CN115221199A

CN115221199A - Information processing method, device, equipment and storage medium

Info

Publication number: CN115221199A
Application number: CN202210900041.0A
Authority: CN
Inventors: 钱志达; 孙裕文
Original assignee: Midea Group Co Ltd; Guangdong Midea Kitchen Appliances Manufacturing Co Ltd
Current assignee: Midea Group Co Ltd; Guangdong Midea Kitchen Appliances Manufacturing Co Ltd
Priority date: 2022-07-28
Filing date: 2022-07-28
Publication date: 2022-10-21

Abstract

According to the information processing method, the information processing device, the equipment and the storage medium, the first equipment determines user input information with abnormal processing; converting the user input information which is processed abnormally to obtain query information related to the user input information which is processed abnormally; determining at least one query result matching the query information; receiving a first query result; the first query result is any one of the at least one query result; and sending the abnormal user input information and the first query result to second equipment.

Description

Information processing method, device, equipment and storage medium

Technical Field

The application relates to the field of intelligent back-end cloud platform construction, in particular to an information processing method, device, equipment and storage medium.

Background

With the popularization of various intelligent cooking appliances, various intelligent functions based on visual and voice interaction emerge endlessly, and the realization of the intelligent functions requires a large number of and various types of sample-tags to be manufactured on corresponding tasks to perform parameter learning on a data set so as to ensure the realization of the intelligent functions. However, the operation similar to the exhaustive operation on the learning data enables the existing model to perform well on the trained data category, and the intelligent device is required to have the capability of self-iterative optimization for the data samples which are not subjected to parameter learning, have large difference in the learned data distribution and are not contained in the background knowledge base to perform poorly. The functions of various on-line devices can be continuously learned and updated by means of a powerful cloud platform in the background, but new data samples and accurate sample label labels are needed, the former can be obtained by collecting samples which are not correctly solved after products are on-line, the latter needs a large amount of manpower to label the difficult samples and then parameter learning iteration of intelligent functions can be carried out, and how to solve the problem of obtaining correct sample labels in the self-learning iteration updating process of intelligent cooking devices is a bottleneck problem of self-learning of the cloud platform at present.

Disclosure of Invention

In view of the above, the main object of the present invention is to provide an information processing method, an information processing apparatus, a device and a storage medium.

The technical scheme of the embodiment of the application is realized as follows:

the embodiment of the application provides an information processing method, which is applied to first equipment and comprises the following steps:

determining user input information for handling the exception;

converting the user input information which is processed abnormally to obtain query information related to the user input information which is processed abnormally;

determining at least one query result matching the query information;

receiving a first query result; the first query result is any one of the at least one query result;

sending the abnormal user input information and the first query result to second equipment; and the user input information with abnormal processing and the first query result are used for the second equipment to perform learning updating.

In the foregoing solution, the determining the abnormal user input information includes:

receiving the input information;

judging whether the first preset model can normally process the input information or not;

and under the condition that the first preset model cannot normally process the input information, determining the input information as the user input information with abnormal processing.

In the above scheme, the input information includes voice information and/or picture information; determining that the input information is the user input information with abnormal processing under the condition that the first preset model cannot normally process the input information, wherein the determining comprises the following steps:

and under the condition that the first preset model cannot normally process the voice information and/or the picture information, determining the voice information and/or the picture information as the user input information with abnormal processing.

In the foregoing solution, the converting the abnormal user input information to obtain query information related to the abnormal user input information includes:

and under the condition that the voice information is the user input information with abnormal processing, performing information supplement on the voice information to obtain query information related to the voice information.

and under the condition that the picture information is the user input information with abnormal processing, extracting the information of the picture information to obtain query information related to the picture information.

In the above solution, after determining at least one query result matching the query information, the method further includes:

sequencing the at least one query result to obtain a first sequencing result;

at least one alternative query result is determined based on the first ranked result.

In the foregoing solution, the method further includes:

re-determining at least one alternative query result based on the first ranking result in the event that the user has not selected the at least one alternative query result; the re-determined at least one alternative query result is different from the at least one alternative query result.

In the foregoing solution, the method further includes:

obtaining the number of times of redetermining at least one alternative query result;

and under the condition that the times are larger than or equal to a first preset threshold value, stopping re-determining at least one alternative query result.

In the above scheme, the method further comprises:

and selecting the at least one query result based on a preset mode to obtain a first query result.

In the above scheme, the method further comprises:

sending the user input information with abnormal processing to the second device under the condition that the user does not select the at least one query result; and the abnormal user input information is used for the second equipment to determine a second query result.

In the above scheme, the method further comprises:

and under the condition that the first preset model can normally process the input information, determining a third query result based on the first preset model.

The embodiment of the application also provides an information processing method, which is applied to the second device and comprises the following steps:

receiving user input information and a first query result which are sent by first equipment and used for processing abnormity;

based on the abnormal user input information and the first query result, learning and updating a second preset model to obtain an updated second preset model; the updated second preset model is used for updating the first preset model in the first equipment.

In the above scheme, the method further comprises:

receiving user input information which is sent by the first equipment and used for processing abnormity;

determining a second query result matched with the abnormal user input information based on the abnormal user input information; and the second query result is used for displaying.

An embodiment of the present application further provides an information processing apparatus, where the information processing apparatus is disposed on a first device, and the information processing apparatus includes:

the first determining module is used for determining abnormal user input information;

the conversion module is used for converting the user input information which is processed abnormally to obtain query information related to the user input information which is processed abnormally;

the second determination module is used for determining at least one query result matched with the query information;

the first receiving module is used for receiving a first query result; the first query result is any one of the at least one query result;

the first sending module is used for sending the abnormal user input information and the first query result to the second equipment; and the abnormal user input information and the first query result are used for the second equipment to perform learning updating.

