WO2022188742A1 - Information processing method and apparatus, medium and electronic device - Google Patents

Abstract

Provided are an information processing method and apparatus, a medium and an electronic device. The method comprises: acquiring problem stem information of a word problem to be solved (S101); according to the problem stem information, generating a comprehensive equation for completely solving the word problem to be solved and a plurality of sub-equations for embodying a solving process (S102); for each sub-equation, generating analysis information, corresponding to the sub-equation, according to the problem stem information, the comprehensive equation and the sub-equation (S103); and generating, according to the plurality of sub-equations and the analysis information corresponding to each sub-equation, explanation information corresponding to the plurality of sub-equations (S104).

Description

Information processing method, apparatus, medium and electronic equipment

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is based on the CN application number 202110251093.5 and the filing date is March 8, 2021, and claims its priority. The disclosure of the CN application is hereby incorporated into this application as a whole.

technical field

The embodiments of the present disclosure relate to the field of computer applications, and in particular, to an information processing method, apparatus, medium, and electronic device.

Background technique

Mathematical application problems are extremely important for cultivating students' mathematical abstraction ability, and are the key and difficult points in mathematics teaching. In the related art, most of the application question assisted learning systems search in the database according to the application question stem input by the user to find the target question stem with the highest similarity with the text of the question stem input by the user, and then compare it with the target question stem. The title stem corresponds to the answer and parsing feedback to the user.

SUMMARY OF THE INVENTION

This Summary is provided to introduce concepts in a simplified form that are described in detail in the Detailed Description section that follows. This summary section is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to be used to limit the scope of the claimed technical solution.

Embodiments of the present disclosure provide an information processing method, apparatus, medium, and electronic device.

In a first aspect, an embodiment of the present disclosure provides an information processing method, including:

Obtain the stem information of the application questions to be solved;

According to the question stem information of the applied problem to be solved, a comprehensive formula for completely solving the applied problem to be solved and a plurality of sub-calculations for reflecting the solution process of the applied problem to be solved are generated, wherein the plurality of sub-calculations The calculation formula conforms to the operation logic of the comprehensive calculation formula;

For each of the sub-expressions, generate analysis information corresponding to the sub-expressions according to the question stem information of the application question to be solved, the comprehensive expression and the sub-expressions, wherein the analysis information includes the sub-expressions The meaning of the result of the calculation and the unit of the result of the sub-calculation;

According to the plurality of sub-calculations and the analysis information corresponding to each of the sub-calculations, explanation information corresponding to the plurality of sub-calculations is generated.

In a second aspect, an embodiment of the present disclosure provides an information processing apparatus, including:

The acquisition module is used to acquire the question stem information of the application question to be solved;

The first generation module is configured to generate a comprehensive formula for completely solving the to-be-solved applied problem and a comprehensive formula for reflecting the to-be-solved applied problem according to the question stem information of the to-be-solved applied problem obtained by the acquisition module A plurality of sub-calculations of the solution process, wherein, the plurality of sub-calculations conform to the operation logic of the comprehensive formula;

The second generating module is configured to, for each of the sub-expressions generated by the first generating module, obtain the problem stem information of the applied question to be solved obtained by the obtaining module, the data generated by the first generating module The integrated calculation formula and the sub-calculation generate analysis information corresponding to the sub-calculation, wherein the analysis information includes the meaning of the result of the sub-calculation and the unit of the result of the sub-calculation;

a third generation module, configured to generate explanations corresponding to the multiple sub-expressions according to the multiple sub-expressions generated by the first generation module and the analysis information corresponding to each of the sub-expressions generated by the second generation module information.

In a third aspect, an embodiment of the present disclosure provides a non-volatile computer-readable medium on which a computer program is stored, and when the program is executed by a processing apparatus, implements the steps of the method provided in the first aspect of the present disclosure.

In a fourth aspect, an embodiment of the present disclosure provides an electronic device, including: a storage device on which a computer program is stored; and a processing device for executing the computer program in the storage device to implement the first aspect of the present disclosure Provided the steps of the method.

In a fifth aspect, the present disclosure provides a computer program, comprising: instructions that, when executed by a processor, cause the processor to perform the method provided by the first aspect of the present disclosure.

In a sixth aspect, the present disclosure provides a computer program product comprising instructions that, when executed by a processor, cause the processor to perform the method provided by the first aspect of the present disclosure.

Other features and advantages of the present disclosure will be described in detail in the detailed description that follows.

Description of drawings

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

FIG. 1 is a flowchart of an information processing method according to an exemplary embodiment of the present disclosure.

Fig. 2 is a schematic structural diagram of a binary tree generated according to a target expression and identification information according to an exemplary embodiment of the present disclosure.

FIG. 3 is a flowchart of an information processing method according to another exemplary embodiment of the present disclosure.

FIG. 4 is a block diagram of an information processing apparatus according to an exemplary embodiment of the present disclosure.

FIG. 5 is a block diagram of an electronic device according to an exemplary embodiment of the present disclosure.

Detailed ways

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but rather are provided for the purpose of A more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes, and are not intended to limit the protection scope of the present disclosure.

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

As used herein, the term "including" and variations thereof are open-ended inclusions, ie, "including but not limited to". The term "based on" is "based at least in part on." The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Relevant definitions of other terms will be given in the description below.

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

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

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

Since the application questions are endless and ever-changing, it is impossible to cover all the application questions in the database. In this way, when the application question queried by the user is not in the database, only the answer to the most similar application question can be given, but the answer to the original question cannot be given, and the user's needs cannot be fully met.

In view of the above technical problems, embodiments of the present disclosure provide an information processing method, apparatus, medium, and electronic device.

FIG. 1 is a flowchart of an information processing method according to an exemplary embodiment of the present disclosure. As shown in FIG. 1 , the method includes S101 to S104.

In S101, the question stem information of the application question to be solved is obtained.

In the embodiments of the present disclosure, the application questions describe relevant facts in language or words, reflect a certain mathematical relationship (eg, quantitative relationship, positional relationship, etc.), and solve an unknown number of questions, wherein the application questions to be solved can be in the form of For general application questions that require users to give the solution process, they can also be presented in the form of multiple-choice questions or fill-in-the-blank questions.

The above-mentioned information processing method can be applied to electronic equipment, for example, a terminal equipment or a server communicatively connected to the terminal equipment. The terminal device may be, for example, a smart phone, a tablet computer, a personal computer, or the like. Among them, the user can input the question stem information of the application question to be solved to the terminal device in the form of image, text, audio, etc., and then the terminal device receives the question stem information and automatically solves it according to the question stem information, or the terminal device will The question stem information is sent to the server, so that the server can automatically solve the question according to the question stem information, and feedback the solution result to the user through the terminal device.

In S102, according to the question stem information of the applied problem to be solved, a comprehensive formula for completely solving the applied problem to be solved and a plurality of sub-calculations for reflecting the solution process of the applied problem to be solved are generated.

In the embodiment of the present disclosure, the above-mentioned multiple sub-calculations conform to the operation logic of the comprehensive calculation formula.

In S103, for each sub-expression, according to the question stem information of the application question to be solved, the comprehensive formula, and the sub-expression, the analysis information corresponding to the sub-expression is generated.

In the embodiment of the present disclosure, the analysis information includes the meaning of the result of the sub-expression and the unit of the result of the sub-expression.

In S104, according to the plurality of sub-expressions and the analysis information corresponding to each sub-expression, explanation information corresponding to the plurality of sub-expressions is generated.

