CN116431316A

CN116431316A - Task processing method, system, platform and automatic question-answering method

Info

Publication number: CN116431316A
Application number: CN202310666870.1A
Authority: CN
Inventors: 刘跃辉; 陈颖达; 周文猛; 贾实聪
Original assignee: Alibaba China Co Ltd
Current assignee: Alibaba China Co Ltd
Priority date: 2023-06-06
Filing date: 2023-06-06
Publication date: 2023-07-14
Anticipated expiration: 2043-06-06
Also published as: CN116431316B

Abstract

The embodiment of the specification provides a task processing method, a system, a platform and an automatic question-answering method, wherein the task processing method comprises the following steps: receiving task data to be processed, which is input through a task interface, wherein the task interface is used for receiving task data of different tasks, packaging the task data to be processed according to the data types of the task data to be processed to obtain a target task request aiming at a target task, and scheduling a task processing model to execute the target task on the task data to be processed based on the target task request to obtain a task processing result, wherein the task processing model is a deep learning model which is trained in advance based on sample data of different tasks. The development of multiple interfaces to receive data is avoided, the development cost is reduced, a target task request aiming at a target task is obtained through encapsulation according to the data type of the task data to be processed to schedule a task processing model, the processing task is realized through the unified interface, the universality of task processing is enhanced, and the user experience is improved.

Description

Task processing method, system, platform and automatic question-answering method

Technical Field

The embodiment of the specification relates to the technical field of data processing, in particular to a task processing method.

Background

With the development of computer technology, different tasks in different fields can be executed based on task processing models obtained by training sample data of different tasks, for example, a generated task processing model can be executed, and natural language processing, image generation, audio generation, machine translation, abstract generation, recommendation systems and the like can be executed.

At present, the task processing model integrates various task processing functions, has a large scale, occupies a large amount of software and hardware resources, is difficult to be directly scheduled by a user, and needs to be deployed on a task processing platform. And an application programming interface on the task processing platform dispatches the task processing model to execute tasks on the data when receiving the data of the user. However, the data corresponds to different types, the data of different types corresponds to different tasks, a plurality of corresponding application programming interfaces need to be developed to receive the data of the corresponding tasks, a task processing model is scheduled to execute the corresponding tasks on the data, development cost is increased, a user is required to input corresponding task requests through each task interface, and under the conditions that the interface naming is not uniform and the task processing logic is not uniform, the universality and the user experience of task processing are reduced. Therefore, a task processing method with low development cost, high versatility and high user experience is needed.

Disclosure of Invention

In view of this, the present embodiment provides a task processing method. One or more embodiments of the present specification relate to another task processing method, an automatic document method, a task processing device, another task processing device, an automatic question answering device, a task processing system, a task processing platform, a computing device, a computer-readable storage medium, and a computer program, to solve the technical defects existing in the prior art.

According to a first aspect of embodiments of the present disclosure, there is provided a task processing method, applied to a task processing platform, including:

receiving task data to be processed, which is input through a task interface, wherein the task interface is used for receiving task data of different tasks;

according to the data type of the task data to be processed, the task data to be processed is packaged, and a target task request aiming at a target task is obtained;

based on the target task request, a task processing model is scheduled to execute a target task on task data to be processed to obtain a task processing result, wherein the task processing model is a deep learning model which is obtained by training in advance based on sample data of different tasks.

According to a second aspect of embodiments of the present disclosure, there is provided a task processing method, applied to a client, including:

the method comprises the steps of sending input task data to be processed to a task processing platform, wherein the task processing platform receives the task data to be processed through a task interface, and the task interface is used for receiving task data of different tasks;

and receiving a task processing result fed back by the task processing platform, wherein the task processing result is obtained by executing a target task on task data to be processed by a task processing scheduling model based on a target task request, and the target task request is obtained by packaging the task data to be processed according to the data type of the task data to be processed.

According to a third aspect of embodiments of the present disclosure, there is provided an automatic question-answering method applied to a task processing platform, including:

receiving problem data input through a task interface, wherein the task interface is used for receiving task data of different tasks;

according to the data type of the question data, the question data is packaged to obtain a target task request aiming at the question-answering task;

based on the target task request, a task processing model is scheduled to execute a question-answer task on the question data to obtain answer data, wherein the task processing model is a deep learning model which is obtained by training in advance based on sample data of different tasks.

According to a fourth aspect of embodiments of the present specification, there is provided a task processing device, applied to a task processing platform, including:

the first data receiving module is configured to receive task data to be processed, which is input through a task interface, wherein the task interface is used for receiving task data of different tasks;

the first packaging module is configured to package the task data to be processed according to the data type of the task data to be processed to obtain a target task request aiming at a target task;

the first processing module is configured to execute target tasks on task data to be processed based on target task requests, and obtain task processing results, wherein the task processing model is a deep learning model which is obtained by training sample data of different tasks in advance.

According to a fifth aspect of embodiments of the present specification, there is provided a task processing device, applied to a client, including:

the task processing platform receives the task data to be processed through a task interface, and the task interface is used for receiving task data of different tasks;

The result receiving module is configured to receive a task processing result fed back by the task processing platform, wherein the task processing result is obtained by executing a target task on task data to be processed by the task processing scheduling model based on a target task request, and the target task request is obtained by packaging the task data to be processed according to the data type of the task data to be processed.

According to a sixth aspect of embodiments of the present disclosure, there is provided an automatic question-answering apparatus, applied to a task processing platform, including:

the second data receiving module is configured to receive problem data input through a task interface, wherein the task interface is used for receiving task data of different tasks;

the second packaging module is configured to package the problem data according to the data type of the problem data to obtain a target task request aiming at the question-answering task;

and the second processing module is configured to schedule a task processing model to execute a question-answer task on the question data based on the target task request to obtain answer data, wherein the task processing model is a deep learning model which is trained in advance based on sample data of different tasks.

According to a seventh aspect of embodiments of the present specification, there is provided a task processing system comprising a client and a task processing platform;

The client is used for sending the input task data to be processed to the task processing platform;

the task processing platform is used for receiving task data to be processed, which are input through a task interface, wherein the task interface is used for receiving task data of different tasks, packaging the task data to be processed according to the data type of the task data to be processed to obtain a target task request aiming at a target task, scheduling a task processing model to execute the target task on the task data to be processed based on the target task request to obtain a task processing result, and feeding back the task processing result to the client, wherein the task processing model is a deep learning model which is trained in advance based on sample data of different tasks;

and the client is also used for receiving the task processing result fed back by the task processing platform.

According to an eighth aspect of embodiments of the present disclosure, there is provided a task processing platform, including a request end and a server end;

the request end is used for receiving task data to be processed, which are input by the client through the task interface, wherein the task interface is used for receiving task data of different tasks, packaging the task data to be processed according to the data type of the task data to be processed to obtain a target task request aiming at a target task, and processing the target task request to the server;

The server side is used for receiving the target task request sent by the request side, and scheduling a task processing model to execute a target task on task data to be processed to obtain a task processing result, wherein the task processing model is a deep learning model which is obtained by training sample data of different tasks in advance.

According to a ninth aspect of embodiments of the present specification, there is provided a computing device comprising:

a memory and a processor;

the memory is configured to store computer-executable instructions that, when executed by the processor, perform the steps of the method described above.

According to a tenth aspect of the embodiments of the present description, there is provided a computer-readable storage medium storing computer-executable instructions which, when executed by a processor, implement the steps of the above-described method.

According to an eleventh aspect of the embodiments of the present specification, there is provided a computer program, wherein the computer program, when executed in a computer, causes the computer to perform the steps of the above method.