An embodiment of the present application further provides an information processing apparatus, where the information processing apparatus is disposed on a second device, and the information processing apparatus includes:

the second receiving module is used for receiving the abnormal user input information and the first query result sent by the first equipment;

the updating module is used for learning and updating a second preset model based on the abnormal user input information and the first query result to obtain an updated second preset model; the updated second preset model is used for updating the first preset model in the first equipment.

An embodiment of the present application further provides a first device, including: a first processor and a first memory for storing a computer program capable of running on the processor,

wherein the first processor is configured to execute the steps of any of the above-mentioned methods of the first device side when running the computer program.

An embodiment of the present application further provides a second device, including: a second processor and a second memory for storing a computer program capable of running on the processor,

wherein the second processor is configured to execute the steps of any one of the methods of the second device side when running the computer program.

An embodiment of the present application further provides a storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of any one of the above-mentioned methods for the first device side, or implements the steps of any one of the above-mentioned methods for the first device side.

According to the information processing method, the information processing device, the equipment and the storage medium, the first equipment determines user input information with abnormal processing; converting the user input information which is processed abnormally to obtain query information related to the user input information which is processed abnormally; determining at least one query result matching the query information; receiving a first query result; the first query result is any one of the at least one query result; sending the abnormal user input information and the first query result to second equipment; correspondingly, the second equipment receives the abnormal user input information and the first query result sent by the first equipment; based on the abnormal user input information and the first query result, learning and updating a second preset model to obtain an updated second preset model; the updated second preset model is used for updating the first preset model in the first equipment. Matching the user input information by determining the abnormal user input information processed by the first equipment and utilizing massive internet knowledge, and returning at least one query result for the user to select, wherein on one hand, the user obtains a correct answer in the at least one query result through self-selection; and on the other hand, the first equipment sends the user input information with abnormal processing and the correct answer to second equipment as labeled data, and the second equipment updates and learns a second preset model based on the labeled data to obtain the updated second preset model, so that the real-time iterative updating of the second preset model of the second equipment is realized.

Drawings

FIG. 1 is a schematic flow chart illustrating an information processing method according to an embodiment of the present disclosure;

FIG. 2 is a schematic view of a workflow of an information processing method according to an embodiment of the present application;

FIG. 3 is a schematic flow chart illustrating an information processing method according to an embodiment of the present application;

FIG. 4 is a schematic flow chart of an information processing method and an information processing method according to an embodiment of the present application;

FIG. 5 is a schematic structural diagram of an information processing apparatus according to an embodiment of the present application;

FIG. 6 is a schematic structural diagram of an information processing apparatus according to an embodiment of the present application;

FIG. 7 is a schematic diagram of an information processing system and an information processing system according to an embodiment of the present application;

FIG. 8 is a schematic structural diagram of a first apparatus according to an embodiment of the present disclosure;

fig. 9 is a schematic structural diagram of a second apparatus according to an embodiment of the present application.

Detailed Description

The present application will be described in further detail with reference to the following drawings and examples.

In the related art, for various intelligent functions of the online intelligent cooking equipment, the self-learning solution of the intelligent learning cloud platform can be roughly divided into three types:

the most common method is to adopt a manually labeled supervised learning mode, firstly, samples which cannot correctly complete the functions at present are collected through the use of an intelligent product of a user, then, the samples of a background cloud platform reach a certain amount or are regularly processed by the difficult samples, a background maintainer sorts and labels the data to form new training data, then, the new training data is added with supervised parameter learning to obtain a model with strong adaptability to new data, and finally, the new model is re-deployed on line, so that iterative updating of the intelligent functions can be completed by periodically and repeatedly performing. The model learning effect is good in the mode, but a large amount of manual participation is needed, the time cost is high, and the model iterative updating in real time cannot be completed.

Because the underlying implementation logic of many intelligent functions can be merged into the problem of sample classification, another solution is to set learning parameters, learn in an unsupervised manner (i.e., no fixed label needs to be set in the training process), directly perform unsupervised learning on collected samples in a similar clustering manner, and cluster-learn samples of the same category together to identify the same category and different categories. The method saves time and cost consumption of manual labeling and can also achieve real-time updating, but because the distributed learning task of the sample without the label is difficult, the distinguishing of part of similar samples is difficult, and the robustness and the accuracy of the model are relatively low.

The third is a self-learning platform combining a supervision mode and an unsupervised mode, a supervision model is adopted for realizing the actual functions, high precision can be achieved, and when new sample data is processed, the unsupervised method is used for classifying and sorting the new samples, so that a large amount of manual labeling work can be saved, the method is limited by the realization precision of the unsupervised model, and the iterative updating of the model still cannot reach a very reliable engineering degree.

In the related technology, a self-learning cloud platform based on a supervised model needs to perform new sample arrangement labeling and parameter training learning regularly, so that the cost and practice consumption are high, and real-time iterative updating cannot be realized; the self-learning cloud platform based on unsupervised learning and the cloud platform combining unsupervised learning and supervised learning are limited by the robustness and model precision of an unsupervised model, and the reliability of engineering is difficult to realize.

Based on this, in various embodiments of the application, a user is used as a high-level intelligent agent, and the user has the characteristic of relatively accurate correct tag distinguishing capability although the user does not have professional tag marking capability, a user interaction interface with functions of recommending search and selecting answers is arranged at a first equipment end, and a self-learning cloud platform based on supervised learning is arranged at a second equipment end. And recommending and displaying the recommended answers to the user by combining with internet search to form a sample for carrying out real-time iterative updating on the second equipment.