In an embodiment of the present disclosure, the explanation information may include at least one of explanation text, explanation audio, and explanation video. In addition, explanation information corresponding to the multiple sub-calculations can be generated through a rule template according to the multiple sub-calculations and the analysis information corresponding to each sub-calculation.

In the above technical solution, after obtaining the question stem information of the applied problem to be solved, a comprehensive formula for completely solving the above applied problem to be solved and a plurality of sub-calculations used to reflect the solution process of the applied problem to be solved are generated according to it. . In this way, without relying on the database, the application problem to be solved input by the user can be answered in a targeted manner, which improves the accuracy of the answer. In addition, it can not only provide a comprehensive formula for completely solving the above-mentioned applied problem to be solved, but also can provide a plurality of sub-calculations used to reflect the solving process of the applied problem to be solved and the explanation information corresponding to the multiple sub-calculations, and the explanation information It includes the meaning and unit of the result of each sub-expression, which is convenient for users to quickly and thoroughly understand the problem-solving idea of the application problem to be solved, and improve the learning effect and efficiency.

For example, the stem information of the application question to be solved is as follows:

A garment factory makes an average of 1,850 sets of clothes every day in the first 12 days of June, and an average of 2,100 sets of clothes every day in the next 18 days. How many sets of clothes are made every day this month?

Through the above S102, a comprehensive formula (1850×12+2100×18)/(12+18)=2000 for completely solving the above-mentioned applied problem to be solved and four sub-calculations for reflecting the solving process of the applied problem to be solved can be obtained : 1850×12=22200, 2100×18=37800, 12+18=30, (22200+37800)/30=2000.

For each of the above four sub-calculations, the analytical information generated by the above step 103 is shown in Table 1 below:

Table 1 Analytical information corresponding to sub-expressions

sub-expression	Meaning of the result of the subexpression	the unit of the result of the subexpression
1850×12=22200	Clothes made in the previous 12 days	set
2100×18=37800	Clothes made after 18 days	set
12+18=30	total days	sky
(22200+37800)/30＝2000	clothes made every day	set

Exemplarily, the above-mentioned explanation information can be an explanation text. According to the above-mentioned four sub-calculations and the analysis information corresponding to each sub-calculation shown in Table 1, the explanation text corresponding to the four sub-calculations generated by the rule template is as follows:

Dear students, let's take a look at this question together!

In order to obtain the clothes made every day, we first need to ask for the clothes made in the previous 12 days, then the clothes made in the first 12 days are 1850×12=22200 sets.

Next, we ask for the clothes made in the last 18 days, then the clothes made in the last 18 days are 2100×18=37800 sets.

Then, we find the total number of days, then the total number of days is 12+18=30 days.

Finally, let's ask for the clothes we make every day. Using 22,200 sets of clothes made in the first 12 days, 37,800 sets of clothes made in the next 18 days and a total of 30 days to calculate, the clothes made every day are (22,200+37,800)/30=2,000 sets.

In this way, the problem is all solved. Don't forget to write your answer!

In the above technical solution, after obtaining the question stem information of the applied problem to be solved, a comprehensive formula for completely solving the above applied problem to be solved and a plurality of sub-calculations used to reflect the solution process of the applied problem to be solved are generated according to it. . In this way, without relying on the database, the application problem to be solved input by the user can be answered in a targeted manner, which improves the accuracy of the answer. In addition, it can not only provide a comprehensive formula for completely solving the above-mentioned applied problem to be solved, but also can provide a plurality of sub-calculations used to reflect the solving process of the applied problem to be solved and the explanation information corresponding to the multiple sub-calculations, and the explanation information It includes the meaning and unit of the results of each sub-calculation, which is convenient for users to quickly and thoroughly understand the problem-solving idea of the application problem to be solved, and improve the learning effect and efficiency.

The specific implementation of generating a comprehensive formula for completely solving the applied problem to be solved and a plurality of sub-calculations for reflecting the solution process of the applied problem to be solved according to the question stem information of the applied problem to be solved in the above S102 will be described in detail below. . Specifically, it can be realized by the following steps (1) to (3).

(1) Input the question stem information of the application question to be solved into the first coding network of the solving model to obtain the encoded question stem information.

(2) Input the encoded question stem information into the first decoding network of the solving model, and obtain the target expression and identification information used to characterize whether each operator in the target expression is used as a unique operator in a sub-expression.

(3) According to the target expression, a comprehensive formula is generated, and according to the target expression and the identification information, a plurality of sub-calculations are generated.

In the embodiment of the present disclosure, the algebraic expression in the comprehensive calculation formula is composed of each element in the target expression. The solution model may be, for example, a sequence-to-sequence model, in particular, the solution model may include a first encoding network and a first decoding network. Wherein, the first encoding network is used for encoding the question stem information of the application question to be solved, and the first decoding network is used for generating the target expression and representing the target expression according to the encoded question stem information output by the first encoding network Whether each operator in the formula is used as the identification information of the unique operator in a sub-calculation.

Exemplarily, the first encoding network may be a Recurrent Neural Network (RNN), Convolutional Neural Networks (CNN), a pre-trained language model RoBERTa (a Robustly Optimized BERT Pretraining Approach, a more robust The best optimized BERT pre-training method), etc., among which, the full English name of BERT is Bidirectional Encoder Representation from Transformers, that is, a bidirectional language representation model. The decoding network can be, for example, an RNN, a CNN, a multi-layer Transformer, or the like.

In addition, the above-mentioned target expression may be a prefix expression, an infix expression, a postfix expression, or the like.

For example, for the application question to be solved for calculating the daily output of a garment factory in the above example, the first decoding network generates the target expression according to the encoded question stem information as "/+×1850 12×2100 18+12 18( Prefix expression)", "1850×12+2100×18/12+18(infix expression)", "1850 12×2100 18×12 18+/(postfix expression)".

Preferably, the above-mentioned target expression can be a prefix expression, which is more in line with human thinking habits, and does not need to use parentheses to indicate calculation priorities, making the solution model easier to train.

In addition, the above-mentioned solving model can be obtained by training in the following manner: obtaining first reference information of the first reference application question, wherein the first reference information includes the first reference question stem information, and the information used to completely solve the above-mentioned first reference application question. The first reference comprehensive calculation formula and a plurality of first reference sub-calculations used to embody the solution process of the first reference applied problem, the first reference applied problem can be the applied problem in the question bank; then, according to the first reference comprehensive formula, obtain The reference expression and the reference identification information used to characterize whether each operator in the reference expression is used as a unique operator in a sub-expression; finally, by using the first reference question stem information as the input of the solution model, the reference expression The formula and reference identification information are used as the target output of the solution model to train the model, and the solution model is obtained.

The specific implementation of generating a comprehensive formula according to the target expression in the above step (3) will be described in detail below.

Specifically, the target expression can be directly converted into an algebraic expression according to the form of the target expression (the form can be one of a prefix expression, an infix expression, and a postfix expression); then, the converted result is calculated. The result of the algebraic formula; finally, the algebraic formula obtained after conversion, the equal sign, and the result of the algebraic formula obtained after conversion can be connected in turn to obtain a comprehensive formula.

