessly or wired with other devices to exchange data. Although FIG. 8 shows an electronic device 800 with various devices, it should be understood that it is not required to implement or have all the devices shown. More or fewer devices may be implemented or have alternatively.

In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 809, or installed from the memory 808, or installed from the ROM 802. When the computer program is executed by the processor 801, the above-mentioned functions defined in the speech processing method of the embodiment of the present disclosure are executed.

It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device: in response to the received voice control instruction, determines the semantic service object corresponding to the voice control instruction and the first service type of the semantic service object. In the case where there are multiple first service types, determines the second service type of the application currently in the open state, and matches the second service type with the first service type. Then, in the case where there is a first target service type in the second service type that successfully matches the first service type, the application corresponding to the semantic service object is provided through the application corresponding to the first target service type. In an embodiment of the present disclosure, when there are multiple service types of the voice service object corresponding to the voice control instruction, the application corresponding to the voice control instruction is determined by matching the semantic service object corresponding to the voice control instruction with the service type of the opened application, so as to provide voice service based on the corresponding application. In the case of multiple semantics, the correct semantics can be determined in the multiple semantics according to the service type of the opened application, which reduces the interaction with the user for semantic clarification, reduces the time consumption of semantic clarification, and improves the efficiency of voice processing.

The electronic device may be written in one or more programming languages or a combination thereof to write computer program code for performing the operations of the present disclosure, including but not limited to object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and/or flow chart, and the combination of the square boxes in the block diagram and/or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

The units involved in the embodiments described in the present disclosure may be implemented by software or hardware, wherein the name of a unit does not, in some cases, constitute a limitation on the unit itself.

The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chip (SOCs), complex programmable logic devices (CPLDs), and the like.

In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other.

In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

Although the subject matter has been described in language specific to structural features and/or methodological logical actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims.

Claims (10)

1. A speech processing method, characterized in that it comprises the following steps: In response to the received voice control instruction, determining a semantic service object corresponding to the voice control instruction and a first service type of the semantic service object; In the case where there are multiple first service types, determining a second service type of an application currently in an open state, and matching the second service type with the first service type to obtain a matching result; When the matching result is that there is a first target service type in the second service type that successfully matches the first service type, a voice service corresponding to the semantic service object is provided through an application corresponding to the first target service type. 2. The method according to claim 1, wherein determining the semantic service object corresponding to the voice control instruction and the first service type of the semantic service object comprises: Performing semantic recognition according to the voice control instruction to obtain a target semantic recognition result; Identify multiple semantic segmentations included in the target semantic recognition result; Constructing a word segmentation label unit set according to the multiple semantic word segmentations, wherein the word segmentation label unit set is a unit set consisting of row word segmentations and column word segmentations, wherein the row word segmentations and column word segmentations are obtained by setting the same order of the word segmentations of the multiple semantic word segmentations in the target semantic recognition result; Labeling corresponding information category labels for each subunit consisting of the row segmentation and the column segmentation in the segmentation label unit set; The semantic service object and the first service type of the semantic service object are determined according to the information category label in the word segmentation label unit set. 3. The method according to claim 2, characterized in that the step of labeling each sub-unit composed of the row segmentation and the column segmentation in the segmentation label unit set with a corresponding information category label comprises: Determine a first segmentation attribute of a row segmentation and a second segmentation attribute of a column segmentation corresponding to each subunit in the segmentation tag unit set; The first word segmentation attribute and the second word segmentation attribute are input into a preset word segmentation relationship extraction model to obtain information category labels of corresponding sub-units. 4. The method according to claim 2 or 3, characterized in that the step of determining the semantic service object and the first service type of the semantic service object according to the information category label in the word segmentation label unit set comprises: Identify table attribute information of the word segmentation label unit set, determine a table feature vector corresponding to the table attribute information according to the word segmentation label unit set; determine a table segmentation position of the word segmentation label unit set according to the table feature vector, and segment the word segmentation label according to the table segmentation position to obtain a plurality of table segmentation units; Determine, according to the information category label contained in each of the table segmentation units, a first prediction score belonging to the semantic service object and a second prediction score belonging to the first service type in each of the table segmentation units; The semantic service object and the first service type are determined according to the first prediction score and the second prediction score, respectively. 5. The method according to claim 1, characterized in that after matching the second service type with the first service type to obtain a matching result, it further comprises: When the matching result is that there is no first target service type in the second service type that successfully matches the first service type, generating semantic clarification prompt information according to a plurality of the first service types; A semantic clarification prompt process is performed according to the semantic clarification prompt information. 6. The method according to claim 5, further comprising: In response to receiving a second target service type input according to the semantic clarification prompt information within a preset time period, determining an application corresponding to the second target service type; Identifying current driving parameter information of the vehicle, determining a current driving safety level according to the current driving parameter information, and determining a program safety level of an application corresponding to the second target service type; When the program safety level matches the current driving safety level, a voice service corresponding to the semantic service object is provided through an application corresponding to the second target service type. 7. The method according to claim 5 or 6, further comprising: In response to not receiving the second target service type input according to the semantic clarification prompt information within the preset time period, determining popularity information of each of the first service types; Determine a third target service type from the plurality of first service types according to the popularity information, and determine an application corresponding to the third target service type; An application corresponding to the third target service type is started to provide a voice service corresponding to the semantic service object through the application corresponding to the third target service type. 8. A speech processing device, comprising: A determination module, configured to determine, in response to a received voice control instruction, a semantic service object corresponding to the voice control instruction and a first service type of the semantic service object; a matching module, configured to determine, when there are multiple first service types, a second service type of an application currently in an open state, and match the second service type with the first service type to obtain a matching result; A processing module is used to provide a voice service corresponding to the semantic service object through an application corresponding to the first target service type when the matching result is that there is a first target service type in the second service type that successfully matches the first service type. 9. An electronic device, characterized in that the electronic device comprises: A processor; a memory for storing instructions executable by the processor; The processor is used to read the executable instructions from the memory and execute the executable instructions to implement the speech processing method described in any one of claims 1-7. 10. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, and the computer program is used to execute the speech processing method described in any one of claims 1 to 7.