In one or more embodiments of the present disclosure, task data to be processed is received, where the task data is input through a task interface, the task interface is configured to receive task data of different tasks, encapsulate the task data to be processed according to a data type of the task data to be processed, obtain a target task request for a target task, schedule a task processing model to execute the target task on the task data to be processed based on the target task request, and obtain a task processing result, where the task processing model is a deep learning model that is trained in advance based on sample data of the different tasks. The task interface is used for receiving the task data of different tasks, and receiving the input task data to be processed, so that development of a plurality of application programming interfaces to receive the data of the corresponding tasks is avoided, development cost is reduced, the task data to be processed is correspondingly packaged according to the data types of the task data to be processed, a target task request aiming at a target task is obtained, a task processing model is scheduled based on the target task request, the task processing model is scheduled to execute the corresponding processing tasks on the task data to be processed of a plurality of data types through a unified interface, universality of task processing is enhanced, and user experience is improved.

Drawings

FIG. 1 is a flow chart of a method of task processing provided in one embodiment of the present disclosure;

FIG. 2 is a flow chart of another task processing method provided by one embodiment of the present disclosure;

FIG. 3 is a flow chart of an automatic question-answering method provided by one embodiment of the present disclosure;

FIG. 4 is a schematic architecture diagram of a task processing platform in a task processing method according to an embodiment of the present disclosure;

FIG. 5 is a data flow diagram of a task processing platform in a task processing method according to an embodiment of the present disclosure;

FIG. 6 is a schematic flow chart of a task synchronous processing method according to an embodiment of the present disclosure;

FIG. 7 is a schematic flow chart of an asynchronous task processing in a task processing method according to an embodiment of the present disclosure;

FIG. 8 is a schematic flow chart of real-time task processing in a task processing method according to an embodiment of the present disclosure;

FIG. 9 is a process flow diagram of a task processing method applied to speech transcription according to one embodiment of the present disclosure;

FIG. 10 is a schematic diagram of a task processing device according to one embodiment of the present disclosure;

FIG. 11 is a schematic diagram of another task processing device according to an embodiment of the present disclosure;

Fig. 12 is a schematic structural diagram of an automatic question answering device according to one embodiment of the present disclosure;

FIG. 13 is a schematic diagram of a task processing system according to one embodiment of the present disclosure;

FIG. 14 is a schematic diagram of a task processing platform according to one embodiment of the present disclosure;

FIG. 15 is a block diagram of a computing device provided in one embodiment of the present description.

Detailed Description

In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present description. This description may be embodied in many other forms than described herein and similarly generalized by those skilled in the art to whom this disclosure pertains without departing from the spirit of the disclosure and, therefore, this disclosure is not limited by the specific implementations disclosed below.

The terminology used in the one or more embodiments of the specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the one or more embodiments of the specification. As used in this specification, one or more embodiments and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used in one or more embodiments of the present specification refers to and encompasses any or all possible combinations of one or more of the associated listed items.

It should be understood that, although the terms first, second, etc. may be used in one or more embodiments of this specification to describe various information, these information should not be limited by these terms. These terms are only used to distinguish one type of information from another. For example, a first may also be referred to as a second, and similarly, a second may also be referred to as a first, without departing from the scope of one or more embodiments of the present description. The word "if" as used herein may be interpreted as "at … …" or "at … …" or "responsive to a determination", depending on the context.

Furthermore, it should be noted that, user information (including, but not limited to, user equipment information, user personal information, etc.) and data (including, but not limited to, data for analysis, stored data, presented data, etc.) according to one or more embodiments of the present disclosure are information and data authorized by a user or sufficiently authorized by each party, and the collection, use, and processing of relevant data is required to comply with relevant laws and regulations and standards of relevant countries and regions, and is provided with corresponding operation entries for the user to select authorization or denial.

Large models refer to deep-learning models with large-scale model parameters, typically comprising hundreds of millions, billions, trillions, or even more than billions of model parameters. The large Model can be called as a Foundation Model, a training Model is performed by using a large-scale unlabeled corpus, a pre-training Model with more than one hundred million parameters is produced, the Model can adapt to a wide downstream task, and the Model has better generalization capability, such as a large-scale language Model (Large Language Model, LLM), a multi-mode pre-training Model and the like.

When the large model is actually applied, the pretrained model can be applied to different tasks by only slightly adjusting a small number of samples, the large model can be widely applied to the fields of natural language processing (Natural Language Processing, NLP for short), computer vision and the like, and particularly can be applied to the tasks of the computer vision fields such as visual question and answer (Visual Question Answering, VQA for short), image description (IC for short), image generation and the like, and the tasks of the natural language processing fields such as emotion classification based on texts, text abstract generation, machine translation and the like, and main application scenes of the large model comprise digital assistants, intelligent robots, searching, online education, office software, electronic commerce, intelligent design and the like.

First, terms related to one or more embodiments of the present specification will be explained.

The software development kit (SDK, software Development Kit) is part of a software development kit that provides the tools and programming interfaces required by software developers to assist them in building and debugging applications, as well as for scheduling of portions of the functionality in the software. The method can be understood as a part of functional modules of the application program, and the scheduling of the software development kit is finished through an application program programming interface to the outside, so that the corresponding function is realized.

Application programming interface (API, application Programming Interface): is a specification and interface for exchanging data and information provided by the application program. An application programming interface is typically made up of a specific set of rules and protocols that define how an application communicates with other applications or operating systems. In an application programming interface, applications can exchange data and information by sending requests and receiving responses. Application programming interfaces typically involve some type and structure of data, such as type, format, and encoding of request and response data. The application programming interface may also define access rights and security measures for the data to ensure that only authorized applications can access and use the data. Generally, an application programming interface corresponds to a software development kit, and the scheduling of the software development kit is realized.

Hypertext transfer protocol (HTTP, hyper Text Transfer Protocol): is a protocol for transmitting data over the Web. HTTP uses a request line, a request header, and a request body to transfer data. The request line includes a request method (GET, POST, etc.), a request URL, and a protocol version. The request header includes information of the client (browser). The transmission is typically carried out using the TCP protocol, running on the application layer.

WebSocket transport protocol: a protocol for full duplex communication over a single TCP connection. The WebSocket transport protocol allows a server to actively push data to a client. In the application programming interface of WebSocket, the browser and the server only need to complete one handshake, and can directly establish persistent connection between the two, perform full duplex data transmission and operate on a network layer.

Large model (large-scale deep learning model): deep learning models with high complexity and a variety of different functions. Typically consisting of multiple layers and may be trained with large amounts of sample data. Large models are commonly used for many different tasks, such as image classification, natural language processing, speech recognition, etc. These models may be constructed using deep learning techniques, such as convolutional neural networks and recurrent neural networks. Building, training, and running a large model requires a significant amount of hardware and software resources, and therefore, large models are typically deployed on remote platforms that are configured with a significant amount of hardware and software resources.

At present, as the task processing model is deployed on the task processing platform and corresponds to different functions of the task processing model, a software development kit is constructed, a corresponding application programming interface is arranged to receive a task request containing task data to be processed, scheduling of the task processing model is achieved, corresponding tasks are executed, for example, the task processing model deployed on the task processing platform can achieve three functions of text generation, text classification and emotion analysis, 3 software development kits are correspondingly constructed, 3 corresponding application programming interfaces are arranged, and text generation software development kit of the task processing model is scheduled to execute text generation on a subject text based on the text generation task request under the condition that the text generation interface receives the text generation task request containing the subject text. And under the condition that the emotion analysis interface receives an emotion analysis task request containing a text to be analyzed, scheduling an emotion analysis software development kit of a task processing model based on the emotion analysis task request, and executing emotion analysis on the text to be analyzed.