For example, the target expression obtained by the above step (2) is an infix expression, the infix expression is 1850×12+2100×18/12+18, and the infix expression 1850×12+2100× 18/12+18 is converted into the algebraic formula (1850×12+2100×18)/(12+18); then, the calculated algebraic formula (1850×12+2100×18)/(12+18) results in 2000; finally , the converted algebraic formula (1850×12+2100×18)/(12+18), =, and the corresponding result 2000 can be connected in turn to obtain the comprehensive formula (1850×12+2100×18)/(12+18 )=2000.

The specific implementation of generating multiple sub-expressions according to the target expression and the identification information in the above step (3) will be described in detail below.

Specifically, a binary tree can be generated first according to the target expression and identification information; then, a plurality of sub-expressions can be generated according to the binary tree.

In the embodiment of the present disclosure, the nodes of the above binary tree are composed of numbers and operators in the target expression. Wherein, the attribute information of the number node in the binary tree includes an attribute value, and the attribute value is the number itself; the attribute information of the operator node in the binary tree includes a flag used to characterize whether the operator node is the only operator in a sub-calculation operator (that is, used to characterize whether the operator node is split) attribute value. That is, when the identifier indicates that the operator node is split, or, when the identifier indicates that the operator node is not split, and the operator node is the root node, the left child of the operator node can be The attribute value of the operator node, the operator represented by the operator node, and the attribute value of the right child node of the operator node constitute an element of the algebraic expression in the sub-expression. At this time, the attribute value of the operator node is the attribute value of its left child node. The result obtained after the value and the attribute value of the right child node are operated by the operator represented by the operator node; when the identifier indicates that the operator node is not divided, and the operator node is not the root node, the operator A node at least needs to form an algebraic expression of a sub-expression together with its parent node. At this time, the attribute value of the operator node is the left parenthesis, the attribute value of its left child node, the operator represented by the operator node, and the attribute of its right child node. The algebraic expression obtained by connecting the value and the closing parenthesis in sequence.

For example, the above identifier is 0 or 1, wherein 0 is used to indicate that the operator node is not divided, and 1 is used to indicate that the operator node is divided.

Illustratively, for the application problem to be solved for calculating the daily output of a garment factory in the above example, according to the target expression and identification information, the generated binary tree is shown in FIG. 2 . As can be seen from Figure 2, the binary tree includes five operator nodes and six numeric nodes.

Among them, the identifier corresponding to the first operator node "×" on the left side of the third layer of the binary tree shown in Figure 2 is 1, and its attribute value is 22200 (ie 1850×12), then the operator node "×" is divided, at this time, the attribute value 1850 of the left child node "1850" of the operator node "x", the operator "x" represented by the operator node "x", the right side of the operator node "x" The attribute value 12 of the child node "12" constitutes an element of an algebraic expression in a sub-expression, that is, 1850, ×, and 12 constitute an element of an algebraic expression in a sub-expression.

The identifier corresponding to the second operator node "×" on the left side of the third layer of the binary tree shown in Figure 2 is 1, and its attribute value is 37800 (ie 2100×18), then the operator node "×" is divided , at this time, the attribute value 2100 of the left child node "2100" of the operator node "×", the operator "×" represented by the operator node "×", and the right child node of the operator node "×" The attribute value 18 of "18" constitutes an element of an algebraic expression in another sub-expression, that is, 2100, ×, and 18 constitute an element of an algebraic expression in another sub-expression.

The identifier corresponding to the first operator node "+" on the left side of the second layer of the binary tree shown in Figure 2 is 0, and its attribute value is (22200+37800), then the operator node "+" is not divided, Since the operator node "+" is not the root node, at this time, the operator node "+" needs to at least form the algebraic expression of the sub-expression together with its parent node "/".

The identifier corresponding to the second operator node "+" on the left side of the second layer of the binary tree shown in Figure 2 is 1, and its attribute value is 30 (ie 12+18), then the operator node "+" is divided , at this time, the attribute value 12 of the left child node "12" of the operator node "+", the operator "+" represented by the operator node "+", and the right child node of the operator node "+" The attribute value 18 of "18" constitutes the elements of the algebraic part of a sub-expression, that is, 12, +, 18 constitute the elements of the algebraic expression in a sub-expression.

The identifier corresponding to the operator node "/" (ie the root node) of the first layer of the binary tree shown in Figure 2 is 0, and its attribute value is 2000 (ie (22200+37800)/30), then the operator node "/" is not divided. At this time, the attribute value (22200+37800) of the left child node "+" of the operator node "/", the operator "/" represented by the operator node "/", the The attribute value 30 of the right child node "+" of the operator node "/" constitutes the elements of the algebraic part of another sub-expression, namely (22200+37800), /, 30 constitute the elements of the algebraic expression in the other sub-expression.

Specifically, according to the above binary tree, multiple sub-calculations can be generated in the following ways:

First, for each operator in the binary tree: according to the identifier of the operator node, determine whether it is divided; if the operator node is divided, or the operator node is not divided, and the operator node is the root node, then connect the attribute value of the left child node of the operator node, the operator represented by the operator node, and the attribute value of the right child node of the operator node in order to obtain an algebraic expression of a sub-expression, and then, Connect the algebraic expression, the equal sign, and the attribute value of the operator node in turn to obtain a sub-expression; if the operator node is not divided and the operator node is not the root node, the sub-expression will not be generated, that is, it will not be executed. any action.

Illustratively, for the application question to be solved for calculating the daily output of a garment factory in the above example, a binary tree is generated according to the encoded question stem information as shown in FIG. 2 . As can be seen from Figure 2, the binary tree includes five operator nodes and six numeric nodes.

For the first operator node "×" on the left side of the third layer of the binary tree shown in Figure 2, its corresponding identifier is 1, and its attribute value is 22200 (ie 1850×12), then the operator node " ×” is divided, at this time, the attribute value 1850 of the left child node “1850” of the operator node, the operator “×” represented by the operator node “×”, and the right child node of the operator node “ The attribute value 12 of 12" is connected in turn to obtain the algebraic formula 1850×12 of a sub-calculation. After that, the algebraic formula 1850×12, =, and the attribute value 22200 of the operator node "×" can be connected in sequence to obtain a sub-calculation. 1850×12=22200.

For the second operator node "×" on the left side of the third layer of the binary tree shown in Figure 2, its corresponding identifier is 1, and its attribute value is 37800 (ie 2100×18), then the operator node " ×” is divided, at this time, the attribute value 2100 of the left child node “2100” of the operator node “×”, the operator “×” represented by the operator node “×”, and the operator node “×” The attribute value 18 of the right child node "18" of "" is connected in turn to obtain the algebraic formula 2100×18 of another sub-expression. After that, the algebraic formula 2100×18, =, the attribute value of the operator node "×" can be 37800 Connect in sequence to obtain another sub-calculation 2100×18=37800.

For the first operator node "+" on the left side of the second level of the binary tree shown in Figure 2, its corresponding identifier is 0, and its attribute value is (22200+37800), then the operator node "+" It is not divided, because the operator node "+" is not the root node, in this case, no sub-expression is generated, that is, no operation is performed.

For the second operator node "+" on the left side of the second level of the binary tree shown in Figure 2, its corresponding identifier is 1, and its attribute value is 30 (ie 12+18), then the operator node " +" is divided, at this time, the attribute value 12 of the left child node "12" of the operator node "+", the operator "+" represented by the operator node "+", the operator node "+" The attribute value 18 of the right child node "18" of "" is connected in turn to obtain the algebraic formula 12+18 of another sub-expression. After that, the algebraic formula 12+18, =, the attribute value 30 of the operator node "+" Connect in sequence to obtain another sub-calculation 12+18=30.