An embodiment of the present application provides an information processing method, where the method is applied to a first device, and fig. 1 is a schematic flow diagram of the information processing method according to the embodiment of the present application, and as shown in fig. 1, the method includes:

step 101: determining user input information for handling the exception;

step 102: converting the user input information which is processed abnormally to obtain query information related to the user input information which is processed abnormally;

step 103: determining at least one query result matching the query information;

step 104: receiving a first query result; the first query result is any one of the at least one query result;

step 105: sending the abnormal user input information and the first query result to second equipment; and the abnormal user input information and the first query result are used for the second equipment to perform learning updating.

In practical application, the first device may be determined according to practical situations, and is not limited herein. As an example, the first device may be a smart device, which may be various kitchen smart cooking devices, such as an oven, an integrated range, a microwave oven, a steam box, an automatic cooker, and the like. The intelligent equipment has an intelligent function of cooking knowledge question answering, and the intelligent function can be a knowledge question answering assistant or a menu searching function and can also be a food material or food image recognition function.

In step 101, the first device may include a user interaction module, where the user interaction module includes an interaction module and a control module, where the interaction module includes an information acquisition unit, the interaction module is configured to perform information interaction in multiple forms with the user, and the interaction module includes an information acquisition unit, and the information acquisition unit is configured to receive the user input information; the control module comprises a central control unit for controlling the first device. The user input information for determining abnormal processing may be that the user input information is acquired by the information acquisition unit, and the user input information for determining abnormal processing is determined by the central control unit.

In step 102, the conversion processing is performed on the user input information with abnormal processing, and the query information related to the user input information with abnormal processing may be obtained by performing conversion processing on the user input information, and enhancing the correlation between the search result and the user input information, so as to obtain the query information related to the user input information with abnormal processing.

In step 103, the first device may include a search module, and the determining at least one query result matching the query information may be inputting the query information to the search module; searching in the Internet based on the searching module to obtain a query result; and matching the query information with the query result, and determining at least one query result matched with the query information.

In step 104, the interaction module includes an information interaction unit, and the information interaction unit is configured to display the at least one query result to the user. The receiving of the first query result may be that the at least one query result is displayed to the user, the user is used as an advanced intelligent agent to have a strong discrimination capability for a correct answer, so that the user selects the first query result from the at least one query result, and the first query result is returned to the first device, and the first device receives the first query result based on the selection of the user for the at least one query result.

In step 105, the control module includes a data transmission unit, where the data transmission unit is configured to perform data transmission with the second device; the sending of the user input information about the abnormal processing and the first query result to the second device may be sending, by the data transmission unit, the user input information about the abnormal processing and the first query result to the second device.

Based on this, in an embodiment, the determining user input information that handles exceptions includes:

receiving the input information;

In actual application, the input information may be determined according to actual conditions, and is not limited herein. As an example, the input information may be input information of a user, and the input information of the user may be information input to the first device by the user. The receiving of the input information may be receiving, by the information acquisition unit, input information of the user. The control module stores the first preset model, and the first preset model can be determined according to actual conditions, which is not limited herein. As an example, the first preset model may be a local model, and is configured to process the input information and determine query information corresponding to the input information.

The method for judging whether the first preset model can normally process the input information may be determined according to actual conditions, and is not limited herein. As an example, whether the first preset model can normally process the input information may be determined according to whether the first preset model can understand the user input information; whether the first preset model can understand the user input information may be, whether the first preset model can output a correct output result corresponding to the user input information according to the user input information.

As another example, it may also be determined whether the first preset model can normally process the input information according to whether a background knowledge base of the second device includes a query result corresponding to the user input information.

In some embodiments, in a case that the first preset model cannot understand the input information, determining that the input information is the abnormal processing user input information, and taking the abnormal processing user input information and the first query result as a difficulty sample; and determining that the input information is the abnormal user input information, and taking the abnormal user input information and the first query result as a new sample.

Based on this, in an embodiment, the input information includes voice information and/or picture information; the determining that the input information is the user input information with abnormal processing under the condition that the first preset model cannot process the input information normally comprises the following steps:

In practical application, the information acquisition unit may include a voice acquisition unit and an image acquisition unit. As an example, the voice collecting unit may be a microphone having a voice collecting function for collecting user information input of questions or instructions in various audio forms; the image acquisition unit can be a camera with an image acquisition function and is used for acquiring user information input of various image information needing to be processed.

Based on this, in an embodiment, the converting the abnormal user input information to obtain query information related to the abnormal user input information includes:

In actual application, the method for information supplementation of the voice information may be determined according to actual conditions, and is not limited herein. As an example, the voice information may be supplemented with question information, which may be a function currently used by the user. For example, in the case where the user makes a cake using an oven, the voice message of the user is "how do without low gluten? "question information can be supplemented to the voice information, and" do nothing with low-gluten flour when making cake? "is used as the query information.

As an example, the text description in the voice message can be converted into a question form. For example, in the case where the user makes a cake using an oven, the voice message of the user is "no soft flour when making a cake", and the text description in the voice message may be converted into a question form, which results in "no soft flour is done when making a cake? "is used as the query information.

In practical application, the method for extracting the information of the picture information may be determined according to practical situations, and is not limited herein. As an example, the picture information may be subjected to color correction, scaling, and key information region extraction.