However, such a manner requires the development of multiple corresponding application programming interfaces to receive data of corresponding tasks, increasing development costs, and reducing versatility and user experience of task processing in the case of non-uniform interface naming and non-uniform task processing logic.

In view of the above problems, the present specification provides a task processing method, and relates to another task processing method, an automatic question answering method, a task processing device, another task processing device, an automatic question answering device, a task processing system, a task processing platform, a computing device, a computer-readable storage medium, and a computer program, which are described in detail in the following embodiments one by one.

Referring to fig. 1, fig. 1 shows a flowchart of a task processing method according to an embodiment of the present disclosure, where the method is applied to a task processing platform, and includes the following specific steps:

step 102: and receiving task data to be processed, which is input through a task interface, wherein the task interface is used for receiving task data of different tasks.

The task processing platform is a remote platform deployed with a task processing model, and a large amount of software and hardware resources are configured on the task processing platform so as to ensure the normal operation of the task processing model. The task processing platform externally provides an application programming interface for receiving task data and realizing the scheduling of the task processing model. The task processing platform may be a single platform or a distributed platform, which is not limited herein.

The task interface is a unified application programming interface arranged on the task processing platform, and the task interface is used for receiving task data of different tasks, wherein the different tasks correspond to functions of the task processing model, namely the tasks which can be executed, and the functions of the task processing model comprise: the task interface is used for receiving the subject text of the text generation task, the question text of the text question-and-answer task, the picture description text of the picture generation task, the audio data of the voice transcription task and the text content and other contents of the text embedding task.

The task data to be processed is the task data to be processed of the target task, and the task data to be processed can be divided into a plurality of data types according to data modalities, including but not limited to: character strings, files (text files, picture files, audio files, and video files), real-time streaming data (real-time audio streams and real-time video streams). The task data to be processed has multiple data types according to the corresponding target task, for example, the target task is a text generating task, and the text of the task data to be processed is a subject text. For another example, the target task is a picture generation task, and the task data to be processed is a picture description text.

The method for receiving the task data to be processed, which is input through the task interface, comprises the following specific steps: and receiving an initial task request input through a task interface, wherein the initial task request carries task data to be processed.

It should be noted that, the task interface is an application programming interface provided for performing task processing model scheduling, and the following

steps

104 and 106 perform corresponding encapsulation and model scheduling execution, which is transparent to the user, and the user can only perceive that the task request carrying the task data to be processed is sent to the task processing platform, so as to obtain the task processing result.

The task processing platform is a remote platform deployed with the large model, and a unified task interface is arranged on the task processing platform: the unified API receives an initial task request of a text generation task input by a client through the unified API, wherein the initial task request carries a subject text: a text of the prose style for rainy days.

And receiving the task data to be processed, which is input through the task interface, wherein the task interface is used for receiving task data of different tasks, and receiving the input task data to be processed through the task interface used for receiving task data of different tasks, so that the development of a plurality of application programming interfaces to receive data of corresponding tasks is avoided, the development cost is reduced, and meanwhile, a foundation is laid for obtaining a target task request through subsequent encapsulation.

Step 104: and packaging the task data to be processed according to the data type of the task data to be processed to obtain a target task request aiming at the target task.

Generally, different types of task data to be processed have different scheduling modes for task processing models, for example, the task data to be processed is a file, and the task processing models are scheduled to process the task data to be processed, so that corresponding task processing results can be obtained after the file content is extracted and the task processing results are executed. For example, the task data to be processed is real-time stream data, the task processing model is scheduled to process the task data, the task processing model is required to be a scheduling mode of inputting and outputting simultaneously, and the corresponding task processing result can be obtained without executing all the real-time stream data. Therefore, the task data to be processed needs to be encapsulated according to the data type, so that the corresponding target task request is obtained.

The data type of the task data to be processed is a data modality type of the task data to be processed, including but not limited to: character string, file, real-time streaming data. For example, the task data to be processed input by the user is text "please simply introduce a communication protocol", and the data type of the task data to be processed is a character string. For another example, the task data to be processed input by the user is table data, and the data type of the task data to be processed is a file. Also, for example, the task data to be processed input by the user is a piece of audio data, and the data type of the task data to be processed is real-time streaming data.

The target task request for the target task is a request scheduling instruction of a model function determined for the target task, and can be understood as a request scheduling instruction of a software development kit of the target task function on a task processing model. The target task request contains a scheduling mode and a scheduling protocol of task data to be processed corresponding to the task processing model. For example, the task processing model completes scheduling of the character string according to the half-duplex scheduling mode through the hypertext transfer protocol, and the target task request comprises the half-duplex scheduling mode and the hypertext transfer protocol.

According to the data type of the task data to be processed, the task data to be processed is packaged to obtain a target task request aiming at a target task, and the specific mode is as follows: and determining a corresponding encapsulation rule according to the data type of the task data to be processed, and encapsulating the task data to be processed by adopting the encapsulation rule to obtain a target task request aiming at the target task. The packaging rule is a data packaging rule and comprises a scheduling mode and a scheduling protocol of task processing models corresponding to task data to be processed.

Illustratively, according to the subject text: data type of "text in a prose style on rainy days": character strings, determining corresponding encapsulation rules: and (3) a hypertext transfer protocol, a half-duplex scheduling mode (synchronous scheduling mode or asynchronous scheduling mode), and packaging a text subject of 'text in a text style about rainy days' by adopting the packaging rule to obtain a target task request aiming at a text generation task.

According to the data types of the task data to be processed, the task data to be processed is packaged to obtain a target task request aiming at a target task, and the targeted packaging of the data types to be executed of various data types is realized through a unified interface, so that a foundation is laid for a subsequent scheduling task processing model.

Step 106: based on the target task request, a task processing model is scheduled to execute a target task on task data to be processed to obtain a task processing result, wherein the task processing model is a deep learning model which is obtained by training in advance based on sample data of different tasks.

The task processing model is a deep learning model with different functions, different tasks can be executed to realize different functions, and the task processing model is scheduled based on task requests to execute corresponding tasks. The task processing model may be a large model integrating multiple functions, or may be a deep learning model composed of distributed small models with different functions, which is not limited herein. The task processing model is obtained based on sample data of different tasks through pre-training, for example, based on sample data of a text generation task, sample data of a picture generation task and sample data of a voice transcription task, the task processing task is pre-trained to obtain a task processing model with text generation, picture generation and voice transcription functions, and the text generation task, the picture generation task and the voice transcription task can be executed.

The task processing result is a processing result of a target task, and corresponds to different tasks, the task processing result has different modes, for example, for a text generating task, the task processing result is a character string text. For another example, for an image generation task, the task processing result is a picture file.

Based on the target task request, the task processing model is scheduled to execute a target task to be processed task data to obtain a task processing result, and the specific mode is as follows: and executing the target task on the task data to be processed by the scheduling task processing model based on the scheduling mode of the target task request to obtain a task processing result.

By way of example, based on the scheduling mode "half duplex scheduling mode" of the target task request, the scheduling task processing model performs a text generation task on the topic text "text of the text style about rainy days" in the half duplex scheduling mode, and obtains a text generation result: the rainy days are natural artworks, and are full of the strength and beauty of life. In rainy days, people can hear the sound of raindrops falling on the ground and feel the warmth and freshness of the raindrops.