The identifier corresponding to the operator node "/" (ie the root node) of the first layer of the binary tree shown in Figure 2 is 0, and its attribute value is 2000 (ie (22200+37800)/30), then the operation The operator node "/" is not divided. Since the operator node "+" is the root node, at this time, the attribute value (22200+37800) of the left child node "+" of the operator node "/", the operator The operator "/" represented by the symbol node "/" and the attribute value 30 of the right child node "+" of the operator node "/" are connected in turn to obtain the algebraic formula (22200+37800)/30 of another sub-expression, Afterwards, the algebraic formula (22200+37800)/30, =, and the attribute value 2000 of the operator node "/" can be connected in sequence to obtain another sub-calculation (22200+37800)/30=2000.

From this, the following four sub-calculations can be obtained: 1850×12=22200, 2100×18=37800, 12+18=30, (22200+37800)/30=2000.

The specific implementation of generating the analytical information corresponding to the sub-calculation according to the question stem information of the application question to be solved, the comprehensive formula and the sub-calculation in the above S103 will be described in detail below.

Specifically, the question stem information, the comprehensive formula and the sub-calculation can be sequentially spliced and input into the multi-task sequence-to-sequence model to obtain the analytical information corresponding to the sub-calculation. In this way, the meaning of the result of the sub-calculation and the unit of the result of the sub-calculation can be obtained through a model, thereby reducing the complexity of the auxiliary learning system for applied questions, and the meaning of the result of the sub-calculation and the unit of the result of the sub-calculation are related to each other. , which is conducive to mutual learning, thereby improving the accuracy of the analytical information generated by the multi-task sequence-to-sequence model.

In an embodiment of the present disclosure, the multi-task sequence-to-sequence model may include a second encoding network, a second decoding network, and a first classifier. Wherein, the second encoding network is used to encode the splicing information to obtain the coded information, and the splicing information is the information obtained by splicing the question stem information, the comprehensive formula and the sub-calculation of the application question to be solved in turn; the second decoding network , which is used to generate the meaning of the result of the sub-calculation according to the coding information; the first classifier is used to generate the unit of the result of the sub-calculation according to the coding information.

For example, the second encoding network may be RNN, CNN, a pre-trained language model RoBERTa, etc., and the second decoding network may be, for example, RNN, CNN, a multi-layer Transformer, and the like.

The above multi-task sequence-to-sequence model can be obtained by training in the following manner: first, acquiring second reference information of the second reference application question, wherein the second reference information includes the stem information of the second reference question, which is used to completely solve the above-mentioned first question. The second reference comprehensive formula of the second reference application question, a plurality of second reference sub-calculations used to reflect the solution process of the second reference application problem, and the annotation analysis information corresponding to each second reference sub-calculation, the second reference application problem It can be an applied question in the question bank, and the annotation analysis information includes the meaning of the result of each second reference sub-equation and the unit of the result of each second reference sub-equation; The reference question stem information, the second reference comprehensive formula, and the second reference sub-formula spliced in sequence are used as the input of the multi-task sequence-to-sequence model, and the annotation parsing information corresponding to the second reference sub-formula is used as the multi-task. The model is trained in the way of the target output of the sequence-to-sequence model, and a multi-task sequence-to-sequence model is obtained.

It should be noted that the above-mentioned second reference application questions may be the same as or different from the above first reference application questions, which are not specifically limited in this disclosure.

In addition, in order to facilitate the user to more quickly and thoroughly understand the problem-solving idea of the application question to be solved, so as to further improve the learning effect and efficiency, the analysis information of the sub-calculation generated in the above S103 also includes the knowledge points involved in the sub-calculation. In this way, the analysis information may include the meaning of the result of the sub-expression, the unit of the result of the sub-expression, and the knowledge points involved in the sub-expression.

Illustratively, for the application problem to be solved for calculating the daily output of a garment factory in the above example, the above four sub-formulas (1850×12=22200, 2100×18=37800, 12+18=30, (22200+37800)/30= 2000), the analytical information generated by above-mentioned S103 is shown in the following table 2:

Table 2 Analytical information corresponding to sub-expressions

Exemplarily, the above-mentioned explanation information can be an explanation text. According to the above-mentioned four sub-calculations and the analysis information corresponding to each sub-calculation shown in Table 2, the explanation text corresponding to the four sub-calculations generated by the rule template is as follows:

Dear students, let's take a look at this question together!

In order to get the clothes made every day, we first need to ask for the clothes made in the previous 12 days. Through the formula "average × number = total number (×, ÷)", it can be calculated that the clothes made in the first 12 days are 1850 × 12 = 22200 sets.

Next, we asked for clothes made after 18 days. Using the same formula as the previous step, it can be calculated that the clothes made in the next 18 days are 2100×18=37800 sets.

Then, we ask for the total number of days. Through the formula Partial Quantity 1 + Partial Quantity 2 = Total, it can be calculated that the total number of days is 12+18=30 days.

Finally, let's ask for the clothes we make every day. Using 22,200 sets of clothes made in the first 12 days, 37,800 sets of clothes made in the next 18 days and a total of 30 days to calculate, the clothes made every day are (22,200+37,800)/30=2,000 sets.

In this way, the problem is all solved. Don't forget to write your answer!

When the analysis information includes the meaning of the result of the sub-calculation, the unit of the result of the sub-calculation, and the knowledge points involved in the sub-calculation, the above S103 can generate the corresponding sub-calculation in the following manner according to the question stem information, the comprehensive calculation formula and the sub-calculation. Parsing information: The question stem information, the comprehensive formula and the sub-calculation are sequentially spliced and input into the multi-task sequence-to-sequence model to obtain the analytical information corresponding to the sub-calculation. In this way, three kinds of information, including the meaning of the result of the sub-calculation, the unit of the result of the sub-calculation, and the knowledge points involved in the sub-calculation, can be obtained through one model, thereby reducing the complexity of the auxiliary learning system for application questions. The meaning of the result, the unit of the result of the sub-calculation, and the knowledge points involved in the sub-calculation are related to each other, which is conducive to mutual learning, thereby improving the accuracy of the analytical information generated by the multi-task sequence-to-sequence model.

In the embodiment of the present disclosure, the multi-task sequence-to-sequence model includes, in addition to the second encoding network, the second decoding network, and the first classifier, a second classifier, wherein the second classifier is used for encoding information according to the above-mentioned encoding information. , and generate the knowledge points involved in the sub-calculation.

At this time, during the training process of the multi-task sequence-to-sequence model, the annotation parsing information corresponding to each second reference sub-formula includes the meaning of the result of each second reference sub-formula, the meaning of each second reference sub-formula In addition to the unit of the result, it also includes the knowledge points involved in each second reference sub-calculation.

In addition, in order to facilitate the user to more quickly and thoroughly understand the problem-solving ideas of the applied problems to be solved, and to further improve the learning effect and efficiency, the above explanation information can also include the stem information of the applied problems to be solved, and the algebraic expressions in the comprehensive formulas. The meaning of each number that appears. Specifically, as shown in FIG. 3, before the above-mentioned S104, the above-mentioned method further includes the following step S105:

In S105, according to the question stem information of the application question to be solved, the meaning of each number appearing in the algebraic expression in the comprehensive calculation formula in the question stem information of the application question to be solved is generated.