In some embodiments, the search module includes search interfaces in a variety of search formats, such as: searching graphs, searching words, searching answers and the like by using graphs, and calling the existing mature search recommendation engine to perform recommendation search of related contents, wherein the recommendation search engine is mainly used for performing accurate search recommendation of internet knowledge under the condition that the input information which cannot be normally processed by the intelligent function of the first device or the query result corresponding to the user input information is not contained in the background knowledge base of the second device. As an example, in the case that the voice information is the user input information of the processing abnormality, the search interface may be an encyclopedic knowledge question and answer or a menu content search interface; in a case where the picture information is the user input information with abnormal processing, the search interface may be an image content recognition interface.

In some embodiments, the search module may build a new search engine by building a recall, rough, fine module without using an existing full-fledged search engine.

Based on this, in an embodiment, after determining at least one query result matching the query information, the method further includes:

sequencing the at least one query result to obtain a first sequencing result;

In actual application, the step of sorting the at least one query result to obtain a first sorting result may be sorting the at least one query result based on the correlation between the at least one query result and the query information to obtain the first sorting result. Specifically, based on the fact that the relevance between the at least one query result and the query information is from strong to weak, the at least one query result is ranked, and a first ranking result is obtained.

The determining at least one alternative query result based on the first sorting result may be that a preset number of query results are selected from the first sorting result to determine at least one alternative query result. The preset number may be determined according to actual conditions, and is not limited herein. As an example, the preset number may be 3 or more than 3, 3 or more than 3 query results are selected in the first ranking result, and 3 or more than 3 candidate query results are determined.

In some embodiments, the obtaining a first query result based on the user selection of the at least one query result includes:

obtaining a first query result based on the selection of the at least one alternative query result by the user;

the first query result is any one of the at least one alternative query result.

In practical application, the information interaction unit may include a visual interaction unit and a voice broadcasting unit. As an example, the visual interaction unit may be a display screen with a visual interaction function for presenting the at least one alternative query result to the user; the voice broadcasting unit may be a sound box having a voice broadcasting function, and is configured to broadcast the at least one candidate query result to the user.

The obtaining of the first query result based on the selection of the user on the at least one alternative query result may be displaying the at least one alternative query result to the user, implementing that the user selects a correct query result in the at least one alternative query result by using a characteristic that the user has a strong discriminative power on a correct answer as a high-level intelligent agent, and determining the first query result based on the selection of the user on the at least one alternative query result.

Based on this, in an embodiment, the method further comprises:

In actual application, when the user does not select the at least one alternative query result, based on the first ranking result, the re-determining of the at least one alternative query result may be that the at least one alternative query result in the first round is determined according to a multi-round selection manner and the first ranking result; and under the condition that the user does not select at least one alternative query result in the first round, determining at least one alternative query result in a second round according to a multi-round selection mode and the first sequencing result. Wherein the at least one alternative query result in the first round and the at least one alternative query result in the second round are different.

Based on this, in an embodiment, the method further comprises:

In practical application, the number of times of obtaining and re-determining at least one alternative query result may be, under the condition that the user does not select the at least one alternative query result, counting the re-determined at least one alternative query result, and obtaining the number of times of counting the re-determined at least one alternative query result.

The first preset threshold may be determined according to actual conditions, and is not limited herein. As an example, the first preset threshold may be a maximum selection statement, and the setting of the maximum number of selection rounds may prevent a situation where there is no correct option in at least one alternative query result.

Based on this, in an embodiment, the method further comprises:

In actual application, the preset mode may be determined according to an actual situation, and is not limited herein. As an example, the preset manner may be that the at least one query result is selected by using a deep learning model through algorithm learning, so as to obtain a first query result. The selecting of the at least one query result using the deep learning model may be extracting a keyword from the query information, and selecting the at least one query result based on the keyword. For example, in the case where the user uses an oven to make a cake, it is determined that "do nothing with weak flour when making a cake? "in which a keyword of" no weak flour "is determined, and" with normal flour and wheat flour 4: a 1-ratio blend may replace the first query result of weak flour ". At this time, at least one query result matched with the query information is determined, and the first query result can be directly obtained in a preset mode without selection of a user.

Based on this, in an embodiment, the method further comprises:

In actual application, the condition that the user does not select the at least one query result may be determined according to an actual condition, which is not limited herein. As an example, the case that the user does not select the at least one query result may be that, if the at least one query result does not include a relevant result or a correct result related to the user input information, it is determined that the user does not select the at least one query result, at this time, the user transmits a signal that the at least one query result is not selected to the first device, and the first device sends the user input information with abnormal processing to the second device.

Based on this, in an embodiment, the method further comprises:

In actual application, the condition that the first preset model can normally process the input information may be determined according to an actual condition, which is not limited herein. As an example, the first preset model may be capable of normally processing the input information, where parameters of the first preset model are capable of correctly understanding the input information intention of the user, and a background knowledge base of the second device includes the input information intention of the user and a third query result corresponding to the input information intention of the user, and it is determined that the first preset model is capable of normally processing the input information.

In some embodiments, fig. 2 is a schematic diagram of a workflow of an information processing method according to an embodiment of the present application, and as shown in fig. 2, the workflow of the information processing method at least includes: acquiring the user input request information, and judging whether the intelligent function of the first equipment can correctly process the user input request information or not; processing the user input request information which cannot be correctly processed by the intelligent function, and converting the processed user input request information into an accurate query request; inputting different accurate query requests from different search interfaces to a search module according to different functions, and returning results to the user; the user selects a correct result from the returned results through the information interaction unit, wherein the information interaction unit can comprise a visual interaction interface; and taking the correct result selected by the user and the accurate query request as a difficult sample or a new sample, and collecting the difficult sample or the new sample to perform learning and updating of a second preset model of the second equipment and a background knowledge base of the second equipment. The above processes are repeated, the user input request information which is difficult to solve by the current intelligent function is searched by combining with the search engine, the user input request information which is difficult to process by the current first preset model is processed by selecting the search result answer of the visual interaction interface by the user, the original request and the user selection result are used as marked data, parameter learning is carried out on the cloud platform model, iterative updating is carried out on a background knowledge base, the adaptability of the intelligent function to the new sample is enhanced, and the generalization and universality of the model are continuously improved.