In the embodiment of the specification, task data to be processed is received, wherein the task data to be processed is input through a task interface, the task interface is used for receiving task data of different tasks, the task data to be processed is packaged according to the data type of the task data to be processed, a target task request aiming at a target task is obtained, a task processing model is scheduled to execute the target task on the task data to be processed based on the target task request, and a task processing result is obtained, and the task processing model is a deep learning model which is trained in advance based on sample data of the different tasks. The task interface is used for receiving the task data of different tasks, and receiving the input task data to be processed, so that development of a plurality of application programming interfaces to receive the data of the corresponding tasks is avoided, development cost is reduced, the task data to be processed is correspondingly packaged according to the data types of the task data to be processed, a target task request aiming at a target task is obtained, a task processing model is scheduled based on the target task request, the task processing model is scheduled to execute the corresponding processing tasks on the task data to be processed of a plurality of data types through a unified interface, universality of task processing is enhanced, and user experience is improved.

In an alternative embodiment of the present disclosure, step 104 includes the following specific steps:

and packaging the task data to be processed according to the data type of the task data to be processed and the task type of the target task to obtain a target task request aiming at the target task.

Generally, the data types of the task data to be processed and the task types of the target tasks have correlation, but the data types of the task data to be processed are not strictly in one-to-one correspondence, and the task data to be processed can be more accurately packaged by combining the task types of the target tasks on the basis of packaging according to the data types of the task data to be processed. For example, the task data to be processed is a text string, the target task is a text generating task, the scheduling mode of the scheduling task processing model can be a full duplex scheduling mode or a half duplex scheduling mode, all the text generating results can be obtained and then output, and the text generating results can also be generated and output word by word, so that the task data to be processed needs to be packaged more specifically according to the task type of the target task on the basis.

The task type of the target task is the type of task execution mode of the target task, including but not limited to: synchronous processing tasks, asynchronous processing tasks, and real-time processing tasks. The synchronous processing task is a processing task in a half duplex mode, and the scheduling mode is to take a target processing request as input and directly feed back the processing result of the task. The asynchronous processing task is a processing task in a half duplex mode, the scheduling mode is to take a target processing request as input, feedback is not directly carried out after a task processing result is obtained, result inquiry can be received, the task processing result can be fed back only when processing is completed, and the corresponding processing progress can be fed back when the processing is not completed. The real-time processing task is a processing task of a full duplex mode, namely, the request end and the service end can mutually and actively send data information to each other, and in the task execution process, the service end actively feeds back the obtained task processing result to the request end.

According to the data type of the task data to be processed and the task type of the target task, the task data to be processed is packaged to obtain a target task request aiming at the target task, and the specific mode is as follows: and determining a corresponding encapsulation rule according to the data type of the task data to be processed and the task type of the target task, and encapsulating the task data to be processed by adopting the encapsulation rule to obtain a target task request aiming at the target task.

Illustratively, according to the subject text: the data type "character string" of the "text of the text style about the rainy day" and the task type "synchronous processing task" of the text generation task, determine the corresponding encapsulation rule: and (3) a hypertext transfer protocol, a half-duplex scheduling mode (synchronous scheduling mode), and packaging a theme text 'text in a text style about rainy days' by adopting a packaging rule to obtain a target task request aiming at a text generation task.

In the embodiment of the specification, more targeted encapsulation of the data types to be executed of multiple data types is realized, and a foundation is laid for a subsequent scheduling task processing model.

In an alternative embodiment of the present disclosure, a task processing platform includes a request end;

Correspondingly, according to the data type of the task data to be processed and the task type of the target task, the task data to be processed is packaged to obtain a target task request aiming at the target task, and the method comprises the following specific steps:

under the condition that the task type of the target task is a synchronous processing task, the request end extracts data information to be synchronous in task data to be processed;

and packaging the information of the data to be synchronized to obtain a target task request aiming at the synchronous processing task.

The request end is a virtual end for data encapsulation and request transmission on the task processing platform. The request end is configured with a software development kit, wherein the software development kit completes the encapsulation and the transmission of the task data to be processed through extraction, encapsulation, transmission and other routines, and the task processing model is scheduled based on a target task request obtained through encapsulation.

The synchronous processing task is a processing task in a half duplex mode, and the scheduling mode is to take a target processing request as input and directly feed back the processing result of the task. For example, for a text generation task, the task type is determined to be a synchronous processing task, i.e., feedback is directly performed after the target text is generated. For another example, for an image generation task, the task type is determined to be a synchronous processing task, i.e., feedback is directly performed after the target image is generated.

The to-be-synchronized data information is data information for executing synchronization processing in to-be-processed task data, and the to-be-processed task data comprises the to-be-synchronized data information and other information, for example, the to-be-processed task data is a String of ' String a= ' abcdefg '; and the character String 'String A=' used for assignment is other information, and 'abcdefg' is data information to be synchronized. Only "abcdefg" needs to be encapsulated. For another example, the task data to be extracted is a file B, where a sub-file B1 is data information to be synchronized.

In an exemplary case that the task type of the text generation is a synchronous processing task, the request end extracts to-be-synchronized data information of 'text of a text style related to a text of a hydrological weather' in a subject text, and encapsulates the subject text to obtain a text generation request aiming at the synchronous processing task.

In the embodiment of the specification, the encapsulation of the data to be processed is completed through the encapsulation rule corresponding to the synchronous processing task, and a foundation is laid for the subsequent synchronous scheduling of the task processing model.

In an optional embodiment of the present disclosure, the task processing platform further includes a server;

accordingly, step 106 includes the following specific steps:

The server receives a target task request sent by the request end, the scheduling task processing model executes a synchronous processing task on the task data to be processed, and a task processing result of the synchronous processing task is fed back to the request end.

The server is a virtual end with a task processing model specifically deployed on the task processing platform. And scheduling corresponding functional modules on the task processing model on the server side through the software development package on the request side, and executing corresponding target tasks. For example, the task processing model is a distributed task processing model, the text generating function and the image generating function are distributed on different hardware, and the target task is scheduled to be executed by sending a target processing request to a corresponding functional module.

The scheduling task processing model executes synchronous processing tasks on task data to be processed, and the specific mode is as follows: and the corresponding functional module of the scheduling task processing model executes synchronous processing tasks on the task data to be processed.

The server receives a text generation request sent by the request end, and schedules a text generation function module of the task processing model to execute a text generation task of synchronous generation on a topic text 'text of a text style related to a rainy day', so as to obtain a text generation result: the rainy days are natural artworks, and are full of the strength and beauty of life. In rainy days, people can hear the sound of raindrops falling on the ground, feel the softness and freshness of the raindrops, and feed back the text generation result to the request end.

In the embodiment of the specification, synchronous execution of synchronous processing tasks is realized, the universality of task processing is further enhanced, and the user experience is further improved.

and under the condition that the task type of the target task is an asynchronous processing task, the request end packages the complete information of the task data to be processed to obtain a target task request aiming at the asynchronous processing task.

The asynchronous processing task is a processing task in a half duplex mode, the scheduling mode is to take a target processing request as input, feedback is not directly carried out after a task processing result is obtained, result inquiry can be received, the task processing result can be fed back only when processing is completed, and the corresponding processing progress can be fed back when the processing is not completed. For example, for an audio transcription task, the task type is determined to be an asynchronous processing task, namely, a transcribed text obtained by transcription is fed back after transcription is completed, and under the condition that transcription is not completed, a result query is received, and the transcription progress is fed back. For another example, for a text embedding task, the task type is determined to be an asynchronous processing task, that is, after transcription is completed, the multi-mode content file is obtained through feedback embedding, and under the condition that the embedding is not completed, a result query is received, and the embedding progress is fed back.

The complete information is data information for executing asynchronous processing in the task data to be processed, and the complete information comprises complete task processing information in the task data to be processed. For example, the task data to be extracted is a file C, where each sub-file is complete information.