For example, for the application question to be solved for calculating the daily output of a garment factory in the above example, in the question stem information, the algebraic formula (1850 in the comprehensive formula (1850×12+2100×18)/(12+18)=2000) The numbers appearing in ×12+2100×18)/(12+18) include 1850, 12, 2100, and 18. The meaning of each number determined through the above steps is shown in Table 3 below:

Table 3 Number meaning table

number	meaning of numbers
12	none
18	none
1850	Average number of clothes made per day in the first 12 days of June
2100	Average number of clothes made per day for 18 days after June

In this way, the above S104 can generate explanation information corresponding to the multiple sub-calculations according to the multiple sub-calculations, the analysis information corresponding to each sub-calculation, and the meaning of each number.

Exemplarily, the above-mentioned explanation information can be an explanation text, according to the above-mentioned four sub-calculations, the corresponding analysis information of each sub-calculation shown in Table 2, and the meaning of each number shown in Table 3, this generated by the rule template. The explanation texts corresponding to the four sub-calculations are as follows:

Dear students, let's take a look at this question together!

In order to get the clothes made every day, we first need to ask for the clothes made in the previous 12 days. We know that the average number of clothes made every day in the first 12 days of June is 1850. Through the formula "average × number = total number (×, ÷)", we can calculate that the clothes made in the first 12 days are 1850 × 12 = 22200 sets .

Next, we asked for clothes made after 18 days. Using the same formula as the previous step, the average number of clothes made every day in the 18 days after June is 2100 and 18 for calculation, and the clothes made in the next 18 days are 2100×18=37800 sets.

Then, we ask for the total number of days. Through the formula Partial Quantity 1 + Partial Quantity 2 = Total, it can be calculated that the total number of days is 12+18=30 days.

Finally, let's ask for the clothes we make every day. Using 22,200 sets of clothes made in the first 12 days, 37,800 sets of clothes made in the next 18 days and a total of 30 days to calculate, the clothes made every day are (22,200+37,800)/30=2,000 sets.

In this way, the problem is all solved. Don't forget to write your answer!

The specific implementation of generating the meaning of each number appearing in the algebraic formula in the comprehensive formula in the question stem information of the applied problem to be solved according to the question stem information of the applied question to be solved in the above S105 will be described in detail below. Specifically, it can be achieved through the following steps:

First, replace each number in the question stem information with a preset identifier; then, generate a coding vector corresponding to each number according to the question stem information obtained after the replacement; finally, for each number, according to the code corresponding to the number A vector that generates the meaning of that number.

Wherein, a sequence-to-sequence generation model can be used to generate a coding vector corresponding to each digit according to the replaced question stem information, and for each digit, according to the coding vector corresponding to the digit, the meaning of the digit is generated. Specifically, the generation model may include an encoding network and a decoding network, wherein the encoding network is used to generate an encoding vector corresponding to each digit according to the replaced stem information, and the decoding network is used for each digit, according to the digit The corresponding encoded vector, generating the meaning of the number. Illustratively, the above-mentioned generative model may be, for example, a BERT model.

In addition, the above-mentioned generative model can be obtained by training in the following manner: first, obtain third reference information of the third reference application question, wherein the third reference information includes the stem information of the third reference question and the information used to completely solve the above-mentioned third reference application The third reference comprehensive formula of the question, the third reference applied problem can be the applied problem in the question bank; then, in the third reference question stem information, each reference number that appears in the algebraic expression in the third reference comprehensive formula is replaced by Preset identification, and mark the meaning of each reference number; finally, for each reference number, by using the third reference stem information obtained after replacement as the input of the generation model, the meaning of the marked reference number is used as the generation model The model is trained in the way of the target output of the model, and the generative model is obtained.

It should be noted that the above-mentioned third reference application questions may be the same as or different from the above first reference application questions, that is, the training samples for solving the model may be the same as or different from the training samples for generating the model. The application questions may be the same as or different from the second reference application questions above, that is, the training samples of the generated model may be the same as or different from the training samples of the above-mentioned multi-task sequence-to-sequence model, which are not specifically limited in this document.

FIG. 4 is a block diagram of an information processing apparatus according to another exemplary embodiment of the present disclosure. As shown in FIG. 4 , the device 400 includes: an acquisition module 401 for acquiring the question stem information of the application questions to be solved; a first generation module 402 for obtaining the application questions to be solved according to the acquisition module 401 the question stem information, and generate a comprehensive formula for completely solving the applied problem to be solved and multiple sub-calculations for reflecting the solution process of the applied problem to be solved, wherein, the multiple sub-calculations meet the requirements of the comprehensive formula Operation logic; the second generation module 403 is configured to, for each of the sub-calculations generated by the first generation module 402, obtain the question stem information of the application question to be solved obtained by the obtaining module 401, the The comprehensive calculation formula and the sub-calculation generated by the first generation module 402 generate analysis information corresponding to the sub-calculation, wherein the analysis information includes the meaning of the result of the sub-calculation and the result of the sub-calculation. unit; the third generation module 404 is configured to generate the multiple sub-expressions according to the plurality of sub-expressions generated by the first generation module 402 and the analysis information corresponding to each of the sub-expressions generated by the second generation module 403 The explanation information corresponding to the sub-calculation.

In the above technical solution, after obtaining the question stem information of the applied problem to be solved, a comprehensive formula for completely solving the above applied problem to be solved and a plurality of sub-calculations used to reflect the solution process of the applied problem to be solved are generated according to it. . In this way, without relying on the database, the application problem to be solved input by the user can be answered in a targeted manner, which improves the accuracy of the answer. In addition, it can not only provide a comprehensive formula for completely solving the above-mentioned applied problem to be solved, but also can provide a plurality of sub-calculations used to reflect the solving process of the applied problem to be solved and the explanation information corresponding to the multiple sub-calculations, and the explanation information It includes the meaning and unit of the results of each sub-calculation, which is convenient for users to quickly and thoroughly understand the problem-solving idea of the application problem to be solved, and improve the learning effect and efficiency.

Optionally, the first generation module 402 includes: a first input sub-module for inputting the question stem information of the application question to be solved into the first coding network of the solving model to obtain the encoded question stem information, Wherein, the solution model includes the first encoding network and the first decoding network; the second input sub-module is used to input the encoded question stem information into the first decoding network to obtain the target expression and Identification information for characterizing whether each operator in the target expression is used as a unique operator in a sub-calculation; the first generation sub-module is used to generate the comprehensive calculation formula according to the target expression, and according to the The target expression and the identification information are used to generate the plurality of sub-expressions, wherein the algebraic expression in the comprehensive expression is composed of elements in the target expression.

Optionally, the target expression is a prefix expression.

Optionally, the analysis information further includes knowledge points involved in the sub-calculation.

Optionally, the second generation module 403 is used to input the question stem information of the application question to be solved, the comprehensive formula and the sub-calculation into the multi-task sequence-to-sequence model in turn to obtain the result. The analytical information corresponding to the prescriptive expression.

Optionally, the multi-task sequence-to-sequence model includes a second encoding network, a second decoding network, a first classifier, and a second classifier; wherein the second encoding network is used to encode the splicing information , obtain coding information, and the splicing information is the information obtained by splicing the question stem information of the applied question to be solved, the comprehensive formula and the sub-calculation in sequence; the second decoding network is used for according to the coding information, to generate the meaning of the result of the sub-calculation; the first classifier is used to generate the unit of the result of the sub-calculation according to the coding information; the second classifier is used to generate the unit of the result of the sub-calculation according to the Encoding information to generate knowledge points involved in the sub-calculation.