Correspondingly, an embodiment of the present application further provides an information processing method, where the method is applied to a second device, fig. 3 is a schematic flow chart of the information processing method according to the embodiment of the present application, and as shown in fig. 3, the method includes:

step 301: receiving user input information and a first query result which are sent by first equipment and used for processing abnormity;

step 302: based on the abnormal user input information and the first query result, learning and updating a second preset model to obtain an updated second preset model; the updated second preset model is used for updating the first preset model in the first equipment.

In practical application, the second device may be determined according to practical situations, and is not limited herein. As an example, the second device may be a network device, and the network device may be a device with self-learning functionality, e.g., a cloud platform with self-learning functionality.

In step 301, the second device includes a learning module and a storage module, where the learning module is configured with a deep learning network model for multiple intelligent tasks, and is configured to receive the user input information for processing the exception and the first query result, determine a deep learning model corresponding to the user input information for processing the exception and the first query result, and adjust and optimize the deep learning model to adapt to a new sample or a difficult sample; the storage module is used for storing a cloud platform model, intelligent function parameters of different versions, a sample database and background knowledge base data.

In step 302, the learning and updating of the second preset model based on the abnormal user input information and the first query result may be performed by using a preset learning model to learn the abnormal user input information and the first query result, and updating parameters of the second preset model. The preset learning model may be a model with an information learning function, and may be determined according to an actual situation, which is not limited herein. As an example, in the case that the voice information is the user input information with abnormal processing, the preset learning model may be a text understanding model; when the picture information is the user input information with abnormal processing, the preset learning model may be an image recognition model.

In some embodiments, the user input information for processing the exception and the first query result are added to a background knowledge base of the second device, and the background knowledge base of the second device is updated.

Wherein, in an embodiment, the method further comprises:

In practical application, the determining, based on the user input information with abnormal processing, the second query result matched with the user input information with abnormal processing may be determining, based on a preset processing mode and the user input information with abnormal processing, the second query result matched with the user input information with abnormal processing. The preset processing mode may be determined according to an actual situation, and is not limited herein. As an example, the preset processing manner may be a manual processing manner. And the second equipment receives the user input information for processing the abnormity, which is sent by the first equipment, and determines a second query result matched with the user input information for processing the abnormity based on a manual processing mode and the user input information for processing the abnormity.

And the second equipment sends the second query result to the first equipment, displays the second query result through the first equipment, displays a second query result matched with the user input information with abnormal processing to the user, and enables the user to obtain a correct result corresponding to the user input information with abnormal processing.

For understanding the embodiment of the present invention, the first device is an oven, and the second device is a cloud platform.

Fig. 4 is a schematic flowchart of an information processing method and an information processing method according to an embodiment of the present application, and as shown in fig. 4, the information processing method and the information processing method mainly include the following steps:

the first step is as follows: the oven obtains the user input information and judges whether the local model of the oven can correctly process the user input information. The oven acquires user input information through an information acquisition unit in the interactive module.

Under the condition that a user makes a cake by using an oven and uses an intelligent function of cooking knowledge question and answer, the user inputs 'how do without low-gluten flour' through a microphone with a voice acquisition function? The parameters of the local model of the oven can correctly understand the will of the user, and the background knowledge base of the cloud platform contains the question and the answer, namely, question: no weak flour, answer: mixing common flour with wheat flour 4:1 proportional mixing is replaceable ", then return the answer that corresponds with the question to report the answer through the stereo set that possesses the voice broadcast function.

The second step: the oven converts the user input information which cannot be normally processed into query information with a query function.

In the case that the local model of the oven cannot understand the user's question, or the background knowledge base of the cloud platform does not contain the question and the answer, the information of the user's question is subjected to diversified transformation processing, and the user's question is expanded into "do nothing with low gluten flour during cake making? "to enhance the relevance of the search results to the user's question.

And thirdly, the oven inputs the query information corresponding to the user input information into a search module of the oven from different search interfaces, and the relevant results of the user input information are obtained from the Internet.

The oven will "do nothing with weak flour when making cake? "the query statement is input into an open source encyclopedia knowledge question-answer search interface to obtain at least one answer.

And fourthly, processing at least one answer obtained by searching from the Internet by the oven and then sending the processed answer to the information interaction unit for the user to select.

The oven sorts at least one answer obtained by searching from the Internet according to the relevance of the question of the user, selects 3 or more than 3 answers in the at least one answer as alternative options, and displays the 3 or more than 3 answers on a display screen with a visual interaction function for the user to select.

In the case that there is no answer that the user considers correct in 3 or more answers of the first round, a multi-round selection mode may be adopted to jump to the next 3 or more answers, and the maximum number of selection rounds is set to prevent the case that there is no correct choice in the 3 or more answers.

And fifthly, judging whether at least one answer has a correct result or a related result by the user.

In the case that 3 or more than 3 answers recommended to the user do not include answers related to questions or correct answers, the user selects a button of "no suitable answer" on the display screen having the visual interaction function, and at this time, the toaster will "do nothing with weak flour when making cake? The query sentence is transmitted to the cloud platform, the cloud platform carries out manual processing on the query sentence to obtain a correct answer, the correct answer is returned to the oven, and the oven displays the correct answer on a display screen with a visual interaction function to provide the correct answer for a user.