In an exemplary case that the task type of the image generation task is an asynchronous processing task, the request end extracts the completion information of the image file, performs image classification, and encapsulates the complete information to obtain an image generation request for the asynchronous processing task.

In the embodiment of the specification, the encapsulation of the data to be processed is completed through the encapsulation rule corresponding to the asynchronous processing task, and a foundation is laid for asynchronous scheduling of a task processing model in the follow-up process.

accordingly, step 106 includes the following specific steps:

the server responds to a target task request sent by the request end, feeds back a task identifier of an asynchronous processing task to the request end, and dispatches a task processing model to execute the asynchronous processing task on task data to be processed;

the request end sends a result query request to the server end based on the task identifier;

And under the condition that the execution of the asynchronous processing task is completed, the server feeds back the task processing result of the asynchronous processing task to the request end.

The task identifier is an identifier distributed to the asynchronous processing task by the server and is used for inquiring the processing progress of the asynchronous processing task.

The scheduling task processing model executes asynchronous processing tasks on task data to be processed, and the specific mode is as follows: and the corresponding functional module of the scheduling task processing model executes asynchronous processing tasks on the task data to be processed.

Optionally, after sending the result query request to the server, the method further includes:

and under the condition that the asynchronous processing task is not executed to be completed, the server feeds back the processing progress of the asynchronous processing task to the request end.

The server responds to the target Task request sent by the request end, feeds back a Task identification Task ID of the asynchronous processing Task to the request end, schedules an image classification function module of the Task processing model, and executes the image classification Task on the complete information of the image file. The request end sends a result query request to the server end based on the Task identification Task ID, the server end feeds back the processing progress 'Running, 70 percent' of the asynchronous processing Task to the request end under the condition that the asynchronous processing Task is not executed and completed, and the server end feeds back the image classification result to the request end under the condition that the image classification Task is executed and completed.

In the embodiment of the specification, the asynchronous execution of the asynchronous processing task is realized, the universality of task processing is further enhanced, and the user experience is further improved.

according to the data type of the task data to be processed and the task type of the target task, the task data to be processed is packaged to obtain a target task request aiming at the target task, and the method comprises the following specific steps:

and under the condition that the task type of the target task is a real-time processing task, packaging real-time stream data of the task data to be processed to obtain a target task request aiming at the real-time processing task.

The real-time processing task is a processing task of a full duplex mode, namely, the request end and the service end can mutually and actively send data information to each other, and in the task execution process, the service end actively feeds back the obtained task processing result to the request end. For example, for text generation tasks, the task type is determined to be a real-time processing task, i.e., each word generated is fed back to the requesting end. For another example, for a question-answer task, the task type is determined to be a real-time processing task, i.e., each word is generated and fed back to the requesting end. Also for example, for a voice dialog task, the task type is determined to be a real-time processing task, i.e., each time a piece of audio is generated, it is fed back to the requesting end.

The real-time streaming data is data, such as audio, video, etc., transmitted in real time in the task data to be processed. Real-time property is realized, namely, real-time stream data can only be processed and analyzed in the transmission process and cannot be stored. For example, for voice conversational tasks, conversational voice audio is real-time streaming data. For another example, for audio transcription to generate real-time subtitles, the audio is real-time streaming data.

In an exemplary embodiment, when the task type of the Audio transcription task for subtitle generation is a real-time processing task, audio Stream data audio_stream of the live Stream is encapsulated, so as to obtain an Audio transcription request for the real-time processing task.

In the embodiment of the specification, the encapsulation of the data to be processed is completed through the encapsulation rule corresponding to the real-time processing task, and a foundation is laid for the subsequent real-time scheduling of the task processing model.

accordingly, step 106 includes the following specific steps:

the server receives a target task request sent by the request end in real time, the scheduling task processing model executes a real-time processing task on the task data to be processed, and a task processing result of the real-time processing task is fed back to the request end in real time.

And real-time data stream transmission is carried out by real-time sending and real-time feedback, namely, through the real-time data connection between the request end and the service end.

The scheduling task processing model executes real-time processing tasks on task data to be processed, and the specific mode is as follows: and the corresponding functional module of the scheduling task processing model executes real-time processing tasks on the task data to be processed.

The server receives an Audio transcription request sent by the request end in real time, and schedules an Audio transcription function module of the task processing model to execute an Audio transcription task of real-time transcription on Audio Stream data audio_stream to obtain an Audio transcription result: according to the latest report of the former, … … is currently reported.

In the embodiment of the specification, the real-time execution of the real-time processing task is realized, the universality of task processing is further enhanced, and the user experience is further improved

Referring to fig. 2, fig. 2 shows a flowchart of another task processing method provided in an embodiment of the present disclosure, where the method is applied to a client, and includes the following specific steps:

step 202: and sending the input task data to be processed to a task processing platform, wherein the task processing platform receives the task data to be processed through a task interface, and the task interface is used for receiving the task data of different tasks.

Step 204: and receiving a task processing result fed back by the task processing platform, wherein the task processing result is obtained by executing a target task on task data to be processed by a task processing scheduling model based on a target task request, and the target task request is obtained by packaging the task data to be processed according to the data type of the task data to be processed.

The embodiment of the specification is applied to the client side of the application program, the webpage or the applet with the task processing function, the data connection is established between the client side and the task processing platform, the client side executes the target task by sending the task data to be processed to the task processing platform and by dispatching the task processing model deployed on the server side, the task processing result is obtained, and the task processing result fed back by the task processing platform is received. Scheduling task processing models on a task processing platform to execute target tasks is transparent to users on clients. The embodiments of the present disclosure and the embodiment of fig. 1 are derived from the same inventive concept, and the details of the embodiments of fig. 1 are not described herein.

Illustratively, a text generation result fed back by the task processing platform is received by sending and inputting a subject text "text in a text generation task, such as a text in a text style about rainy days" to the task processing platform: the rainy days are natural artworks, and are full of the strength and beauty of life. In rainy days, people can hear the sound of raindrops falling on the ground and feel the warmth and freshness of the raindrops.

In the embodiment of the specification, task data to be processed is sent to a task processing platform, the task data to be processed is received on the task processing platform through task interfaces for receiving task data of different tasks, development of a plurality of application programming interfaces to receive data of corresponding tasks is avoided, development cost is reduced, task processing results fed back by the task processing platform are received, the task processing results are obtained by executing target tasks based on a target task request scheduling task processing model, the target task requests are obtained by correspondingly packaging the task data to be processed according to data types of the task data to be processed, corresponding processing tasks of the task data to be processed of various data types are executed by the scheduling task processing model through a unified interface, and user experience is improved.

Referring to fig. 3, fig. 3 shows a flowchart of an automatic question-answering method provided in an embodiment of the present disclosure, where the method is applied to a task processing platform, and includes the following specific steps:

step 302: question data input through a task interface is received, wherein the task interface is used for receiving task data of different tasks.

Step 304: and packaging the problem data according to the data type of the problem data to obtain a target task request aiming at the question-answering task.

Step 306: based on the target task request, a task processing model is scheduled to execute a question-answer task on the question data to obtain answer data, wherein the task processing model is a deep learning model which is obtained by training in advance based on sample data of different tasks.

The question-answering task is a dialogue task executed in a question-answering mode, a user inputs question data on a client of a task service platform, and after the question-answering task is executed, answer data is received. The input question data is to-be-replied question data in a question-and-answer task, and is data of a character string mode. The answer data is the processing result of the question-answer task, and is data of at least one medium mode aiming at the question data. For example, the question data is "what is the temperature of the tomorrow a to be asked? The "reply data is" the air temperature of tomorrow A is 30 degrees "and the Wen Shixu chart. For the embodiment of the present disclosure, the specific manner of steps 302 to 306 refers to steps 102 to 106 in the embodiment of fig. 1, and is not described herein.