Optionally, the apparatus 400 further includes: a fourth generation module, configured to generate, according to the question stem information of the applied question to be solved, the question stem information of the applied question to be solved and in the comprehensive formula. The meaning of each number appearing in the algebraic formula; the third generation module 404 is configured to generate the plurality of sub-calculations according to the plurality of sub-calculations, the analytical information corresponding to each of the sub-calculations, and the meaning of each of the numbers The explanation information corresponding to the formula.

Optionally, the fourth generation module includes: a replacement submodule for replacing each of the numbers in the question stem information of the applied question to be solved with a preset identifier; the second generation submodule for According to the question stem information obtained after the replacement, a coding vector corresponding to each of the numbers is generated; the third generation sub-module is used to generate the meaning of the numbers according to the coding vector corresponding to the numbers for each of the numbers .

Embodiments of the present disclosure further provide a computer-readable medium on which a computer program is stored, and when the program is executed by a processing apparatus, implements the steps of the above-mentioned information processing method provided by the embodiments of the present disclosure.

Referring next to FIG. 5 , it shows a schematic structural diagram of an electronic device (eg, a terminal device or a server) 500 suitable for implementing an embodiment of the present disclosure. Terminal devices in the embodiments of the present disclosure may include, but are not limited to, such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablets), PMPs (portable multimedia players), vehicle-mounted terminals (eg, mobile terminals such as in-vehicle navigation terminals), etc., and stationary terminals such as digital TVs, desktop computers, and the like. The electronic device shown in FIG. 5 is only an example, and should not impose any limitation on the function and scope of use of the embodiments of the present disclosure.

As shown in FIG. 5 , an electronic device 500 may include a processing device (eg, a central processing unit, a graphics processor, etc.) 501 that may be loaded into random access according to a program stored in a read only memory (ROM) 502 or from a storage device 508 Various appropriate actions and processes are executed by the programs in the memory (RAM) 503 . In the RAM 503, various programs and data required for the operation of the electronic device 500 are also stored. The processing device 501, the ROM 502, and the RAM 503 are connected to each other through a bus 504. An input/output (I/O) interface 505 is also connected to bus 504 .

Typically, the following devices can be connected to the I/O interface 505: input devices 506 including, for example, a touch screen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; including, for example, a liquid crystal display (LCD), speakers, vibration An output device 507 such as a computer; a storage device 508 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 509 . Communication means 509 may allow electronic device 500 to communicate wirelessly or by wire with other devices to exchange data. While FIG. 5 shows electronic device 500 having various means, it should be understood that not all of the illustrated means are required to be implemented or provided. More or fewer devices may alternatively be implemented or provided.

In particular, according to embodiments of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program carried on a non-transitory computer readable medium, the computer program containing program code for performing the method illustrated in the flowchart. In such an embodiment, the computer program may be downloaded and installed from the network via the communication device 509, or from the storage device 508, or from the ROM 502. When the computer program is executed by the processing apparatus 501, the above-mentioned functions defined in the methods of the embodiments of the present disclosure are executed.

It should be noted that the computer-readable medium mentioned above in the present disclosure may be a computer-readable signal medium or a computer-readable storage medium, or any combination of the above two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus or device, or a combination of any of the above. More specific examples of computer readable storage media may include, but are not limited to, electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable Programmable read only memory (EPROM or flash memory), fiber optics, portable compact disk read only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing. In the embodiments of the present disclosure, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. Rather, in embodiments of the present disclosure, a computer-readable signal medium may include a data signal in baseband or propagated as part of a carrier wave, carrying computer-readable program code therein. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device . Program code embodied on a computer readable medium may be transmitted using any suitable medium including, but not limited to, electrical wire, optical fiber cable, RF (radio frequency), etc., or any suitable combination of the foregoing.

In some embodiments, terminal devices and servers can use any currently known or future developed network protocols such as HTTP (HyperText Transfer Protocol) to communicate, and can communicate with digital data in any form or medium Communication (eg, a communication network) interconnects. Examples of communication networks include local area networks ("LAN"), wide area networks ("WAN"), the Internet (eg, the Internet), and peer-to-peer networks (eg, ad hoc peer-to-peer networks), as well as any currently known or future development network of.

The above-mentioned computer-readable medium may be included in the above-mentioned electronic device; or may exist alone without being assembled into the electronic device.

The above-mentioned computer-readable medium carries one or more programs, and when the above-mentioned one or more programs are executed by the electronic device, the electronic device: obtains the question stem information of the application question to be solved; Question stem information to generate a comprehensive formula for completely solving the applied problem to be solved and a plurality of sub-calculations for reflecting the solution process of the applied problem to be solved, wherein, the plurality of sub-calculations conform to the operation of the comprehensive formula Logic; for each of the sub-expressions, generate analysis information corresponding to the sub-expressions according to the question stem information of the applied question to be solved, the comprehensive expression and the sub-expressions, wherein the analysis information includes all Describe the meaning of the result of the sub-calculation and the unit of the result of the sub-calculation; according to the multiple sub-calculations and the analysis information corresponding to each of the sub-calculations, generate explanation information corresponding to the multiple sub-calculations.

Computer program code for performing operations of the present disclosure may be written in one or more programming languages, including but not limited to object-oriented programming languages—such as Java, Smalltalk, C++, and This includes conventional procedural programming languages - such as the "C" language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (eg, using an Internet service provider to via Internet connection).

The flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code that contains one or more logical functions for implementing the specified functions executable instructions. It should also be noted that, in some alternative implementations, the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It is also noted that each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations, can be implemented in dedicated hardware-based systems that perform the specified functions or operations , or can be implemented in a combination of dedicated hardware and computer instructions.

The modules involved in the embodiments of the present disclosure may be implemented in software or hardware. Wherein, the name of the module does not constitute a limitation of the module itself under certain circumstances, for example, the acquisition module can also be described as "a module for acquiring the question stem information of the applied question to be solved".

The functions described herein above may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), Systems on Chips (SOCs), Complex Programmable Logical Devices (CPLDs) and more.

In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in connection with the instruction execution system, apparatus or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media may include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media would include one or more wire-based electrical connections, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), fiber optics, compact disk read only memory (CD-ROM), optical storage, magnetic storage, or any suitable combination of the foregoing.

According to one or more embodiments of the present disclosure, Example 1 provides an information processing method, including: acquiring question stem information of an applied question to be solved; The comprehensive formula of the applied problem to be solved and a plurality of sub-calculations used to embody the solution process of the applied problem to be solved, wherein the plurality of sub-calculations conform to the operation logic of the comprehensive formula; for each of the sub-calculations , according to the question stem information of the application question to be solved, the comprehensive calculation formula and the sub-calculation, generate the analysis information corresponding to the sub-calculation, wherein the analysis information includes the meaning of the result of the sub-calculation and all The unit of the result of the sub-calculation is described; according to the plurality of sub-calculations and the analysis information corresponding to each of the sub-calculations, explanation information corresponding to the plurality of sub-calculations is generated.