The 3 or more answers recommended to the user include "ordinary flour and wheat flour 4:1 proportion mix can replace "similar answer, then the user check this answer based on more accurate correct label discrimination ability to click" confirm select "button, then the oven will question and answer, promptly" ask: no weak flour, answer: mixing common flour with wheat flour 4: a 1-ratio blend may replace the "pass back to cloud platform.

And a sixth step: the cloud platform collects the questions and answers sent by the oven opening in real time, new sample training learning is carried out through a supervised learning network, an updated cloud platform model is obtained, the local model of the oven is updated in real time through the cloud platform model of the cloud platform, and meanwhile, a background knowledge base of the cloud platform is updated.

The cloud platform receives the questions and answers sent by the oven, namely' question: no weak flour, answer: mixing common flour with wheat flour 4: and 1, after proportional mixing can replace the' step, learning the question and the answer by using a text understanding model and updating parameters of a cloud platform model of a cloud platform to obtain an updated cloud platform model, updating a local model of the oven in real time by using the cloud platform model of the cloud platform, and adding the question and the answer into a background knowledge base of the cloud platform by the oven to complete updating.

And seventhly, circulating the first step to the sixth step by the oven and the cloud platform, and continuously performing self-learning updating on the cloud platform model of the cloud platform and the background knowledge base of the cloud platform.

According to the information processing method and the information processing method provided by the embodiment of the application, the abnormal user input information processed by the first equipment is determined, the user input information is matched by using massive internet knowledge, at least one query result is returned for the user to select, and on one hand, the user obtains a correct answer in the at least one query result through self-selection; and on the other hand, the first equipment sends the user input information with abnormal processing and the correct answer to second equipment as labeled data, and the second equipment updates and learns a second preset model based on the labeled data to obtain the updated second preset model, so that the real-time iterative updating of the second preset model of the second equipment is realized.

The method has the advantages that the defect that the knowledge quantity of the original database is limited is overcome by using a search engine which is rich in knowledge and accurate in search recommendation, and the defect that the labor force is consumed due to the fact that the traditional database expansion needs to be carried out by secondary manual data marking is overcome by using the characteristic that a user has strong distinguishing capability on a correct answer as an advanced intelligent main body; the parameter updating is realized in real time by matching with a background self-learning cloud platform, so that the method is suitable for wider knowledge plane interaction and more accurate interaction effect. For users, the method and the system break through the situation that the traditional fixed function is limited in application condition and poor in updating, iteration and slow use experience, the range of solving problems of the users is expanded through a search and recommendation mode, for enterprises, manual consumption of accurate data marking is saved, and the function iteration process is accelerated.

In order to implement the information processing method according to the embodiment of the present application, an embodiment of the present application further provides an information processing apparatus, where the information processing apparatus is disposed on a first device, fig. 5 is a schematic structural diagram of the information processing apparatus according to the embodiment of the present application, and as shown in fig. 5, the information processing apparatus 500 includes:

a first determining module 501, configured to determine user input information for handling an exception;

a conversion module 502, configured to perform conversion processing on the abnormal user input information to obtain query information related to the abnormal user input information;

a second determining module 503, configured to determine at least one query result matching the query information;

a first receiving module 504, configured to receive a first query result; the first query result is any one of the at least one query result;

a first sending module 505, configured to send the abnormal user input information and the first query result to the second device; and the abnormal user input information and the first query result are used for the second equipment to perform learning updating.

In an embodiment, the first determining module 501 is further configured to receive the input information; judging whether the first preset model can normally process the input information or not; and under the condition that the first preset model cannot normally process the input information, determining the input information as the user input information with abnormal processing.

In one embodiment, the input information includes voice information and/or picture information; the first determining module 501 is further configured to determine, when the first preset model cannot normally process the voice information and/or the picture information, that the voice information and/or the picture information is the user input information with abnormal processing.

In an embodiment, the conversion module 502 is further configured to, when the voice information is the user input information with abnormal processing, perform information supplementation on the voice information to obtain query information related to the voice information.

In an embodiment, the conversion module 502 is further configured to, when the picture information is the user input information with abnormal processing, perform information extraction on the picture information to obtain query information related to the picture information.

In an embodiment, the information processing apparatus 500 further includes: a sorting module and a third determining module,

the sorting module is used for sorting the at least one query result to obtain a first sorting result;

the third determining module is configured to determine at least one alternative query result based on the first ranking result.

In an embodiment, the information processing apparatus 500 further includes: a fourth determining module, configured to re-determine at least one alternative query result based on the first ranking result if the user has not selected the at least one alternative query result; the re-determined at least one alternative query result is different from the at least one alternative query result.

In an embodiment, the information processing apparatus 500 further includes: an acquisition module and a stop module, wherein the acquisition module and the stop module,

the acquisition module is used for acquiring the times of re-determining at least one alternative query result;

the stopping module is configured to stop re-determining the at least one alternative query result when the number of times is greater than or equal to a first preset threshold.

In an embodiment, the information processing apparatus 500 further includes: and the selection module is used for selecting the at least one query result based on a preset mode to obtain a first query result.

In an embodiment, the information processing apparatus 500 further includes: a second sending module, configured to send the user input information with abnormal processing to the second device when the user does not select the at least one query result; and the abnormal user input information is used for the second equipment to determine a second query result.

In an embodiment, the information processing apparatus 500 further includes: and the fifth determining module is used for determining a third query result based on the first preset model under the condition that the first preset model can normally process the input information.

It should be noted that: in the information processing apparatus provided in the above embodiment, when performing information processing, only the division of each program module is illustrated, and in practical applications, the processing allocation may be completed by different program modules according to needs, that is, the internal structure of the apparatus may be divided into different program modules to complete all or part of the processing described above. In addition, the information processing apparatus and the information processing method provided by the above embodiments belong to the same concept, and specific implementation processes thereof are described in the method embodiments and are not described herein again.