In the embodiment of the specification, the task interfaces for receiving task data of different tasks are used for receiving input problem data, so that development of a plurality of application programming interfaces to receive data of corresponding tasks is avoided, development cost is reduced, the problem data is correspondingly packaged according to the data types of the problem data to obtain target task requests aiming at the question-answer tasks, a task processing model is scheduled based on the target task requests, the task processing model is scheduled to execute the corresponding processing tasks on the question-answer data through a unified interface to obtain targeted answer data, a user is prevented from calling the task processing model through a special question-answer interface, the universality of task processing is enhanced, and user experience is improved.

Fig. 4 is a schematic architecture diagram of a task processing platform in a task processing method according to an embodiment of the present disclosure, where the task processing platform is shown in fig. 4:

the task processing platform comprises a request end and a service end of the task processing model. The request end receives task data to be processed of different tasks (task data of a generated task and task data … … of a conversation task) through unified task interfaces of synchronous processing, asynchronous processing and real-time processing, determines corresponding scheduling modes (hypertext transfer protocol service scheduling and WebSocket transfer protocol service scheduling), sends a target task request to a service end of a task processing model, and schedules the task processing model to execute the target processing task.

Fig. 5 shows a data flow diagram of a task processing platform in a task processing method according to an embodiment of the present disclosure, where the data flow diagram is shown in fig. 5:

the task processing platform receives different task data to be processed (character string input, file data, content generator input or other types of input) through a unified task interface, encapsulates the task data to be processed according to the data types of the input task data to be processed, obtains a target task request, and dispatches a task processing model from a server side of the task processing model to execute the target task based on the target task request.

Fig. 6 is a schematic flow chart of a task synchronization processing in a task processing method according to an embodiment of the present disclosure, where the flow chart is shown in fig. 6:

the request end (SDK) sends the target task request obtained by encapsulation to the service end, and the service end returns a task processing result to the request end (SDK) after the execution is completed.

Fig. 7 is a schematic flow chart of asynchronous task processing in a task processing method according to an embodiment of the present disclosure, where the flow chart is shown in fig. 7:

the method comprises the steps that a request end (SDK) sends a target task request obtained through encapsulation to a server end, the server end returns a task identifier to the request end (SDK), the request end (SDK) sends a result query request to the server end according to the task identifier, the request end (SDK) returns a task processing progress to the request end (SDK) under the condition that target task execution is incomplete, and the request end (SDK) returns a task processing result to the request end (SDK) under the condition that target task execution is complete.

Fig. 8 is a schematic flow chart of real-time task processing in a task processing method according to an embodiment of the present disclosure, where the flow chart is shown in fig. 8:

the request end and the service end complete the dispatching of the task processing model through the real-time data flow, and the task processing result is returned to the request end (SDK) directly through the real-time data flow after the task processing result is obtained.

The task processing method provided in the present specification will be further described with reference to fig. 9 by taking an application of the task processing method to speech transcription as an example. FIG. 9 is a flowchart of a task processing method for speech transcription, where the method is applied to a human task processing platform with executable tasks for speech transcription of different types, and includes the following specific steps:

step 902: and receiving voice data to be transcribed, which is transmitted by the client and is input through a task interface, wherein the task interface is used for receiving task data of different tasks.

Step 904: and judging the task type of the voice transcription task.

In the case that the task type of the speech transcription task is a synchronous processing task, for example, text content is generated for recording, step 906 is performed; in the case that the task type of the speech transcription task is an asynchronous processing task, for example, a text document is generated for storage, step 912 is performed; in the case where the task type of the voice transcription task is a real-time processing task, for example, a live caption is generated for display, step 922 is performed.

Step 906: the request end extracts the data information to be synchronized in the voice data to be transcribed.

Step 908: and packaging the information of the data to be synchronized to obtain a voice transcription task request aiming at the synchronous processing task.

Step 910: the server receives a voice transcription task request sent by the request end, the scheduling large model executes synchronous processing tasks on voice data to be transcribed, and feeds back voice transcription results of the synchronous processing tasks to the request end, and the request end feeds back the voice transcription results to the client end.

Step 912: the request end encapsulates complete information of voice data to be transcribed to obtain a voice transcription task request aiming at an asynchronous processing task.

Step 914: the server responds to the voice transcription task request sent by the request end, feeds back the task identification of the asynchronous processing task to the request end, and dispatches the large model to execute the asynchronous processing task on the voice data to be transcribed.

Step 916: the request end sends a result query request to the server end based on the task identification.

Step 918: and under the condition that the asynchronous processing task is not executed, the server feeds back the voice transcription progress of the asynchronous processing task to the request terminal, and the request terminal feeds back the voice transcription progress to the client.

Step 920: under the condition that the execution of the asynchronous processing task is completed, the server feeds back a voice transcription result of the asynchronous processing task to the request end, and the request end feeds back the voice transcription result to the client.

Step 922: and packaging the real-time stream data of the voice data to be transcribed to obtain a voice transcription task request aiming at the real-time processing task.

Step 924: the server receives a voice transcription task request sent by the request end in real time, schedules the large model to execute a real-time processing task on voice data to be transcribed, feeds back a voice transcription result of the real-time processing task to the request end in real time, and feeds back the voice transcription result to the client end.

In the embodiment of the specification, the task interfaces for receiving the task data of different types are used for receiving the input voice data to be transcribed, so that the development of a plurality of application programming interfaces to receive the data of corresponding tasks is avoided, the development cost is reduced, the voice data to be transcribed is correspondingly packaged according to the data type of the voice data to be transcribed, the voice transcription task request aiming at the voice transcription task is obtained, the large scheduling model is based on the voice transcription task request, the large scheduling model is used for executing the corresponding processing tasks on the voice data to be transcribed of various types through the unified interface, the universality of voice transcription is enhanced, and the user experience is improved.

Corresponding to the method embodiment, the present disclosure further provides an embodiment of a task processing device, and fig. 10 shows a schematic structural diagram of the task processing device provided in one embodiment of the present disclosure. As shown in fig. 10, the apparatus is applied to a task processing platform, and includes:

a first data receiving module 1002 configured to receive task data to be processed input through a task interface, where the task interface is configured to receive task data of different tasks;

the first packaging module 1004 is configured to package the task data to be processed according to the data type of the task data to be processed, so as to obtain a target task request aiming at a target task;

the first processing module 1006 is configured to execute a target task on the task data to be processed based on the target task request, and obtain a task processing result, where the task processing model is a deep learning model trained in advance based on sample data of different tasks.

Optionally, the first encapsulation module 1004 is further configured to: and packaging the task data to be processed according to the data type of the task data to be processed and the task type of the target task to obtain a target task request aiming at the target task.

Optionally, the task processing platform includes a request end; accordingly, the first encapsulation module 1004 is further configured to: under the condition that the task type of the target task is a synchronous processing task, the request end extracts data information to be synchronous in task data to be processed; and packaging the information of the data to be synchronized to obtain a target task request aiming at the synchronous processing task.

Optionally, the task processing platform further comprises a server; accordingly, the first processing module 1006 is further configured to: the server receives a target task request sent by the request end, the scheduling task processing model executes a synchronous processing task on the task data to be processed, and a task processing result of the synchronous processing task is fed back to the request end.