According to one or more embodiments of the present disclosure, Example 2 provides the method of Example 1, wherein according to the question stem information of the to-be-solved applied question, a comprehensive formula for completely solving the to-be-solved applied question and the method of The multiple sub-calculations that embody the solution process of the application question to be solved include: inputting the question stem information of the application question to be solved into the first coding network of the solving model, and obtaining the encoded question stem information, wherein the The solution model includes the first encoding network and the first decoding network; the encoded question stem information is input into the first decoding network to obtain the target expression and the operations used to characterize the target expression. Whether the operator is used as the identification information of the unique operator in a sub-calculation; according to the target expression, generate the comprehensive calculation formula, and according to the target expression and the identification information, generate the multiple sub-calculations, wherein, The algebraic expression in the comprehensive expression is composed of the elements in the target expression.

According to one or more embodiments of the present disclosure, Example 3 provides the method of Example 2, the target expression being a prefix expression.

According to one or more embodiments of the present disclosure, Example 4 provides the method of Example 1, and the parsing information further includes knowledge points involved in the sub-calculation.

According to one or more embodiments of the present disclosure, Example 5 provides the method of Example 4, wherein according to the question stem information of the applied question to be solved, the comprehensive formula and the sub-calculation, the corresponding sub-calculation is generated The analytical information, including: the question stem information of the application question to be solved, the comprehensive formula and the sub-calculation are sequentially spliced and input into the multi-task sequence-to-sequence model, and the analytical information corresponding to the sub-calculation is obtained. .

According to one or more embodiments of the present disclosure, Example 6 provides the method of Example 5, the multi-task sequence-to-sequence model comprising a second encoding network, a second decoding network, a first classifier, and a second classifier; Wherein, the second encoding network is used to encode the splicing information to obtain the coded information, and the splicing information is obtained by splicing the stem information of the application question to be solved, the comprehensive formula and the sub-calculation in sequence. the obtained information; the second decoding network is used for generating the meaning of the result of the sub-calculation according to the coding information; the first classifier is used for generating the sub-calculation according to the coding information The unit of the result; the second classifier is configured to generate the knowledge points involved in the sub-calculation according to the encoded information.

According to one or more embodiments of the present disclosure, Example 7 provides the method of any one of Examples 1-6, the method further comprising: generating the application to be solved according to the question stem information of the application to be solved The meaning of each number that appears in the algebraic formula in the question stem information of the question; the multiple sub-calculations are generated according to the multiple sub-calculations and the analytical information corresponding to each of the sub-calculations The corresponding explanation information includes: generating explanation information corresponding to the plurality of sub-calculations according to the plurality of sub-calculations, the analysis information corresponding to each of the sub-calculations, and the meaning of each of the numbers.

According to one or more embodiments of the present disclosure, Example 8 provides the method of Example 7, wherein generating the question stem information of the applied question to be solved according to the question stem information of the applied question to be solved, in the The meaning of each number that appears in the algebraic formula in the comprehensive formula includes: replacing each of the numbers in the question stem information of the applied question to be solved with a preset identifier; according to the question stem information obtained after the replacement, generating A coding vector corresponding to each of the numbers; for each of the numbers, the meaning of the number is generated according to the coding vector corresponding to the number.

According to one or more embodiments of the present disclosure, Example 9 provides an information processing apparatus, including: an acquisition module for acquiring the question stem information of an application question to be solved; a first generation module for obtaining the information according to the acquisition module The obtained question stem information of the applied problem to be solved generates a comprehensive formula for completely solving the applied problem to be solved and a plurality of sub-calculations for reflecting the solution process of the applied problem to be solved, wherein the A plurality of sub-calculations conform to the operation logic of the comprehensive calculation formula; a second generation module is used for each of the sub-calculations generated by the first generation module, according to the obtained application problem to be solved by the acquisition module Question stem information, the comprehensive formula and the sub-calculation generated by the first generating module, generate analysis information corresponding to the sub-calculation, wherein the analysis information includes the meaning of the result of the sub-calculation and the The unit of the result of the sub-calculation; the third generation module is configured to generate all the sub-calculations according to the plurality of sub-calculations generated by the first generation module and the analysis information corresponding to each of the sub-calculations generated by the second generation module. Describe the explanation information corresponding to the multiple sub-calculations.

According to one or more embodiments of the present disclosure, Example 10 provides a non-transitory computer-readable medium having stored thereon a computer program that, when executed by a processing apparatus, implements any of Examples 1-8. steps of the method described.

According to one or more embodiments of the present disclosure, Example 11 provides an electronic device, including: a storage device on which a computer program is stored; and a processing device for executing the computer program in the storage device to The steps of implementing the following method: obtaining the question stem information of the application question to be solved; generating a comprehensive formula for completely solving the application question to be solved and a comprehensive formula used to represent the application question to be solved according to the question stem information of the application question to be solved. A plurality of sub-calculations of the solving process of an applied problem, wherein the plurality of sub-calculations conform to the operation logic of the comprehensive formula; for each of the sub-calculations, according to the question stem information of the applied problem to be solved, the comprehensive formula and the sub-calculation, generating analysis information corresponding to the sub-calculation, wherein the analysis information includes the meaning of the result of the sub-calculation and the unit of the result of the sub-calculation; according to the multiple sub-calculations and each The analysis information corresponding to the sub-expressions is used to generate explanation information corresponding to the plurality of sub-expressions.

According to one or more embodiments of the present disclosure, Example 12 provides the electronic device of Example 11, the comprehensive formula and The multiple sub-calculations used to embody the solution process of the applied problem to be solved include: inputting the problem stem information of the applied problem to be solved into the first coding network of the solving model, and obtaining the encoded problem stem information, wherein, The solution model includes the first encoding network and the first decoding network; the encoded question stem information is input into the first decoding network to obtain a target expression and a target expression for characterizing each of the target expressions. Whether the operator is used as the identification information of the unique operator in a sub-calculation; according to the target expression, generate the comprehensive calculation formula, and according to the target expression and the identification information, generate the multiple sub-calculations, wherein , the algebraic expression in the comprehensive calculation formula is composed of each element in the target expression.

According to one or more embodiments of the present disclosure, Example 13 provides the electronic device of Example 12, the target expression being a prefix expression.

According to one or more embodiments of the present disclosure, Example 14 provides the electronic device of Example 11, and the analysis information further includes knowledge points involved in the sub-calculation.

According to one or more embodiments of the present disclosure, Example 15 provides the electronic device of Example 14, the sub-equation is generated according to the stem information of the application question to be solved, the comprehensive equation and the sub-equation Corresponding analysis information includes: splicing the question stem information of the application question to be solved, the comprehensive formula and the sub-calculation in sequence, and then inputting them into the multi-task sequence-to-sequence model to obtain the analysis corresponding to the sub-calculation information.

According to one or more embodiments of the present disclosure, Example 16 provides the electronic device of Example 15, the multitasking sequence-to-sequence model including a second encoding network, a second decoding network, a first classifier, and a second classifier Wherein, the second coding network is used to encode the splicing information to obtain coding information, and the splicing information is to sequentially splicing the question stem information, the comprehensive formula and the sub-calculation of the application question to be solved. The second decoding network is used to generate the meaning of the result of the sub-calculation according to the coding information; the first classifier is used to generate the sub-calculation according to the coding information The unit of the result; the second classifier is configured to generate the knowledge points involved in the sub-calculation according to the encoding information.

According to one or more embodiments of the present disclosure, Example 17 provides the electronic device of any one of Examples 11-16, and the method further includes the step of: generating the The meaning of each number that appears in the algebraic formula in the comprehensive formula in the question stem information of the application question to be solved; the generation of the The explanation information corresponding to the multiple sub-calculations includes: generating explanation information corresponding to the multiple sub-calculations according to the multiple sub-calculations, the analysis information corresponding to each of the sub-calculations, and the meaning of each of the numbers.