In order to implement the information processing method according to the embodiment of the present application, an embodiment of the present application further provides an information processing apparatus, where the information processing apparatus is disposed on a second device, fig. 6 is a schematic structural diagram of the information processing apparatus according to the embodiment of the present application, and as shown in fig. 6, the information processing apparatus 600 includes:

a second receiving module 601, configured to receive user input information and a first query result, sent by the first device, for handling an exception;

an updating module 602, configured to perform learning updating on a second preset model based on the abnormal user input information and the first query result, so as to obtain an updated second preset model; the updated second preset model is used for updating the first preset model in the first equipment.

In one embodiment, the information processing apparatus 600 includes: a third receiving module and a sixth determining module,

the third receiving module is configured to receive user input information sent by the first device and used for processing the exception;

the sixth determining module is configured to determine, based on the abnormal user input information, a second query result matched with the abnormal user input information; and the second query result is used for displaying.

In an actual application scenario, fig. 7 is a schematic structural diagram of an information processing and information processing system according to an embodiment of the present application, and as shown in fig. 7, the system includes a first device 701 and a second device 702.

The first device 701 includes a user interaction module and a search module, and the user interaction module includes an interaction module and a control module. The interaction module comprises an information acquisition unit and an information interaction unit, wherein the information acquisition unit is used for acquiring the user input information; the information interaction unit is used for displaying the at least one query result to the user. The control module comprises a central control unit and a data transmission unit, wherein the central control unit is used for determining abnormal user input information, and the data transmission unit is used for carrying out data transmission with the second equipment. The search module is used for determining at least one query result matched with the query information from the Internet.

The second device 702 comprises a learning module and a storage module, wherein the learning module is provided with a deep learning network model for multiple intelligent tasks, and is used for receiving the abnormal user input information and the first query result, determining a deep learning model corresponding to the abnormal user input information and the first query result, and adjusting and optimizing the deep learning model to adapt to a new sample or a difficult sample; the storage module is used for storing a cloud platform model, intelligent function parameters of different versions, a sample database and background knowledge base data.

It should be noted that: in the information processing apparatus provided in the above embodiment, when performing information processing, only the division of each program module is exemplified, and in practical applications, the processing may be distributed to different program modules according to needs, that is, the internal structure of the apparatus may be divided into different program modules to complete all or part of the processing described above. In addition, the information processing apparatus and the information processing method provided in the above embodiments belong to the same concept, and specific implementation processes thereof are described in the method embodiments, and are not described herein again.

Based on the hardware implementation of the program module, and in order to implement the method on the first device side in the embodiment of the present application, an embodiment of the present application further provides a first device, and fig. 8 is a schematic structural diagram of the first device in the embodiment of the present application, and as shown in fig. 8, the first device 800 includes:

a first communication interface 801 capable of performing information interaction with a second device;

the first processor 802 is connected to the first communication interface 801 to implement information interaction with the second device, and is configured to execute a method provided by one or more technical solutions of the first device side when running a computer program. And the computer program is stored on the first memory 803.

Specifically, the first processor 802 is configured to determine user input information for handling an exception; converting the user input information which is processed abnormally to obtain query information related to the user input information which is processed abnormally; determining at least one query result matched with the query information from the Internet; receiving a first query result; the first query result is any one of the at least one query result;

the first communication interface 802 is configured to send the abnormal user input information and the first query result to a second device; and the abnormal user input information and the first query result are used for the second equipment to perform learning updating.

It should be noted that: the specific processing of the first processor 802 and the first communication interface 801 may be understood with reference to the methods described above.

Of course, in practice, the various components in the first device 800 are coupled together by the bus system 804. It is understood that the bus system 804 is used to enable communications among the components. The bus system 704 includes a power bus, a control bus, and a status signal bus in addition to a data bus. For clarity of illustration, however, the various buses are labeled as bus system 804 in FIG. 8.

The first memory 803 in the present embodiment is used to store various types of data to support the operation of the first device 800. Examples of such data include: any computer program for operating on the first device 800.

The method disclosed in the embodiments of the present application may be applied to the first processor 802, or implemented by the first processor 802. The first processor 802 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be implemented by integrated logic circuits of hardware or instructions in the form of software in the first processor 802. The first Processor 802 may be a general purpose Processor, a Digital Signal Processor (DSP), or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, etc. The first processor 802 may implement or perform the methods, steps, and logic blocks disclosed in the embodiments of the present application. The general purpose processor may be a microprocessor or any conventional processor or the like. The steps of the method disclosed in the embodiments of the present application may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in a storage medium located in the first memory 803, and the first processor 802 reads the information in the first memory 803, and completes the steps of the foregoing method in combination with its hardware.

In an exemplary embodiment, the first Device 800 may be implemented by one or more Application Specific Integrated Circuits (ASICs), DSPs, programmable Logic Devices (PLDs), complex Programmable Logic Devices (CPLDs), field Programmable Gate Arrays (FPGAs), general purpose processors, controllers, micro Controllers (MCUs), microprocessors (microprocessors), or other electronic components for performing the aforementioned methods.

Based on the hardware implementation of the program module, and in order to implement the method on the second device side in the embodiment of the present application, an embodiment of the present application further provides a second device, and fig. 9 is a schematic structural diagram of the second device in the embodiment of the present application, and as shown in fig. 9, the second device 900 includes:

a second communication interface 901, which is capable of performing information interaction with the first device;

the second processor 902 is connected to the second communication interface 901 to implement information interaction with the first device, and is configured to execute the method provided by one or more technical solutions of the second device side when running a computer program. And the computer program is stored on the second memory 903.