Optionally, the task processing platform includes a request end; accordingly, the first encapsulation module 1004 is further configured to: and under the condition that the task type of the target task is an asynchronous processing task, the request end packages the complete information of the task data to be processed to obtain a target task request aiming at the asynchronous processing task.

Optionally, the task processing platform further comprises a server; accordingly, the first processing module 1006 is further configured to: the server responds to a target task request sent by the request end, feeds back a task identifier of an asynchronous processing task to the request end, and dispatches a task processing model to execute the asynchronous processing task on task data to be processed; the request end sends a result query request to the server end based on the task identifier; and under the condition that the execution of the asynchronous processing task is completed, the server feeds back the task processing result of the asynchronous processing task to the request end.

Optionally, the task processing platform includes a request end; accordingly, the first encapsulation module 1004 is further configured to: and under the condition that the task type of the target task is a real-time processing task, packaging real-time stream data of the task data to be processed to obtain a target task request aiming at the real-time processing task.

Optionally, the task processing platform further comprises a server; accordingly, the first processing module 1006 is further configured to: the server receives a target task request sent by the request end in real time, the scheduling task processing model executes a real-time processing task on the task data to be processed, and a task processing result of the real-time processing task is fed back to the request end in real time.

In the embodiment of the specification, the task interfaces for receiving the task data of different tasks are used for receiving the input task data to be processed, so that the development of a plurality of application programming interfaces to receive the data of the corresponding tasks is avoided, the development cost is reduced, the task data to be processed is correspondingly packaged according to the data types of the task data to be processed, the target task request aiming at the target task is obtained, the task processing model is scheduled based on the target task request, the task processing model to be scheduled is used for executing the corresponding processing tasks on the task data to be processed of a plurality of data types through the unified interface, the universality of task processing is enhanced, and the user experience is improved.

The above is a schematic solution of a task processing device of the present embodiment. It should be noted that, the technical solution of the task processing device and the technical solution of the task processing method belong to the same concept, and details of the technical solution of the task processing device, which are not described in detail, can be referred to the description of the technical solution of the task processing method.

Corresponding to the above method embodiments, the present disclosure further provides an embodiment of a task processing device, and fig. 11 shows a schematic structural diagram of another task processing device provided in one embodiment of the present disclosure. As shown in fig. 11, the apparatus is applied to a task processing platform, and includes:

the sending module 1102 is configured to send the input task data to be processed to the task processing platform, wherein the task processing platform receives the task data to be processed through a task interface, and the task interface is used for receiving task data of different tasks;

the result receiving module 1104 is configured to receive a task processing result fed back by the task processing platform, where the task processing result is obtained by executing a target task on task data to be processed by the task processing scheduling model based on a target task request, and the target task request is obtained by packaging the task data to be processed according to a data type of the task data to be processed.

Corresponding to the method embodiment, the present disclosure further provides an automatic question-answering device embodiment, and fig. 12 shows a schematic structural diagram of an automatic question-answering device provided in one embodiment of the present disclosure. As shown in fig. 12, the apparatus is applied to a task processing platform, and includes:

a second data receiving module 1202 configured to receive question data input through a task interface, wherein the task interface is used for receiving task data of different tasks;

the second packaging module 1204 is configured to package the problem data according to the data type of the problem data, so as to obtain a target task request for the question-answering task;

and a second processing module 1206 configured to schedule a task processing model to execute a question-answer task on the question data based on the target task request, and obtain answer data, wherein the task processing model is a deep learning model trained in advance based on sample data of different tasks.

The above is a schematic scheme of an automatic question answering apparatus of this embodiment. It should be noted that, the technical solution of the automatic question-answering device and the technical solution of the automatic question-answering method belong to the same concept, and details of the technical solution of the automatic question-answering device, which are not described in detail, can be referred to the description of the technical solution of the automatic question-answering method.

Corresponding to the above method embodiments, the present disclosure further provides a task processing system embodiment, and fig. 13 shows a schematic structural diagram of a task processing system provided in one embodiment of the present disclosure. As shown in fig. 13, the system includes a client 1302 and a task processing platform 1304;

a client 1302, configured to send input task data to be processed to a task processing platform 1304;

the task processing platform 1304 is configured to receive task data to be processed, which is input through a task interface, where the task interface is configured to receive task data of different tasks, encapsulate the task data to be processed according to a data type of the task data to be processed, obtain a target task request for a target task, schedule a task processing model to execute the target task on the task data to be processed based on the target task request, obtain a task processing result, and feed back the task processing result to the client 1302, where the task processing model is a deep learning model that is trained in advance based on sample data of different tasks;

The client 1302 is further configured to receive a task processing result fed back by the task processing platform 1304.

In the embodiment of the specification, the task processing platform receives the input task data to be processed through the task interfaces for receiving the task data of different tasks, so that development of a plurality of application programming interfaces to receive the data of the corresponding tasks is avoided, development cost is reduced, the task data to be processed is correspondingly packaged according to the data types of the task data to be processed, a target task request aiming at a target task is obtained, a task processing model is scheduled based on the target task request, the task processing model is scheduled to execute the corresponding processing tasks on the task data to be processed of a plurality of data types through a unified interface, universality of task processing is enhanced, and user experience is improved.

The above is a schematic solution of a task processing system of the present embodiment. It should be noted that, the technical solution of the task processing system and the technical solution of the task processing method belong to the same concept, and details of the technical solution of the task processing system, which are not described in detail, can be referred to the description of the technical solution of the task processing method.

Corresponding to the above method embodiments, the present disclosure further provides a task processing platform embodiment, and fig. 14 shows a schematic structural diagram of a task processing platform according to one embodiment of the present disclosure. As shown in fig. 14, the platform includes a request end 1402 and a service end 1404;

a request end 1402, configured to receive task data to be processed, which is input by a client through a task interface, where the task interface is configured to receive task data of different tasks, encapsulate the task data to be processed according to a data type of the task data to be processed, obtain a target task request for a target task, and process the target task request to a server end 1404;

the server 1404 is configured to receive a target task request sent by the request 1402, and schedule a task processing model to execute a target task on task data to be processed, so as to obtain a task processing result, where the task processing model is a deep learning model that is trained in advance based on sample data of different tasks.

Optionally, data transmission is performed between the request end 1402 and the service end 1404 through a full duplex streaming manner.

The above is a schematic solution of a task processing platform of the present embodiment. It should be noted that, the technical solution of the task processing platform and the technical solution of the task processing method belong to the same concept, and details of the technical solution of the task processing platform which are not described in detail can be referred to the description of the technical solution of the task processing method.

FIG. 15 illustrates a block diagram of a computing device provided in one embodiment of the present description. Components of the computing device 1500 include, but are not limited to, a memory 1510 and a processor 1520. Processor 1520 is coupled to memory 1510 via bus 1530 and database 1550 is used to hold data.

Computing device 1500 also includes an access device 1540, the access device 1540 enabling the computing device 1500 to communicate via one or more networks 1560. Examples of such networks include public switched telephone networks (PSTN, public Switched Telephone Network), local area networks (LAN, local Area Network), wide area networks (WAN, wide Area Network), personal area networks (PAN, personal Area Network), or combinations of communication networks such as the internet. Access device 1540 may include one or more of any type of network interface, wired or wireless (e.g., network interface card (NIC, network interface controller)), such as an IEEE802.11 wireless local area network (WLAN, wireless Local Area Network) wireless interface, a worldwide interoperability for microwave access (Wi-MAX, worldwide Interoperability for Microwave Access) interface, an ethernet interface, a universal serial bus (USB, universal Serial Bus) interface, a cellular network interface, a bluetooth interface, near field communication (NFC, near Field Communication).