According to one or more embodiments of the present disclosure, Example 18 provides the electronic device of Example 17, wherein the question stem information of the applied question to be solved is generated according to the question stem information of the applied question to be solved, in all The meaning of each number that appears in the algebraic formula in the comprehensive formula includes: replacing each of the numbers in the question stem information of the applied question to be solved with a preset identifier; according to the question stem information obtained after the replacement, A coding vector corresponding to each of the numbers is generated; for each of the numbers, the meaning of the number is generated according to the coding vector corresponding to the number.

According to one or more embodiments of the present disclosure, there is provided a computer program comprising: instructions which, when executed by a processor, cause the processor to perform the method provided by the first aspect of the present disclosure.

According to one or more embodiments of the present disclosure, there is provided a computer program product comprising instructions which, when executed by a processor, cause the processor to perform the method provided by the first aspect of the present disclosure.

The above description is merely a preferred embodiment of the present disclosure and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of the disclosure involved in the present disclosure is not limited to the technical solutions formed by the specific combination of the above-mentioned technical features, and should also cover, without departing from the above-mentioned disclosed concept, the technical solutions formed by the above-mentioned technical features or Other technical solutions formed by any combination of its equivalent features. For example, a technical solution is formed by replacing the above features with the technical features disclosed in the present disclosure (but not limited to) with similar functions.

Additionally, although operations are depicted in a particular order, this should not be construed as requiring that the operations be performed in the particular order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Likewise, although the above discussion contains several implementation-specific details, these should not be construed as limitations on the scope of the present disclosure. Certain features that are described in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination.

Although the subject matter has been described in language specific to structural features and/or logical acts of method, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are merely example forms of implementing the claims. Regarding the apparatus in the above-mentioned embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be described in detail here.

Claims (14)

1. An information processing method, comprising:

   Obtain the stem information of the application questions to be solved;

   According to the question stem information of the applied problem to be solved, a comprehensive formula for completely solving the applied problem to be solved and a plurality of sub-calculations for reflecting the solution process of the applied problem to be solved are generated, wherein the plurality of sub-calculations The calculation formula conforms to the operation logic of the comprehensive calculation formula;

   For each of the plurality of sub-expressions, according to the question stem information of the applied question to be solved, the comprehensive formula and the sub-expressions, the analysis information corresponding to the sub-expression is generated, wherein the analysis information includes the the meaning of the result of the sub-expression and the unit of the result of said sub-expression;

   According to the plurality of sub-calculations and the analysis information corresponding to each of the plurality of sub-calculations, explanation information corresponding to the plurality of sub-calculations is generated.
2. The method according to claim 1, wherein, according to the question stem information of the applied problem to be solved, a comprehensive formula for completely solving the applied problem to be solved and a formula for reflecting the applied problem to be solved are generated. Multiple sub-calculations of the solution process, including:

   Input the question stem information of the application question to be solved into the first coding network of the solving model, and obtain the encoded question stem information, wherein the solving model includes the first coding network and the first decoding network;

   The question stem information after described coding is input in described first decoding network, obtains target expression and is used to characterize whether each operator in described target expression is as the identification information of the unique operator in a sub-calculation;

   According to the target expression, the comprehensive calculation formula is generated, and according to the target expression and the identification information, the plurality of sub-calculations are generated, wherein the algebraic formula in the comprehensive calculation formula is determined by the target expression. composition of each element.
3. The method of claim 2, wherein the target expression is a prefix expression.
4. The method according to claim 1, wherein the analysis information further includes knowledge points involved in the sub-calculation.
5. The method according to claim 4, wherein, generating the analytical information corresponding to the sub-calculation according to the question stem information of the applied problem to be solved, the comprehensive formula and the sub-calculation, comprising:

   The stem information of the application question to be solved, the comprehensive formula and the sub-calculation are sequentially spliced and input into the multi-task sequence-to-sequence model to obtain the analytical information corresponding to the sub-calculation.
6. The method of claim 5, wherein the multi-task sequence-to-sequence model includes a second encoding network, a second decoding network, a first classifier, and a second classifier,

   The second encoding network is used to encode the splicing information to obtain the coded information, and the splicing information is obtained by sequentially splicing the stem information of the applied problem to be solved, the comprehensive formula and the sub-calculation. information,

   The second decoding network is used to generate the meaning of the result of the sub-calculation according to the encoding information,

   the first classifier, configured to generate the unit of the result of the sub-calculation according to the encoding information,

   The second classifier is configured to generate the knowledge points involved in the sub-calculation according to the encoding information.
7. The method of any one of claims 1-6, further comprising:

   According to the question stem information of the applied question to be solved, the meaning of each number in the question stem information of the applied question to be solved and the meaning of each number appearing in the algebraic formula in the comprehensive calculation formula are generated;

   Wherein, generating the explanation information corresponding to the multiple sub-calculations according to the multiple sub-calculations and the analysis information corresponding to each of the multiple sub-calculations includes:

   According to the plurality of sub-calculations, analysis information corresponding to each of the plurality of sub-calculations, and the meaning of each number, explanation information corresponding to the plurality of sub-calculations is generated.
8. The method according to claim 7, wherein the meaning of each number in the question stem information of the applied question to be solved is generated according to the question stem information of the applied question to be solved, in the comprehensive calculation formula The meaning of each number that appears in the algebraic expression of , including:

   Replace each number in the question stem information of the applied question to be solved with a preset identifier;

   According to the question stem information obtained after the replacement, the coding vector corresponding to each number is generated;

   For each number, the meaning of the number is generated according to the encoding vector corresponding to the number.
9. The method according to any one of claims 1-6, wherein the question stem information includes image information, text information or audio information, and the explanation information includes at least one of explanation text, explanation audio, and explanation video.
10. An information processing device, comprising:

    The acquisition module is used to acquire the question stem information of the application question to be solved;

    The first generation module is configured to generate a comprehensive formula for completely solving the to-be-solved applied problem and a comprehensive formula for reflecting the to-be-solved applied problem according to the question stem information of the to-be-solved applied problem obtained by the acquisition module A plurality of sub-calculations of the solution process, wherein, the plurality of sub-calculations conform to the operation logic of the comprehensive formula;

    The second generation module is configured to, for each of the plurality of sub-expressions, generate analytical information corresponding to the sub-expressions according to the question stem information of the applied question to be solved, the comprehensive expression and the sub-expressions, wherein, The analysis information includes the meaning of the result of the sub-expression and the unit of the result of the sub-expression;

    The third generating module is configured to generate explanation information corresponding to the multiple sub-calculations according to the multiple sub-calculations and the analysis information corresponding to each of the multiple sub-calculations.
11. A non-volatile computer-readable medium having stored thereon a computer program, wherein the program, when executed by a processing device, implements the steps of the method of any one of claims 1-9.
12. An electronic device comprising:

    a storage device on which a computer program is stored;

    A processing device is configured to execute the computer program in the storage device to implement the steps of the method of any one of claims 1-9.
13. A computer program comprising:

    Instructions which, when executed by a processor, cause the processor to perform the method of any of claims 1-9.
14. A computer program product comprising instructions which, when executed by a processor, cause the processor to perform the method of any of claims 1-9.

Images