Specifically, the second communication interface 901 is configured to receive user input information and a first query result that are sent by the first device and used for handling exception;

the second processor 902 is configured to perform learning update on a second preset model based on the abnormal user input information and the first query result, so as to obtain an updated second preset model; the updated second preset model is used for updating the first preset model in the first equipment.

It should be noted that: the specific processing procedures of the second communication interface 901 and the second processor 902 can be understood with reference to the above-described methods.

Of course, in practice, the various components of the second device 900 are coupled together by a bus system 904. It is understood that the bus system 904 is used to enable communications among the components. The bus system 904 includes a power bus, a control bus, and a status signal bus in addition to a data bus. But for clarity of illustration the various buses are labeled as bus system 904 in figure 9.

The second memory 903 in the embodiments of the present application is used to store various types of data to support the operation of the second device 900. Examples of such data include: any computer program for operating on the second device 900.

The method disclosed in the embodiments of the present application may be applied to the second processor 902, or implemented by the second processor 902. The second processor 902 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be implemented by an integrated logic circuit of hardware or an instruction in the form of software in the second processor 902. The second processor 902 described above may be a general purpose processor, a DSP, or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, or the like. The second processor 902 may implement or perform the methods, steps and logic blocks disclosed in the embodiments of the present application. A general purpose processor may be a microprocessor or any conventional processor or the like. The steps of the method disclosed in the embodiments of the present application may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software modules may be located in a storage medium located in the second memory 903, and the second processor 902 reads the information in the second memory 903 and performs the steps of the foregoing methods in combination with its hardware.

In an exemplary embodiment, the second device 900 may be implemented by one or more ASICs, DSPs, PLDs, CPLDs, FPGAs, general-purpose processors, controllers, MCUs, microprocessors, or other electronic components for performing the aforementioned methods.

It is understood that the memories (the first memory 803 and the second memory 903) of the embodiments of the present application may be volatile memories or nonvolatile memories, and may also include both volatile and nonvolatile memories. Among them, the nonvolatile Memory may be a Read Only Memory (ROM), a Programmable Read Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), an Electrically Erasable Programmable Read-Only Memory (EEPROM), a magnetic random access Memory (FRAM), a magnetic random access Memory (Flash Memory), a magnetic surface Memory, an optical Disc, or a Compact Disc Read-Only Memory (CD-ROM); the magnetic surface storage may be disk storage or tape storage. Volatile Memory can be Random Access Memory (RAM), which acts as external cache Memory. By way of illustration, and not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), synchronous Static Random Access Memory (SSRAM), dynamic Random Access Memory (DRAM), synchronous Dynamic Random Access Memory (SDRAM), double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), double Data Rate Synchronous Random Access Memory (ESDRAM), enhanced Synchronous Dynamic Random Access Memory (ESDRAM), enhanced Synchronous Random Access Memory (DRAM), synchronous Random Access Memory (DRAM), direct Random Access Memory (DRmb Access Memory). The memories described in the embodiments of the present application are intended to comprise, without being limited to, these and any other suitable types of memory.

It should be noted that: "first," "second," and the like are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.

The technical means described in the embodiments of the present application may be arbitrarily combined without conflict.

The above description is only a preferred embodiment of the present application, and is not intended to limit the scope of the present application.

Claims

1. An information processing method, applied to a first device, the method comprising:

determining user input information for handling the exception;

determining at least one query result matching the query information;

sending the abnormal user input information and the first query result to second equipment; and the abnormal user input information and the first query result are used for the second equipment to perform learning updating.

2. The method of claim 1, wherein determining user input information to handle the exception comprises:

receiving the input information;

3. The method according to claim 2, characterized in that the input information comprises voice information and/or picture information; determining that the input information is the user input information with abnormal processing under the condition that the first preset model cannot normally process the input information, wherein the determining comprises the following steps:

4. The method of claim 3, wherein the transforming the exception handling user input information to obtain query information associated with the exception handling user input information comprises:

5. The method of claim 3, wherein the transforming the exception handling user input information to obtain query information associated with the exception handling user input information comprises:

6. The method of claim 1, wherein after determining at least one query result that matches the query information, the method further comprises:

sequencing the at least one query result to obtain a first sequencing result;

7. The method of claim 6, further comprising:

8. The method of claim 7, further comprising:

obtaining the times of redetermining at least one alternative query result;

and under the condition that the times are greater than or equal to a first preset threshold value, stopping re-determining at least one alternative query result.

9. The method of claim 1, further comprising:

10. The method of claim 1, further comprising:

11. The method of claim 2, further comprising:

12. An information processing method applied to a second device, the method comprising:

13. The method of claim 12, further comprising:

14. An information processing apparatus provided in a first device, the information processing apparatus comprising:

15. An information processing apparatus provided in a second device, the information processing apparatus comprising:

the updating module is used for learning and updating the second preset model based on the abnormal user input information and the first query result to obtain an updated second preset model; the updated second preset model is used for updating the first preset model in the first equipment.

16. A first device, comprising: a first processor and a first memory for storing a computer program capable of running on the processor,

wherein the first processor is adapted to perform the steps of the method of any one of claims 1 to 11 when running the computer program.

17. A second apparatus, comprising: a second processor and a second memory for storing a computer program capable of running on the processor,

wherein the second processor is adapted to perform the steps of the method of any of claims 12 to 13 when running the computer program.

18. A storage medium having stored thereon a computer program for performing the steps of the method of any one of claims 1 to 11, or for performing the steps of the method of any one of claims 12 to 13, when the computer program is executed by a processor.