In one embodiment of the present description, the above-described components of computing device 1500, as well as other components not shown in FIG. 15, may also be connected to each other, such as by a bus. It should be understood that the block diagram of the computing device illustrated in FIG. 15 is for exemplary purposes only and is not intended to limit the scope of the present description. Those skilled in the art may add or replace other components as desired.

Computing device 1500 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., tablet, personal digital assistant, laptop, notebook, netbook, etc.), mobile phone (e.g., smart phone), wearable computing device (e.g., smart watch, smart glasses, etc.), or other type of mobile device, or a stationary computing device such as a desktop computer or personal computer (PC, personal Computer). Computing device 1500 may also be a mobile or stationary server.

Wherein the processor 1520 is configured to execute computer-executable instructions that, when executed by the processor, perform the steps of the task processing method or the automatic question-answering method described above.

The foregoing is a schematic illustration of a computing device of this embodiment. It should be noted that, the technical solution of the computing device belongs to the same concept as the technical solutions of the task processing method and the automatic question-answering method, and details of the technical solution of the computing device, which are not described in detail, can be referred to the description of the technical solutions of the task processing method or the automatic question-answering method.

An embodiment of the present disclosure also provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the task processing method or the automatic question-answering method described above.

The above is an exemplary version of a computer-readable storage medium of the present embodiment. It should be noted that, the technical solution of the storage medium and the technical solutions of the task processing method and the automatic question-answering method belong to the same concept, and details of the technical solution of the storage medium which are not described in detail can be referred to the description of the technical solutions of the task processing method or the automatic question-answering method.

An embodiment of the present specification also provides a computer program, wherein the computer program, when executed in a computer, causes the computer to perform the steps of the task processing method or the automatic question-answering method described above.

The above is an exemplary version of a computer program of the present embodiment. It should be noted that, the technical solution of the computer program and the technical solution of the task processing method and the automatic question-answering method belong to the same concept, and details of the technical solution of the computer program, which are not described in detail, can be referred to the description of the technical solution of the task processing method or the automatic question-answering method.

The foregoing describes specific embodiments of the present disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

The computer instructions include computer program code that may be in source code form, object code form, executable file or some intermediate form, etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), an electrical carrier signal, a telecommunications signal, a software distribution medium, and so forth. It should be noted that the content of the computer readable medium can be increased or decreased appropriately according to the requirements of the patent practice, for example, in some areas, according to the patent practice, the computer readable medium does not include an electric carrier signal and a telecommunication signal.

It should be noted that, for simplicity of description, the foregoing method embodiments are all expressed as a series of combinations of actions, but it should be understood by those skilled in the art that the embodiments are not limited by the order of actions described, as some steps may be performed in other order or simultaneously according to the embodiments of the present disclosure. Further, those skilled in the art will appreciate that the embodiments described in the specification are all preferred embodiments, and that the acts and modules referred to are not necessarily all required for the embodiments described in the specification.

In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to the related descriptions of other embodiments.

The preferred embodiments of the present specification disclosed above are merely used to help clarify the present specification. Alternative embodiments are not intended to be exhaustive or to limit the invention to the precise form disclosed. Obviously, many modifications and variations are possible in light of the teaching of the embodiments. The embodiments were chosen and described in order to best explain the principles of the embodiments and the practical application, to thereby enable others skilled in the art to best understand and utilize the invention. This specification is to be limited only by the claims and the full scope and equivalents thereof.

Claims

1. A task processing method is applied to a task processing platform and comprises the following steps:

according to the data type of the task data to be processed, packaging the task data to be processed to obtain a target task request aiming at a target task;

and based on the target task request, scheduling a task processing model to execute the target task on the task data to be processed to obtain a task processing result, wherein the task processing model is a deep learning model which is obtained by training in advance based on sample data of different tasks.

2. The method of claim 1, wherein the encapsulating the task data to be processed according to the data type of the task data to be processed to obtain the target task request for the target task includes:

3. The method of claim 2, the task processing platform comprising a requesting end;

The step of packaging the task data to be processed according to the data type of the task data to be processed and the task type of the target task to obtain a target task request aiming at the target task, comprises the following steps:

under the condition that the task type of the target task is a synchronous processing task, the request end extracts data information to be synchronous in the task data to be processed;

and packaging the data information to be synchronized to obtain a target task request aiming at the synchronous processing task.

4. A method according to claim 3, the task processing platform further comprising a server;

the task processing module is configured to execute the target task on the task data to be processed based on the target task request, and obtain a task processing result, where the task processing module includes:

the server receives the target task request sent by the request end, schedules the task processing model to execute the synchronous processing task on the task data to be processed, and feeds back a task processing result of the synchronous processing task to the request end.

5. The method of claim 2, the task processing platform comprising a requesting end;

6. The method of claim 5, the task processing platform further comprising a server;

the server side responds to the target task request sent by the request side, feeds back a task identifier of the asynchronous processing task to the request side, and dispatches a task processing model to execute the asynchronous processing task on the task data to be processed;

and under the condition that the execution of the asynchronous processing task is completed, the server feeds back a task processing result of the asynchronous processing task to the request end.

7. The method of claim 2, the task processing platform comprising a requesting end;

And under the condition that the task type of the target task is a real-time processing task, packaging the real-time stream data of the task data to be processed to obtain a target task request aiming at the real-time processing task.

8. The method of claim 7, the task processing platform further comprising a server;

the server receives the target task request sent by the request end in real time, schedules the task processing model to execute the real-time processing task on the task data to be processed, and feeds back the task processing result of the real-time processing task to the request end in real time.

9. A task processing method is applied to a client and comprises the following steps:

and receiving a task processing result fed back by a task processing platform, wherein the task processing result is obtained by scheduling a task processing model to execute a target task on the task data to be processed based on a target task request, and the target task request is obtained by packaging the task data to be processed according to the data type of the task data to be processed.

10. An automatic question-answering method is applied to a task processing platform and comprises the following steps:

according to the data type of the question data, packaging the question data to obtain a target task request aiming at a question-answering task;

and based on the target task request, scheduling a task processing model to execute the question-answer task on the question data to obtain answer data, wherein the task processing model is a deep learning model which is obtained by training in advance based on sample data of different tasks.

11. A task processing system, the system comprising a client and a task processing platform;

the task processing platform is used for receiving the task data to be processed, which are input through a task interface, wherein the task interface is used for receiving task data of different tasks, packaging the task data to be processed according to the data type of the task data to be processed to obtain a target task request aiming at a target task, scheduling a task processing model to execute the target task on the task data to be processed based on the target task request to obtain a task processing result, and feeding back the task processing result to the client, wherein the task processing model is a deep learning model which is trained in advance based on sample data of different tasks;

The client is also used for receiving the task processing result fed back by the task processing platform.

12. The task processing platform comprises a request end and a service end;

the request end is used for receiving task data to be processed, which are input by the client end through a task interface, wherein the task interface is used for receiving task data of different tasks, packaging the task data to be processed according to the data type of the task data to be processed, obtaining a target task request aiming at a target task, and sending the target task processing request to the server end;

the server is used for receiving the target task request sent by the request end, and scheduling a task processing model to execute the target task on the task data to be processed to obtain a task processing result, wherein the task processing model is a deep learning model which is trained in advance based on sample data of different tasks.

13. A computing device, comprising:

a memory and a processor;

the memory is configured to store computer executable instructions, the processor being configured to execute the computer executable instructions, which when executed by the processor, implement the steps of the method of any one of claims 1 to 10.

14. A computer readable storage medium storing computer executable instructions which when executed by a processor implement the steps of the method of any one of claims 1 to 10.