Teaching plan generation method and device, computing equipment and storage medium

Abstract

The invention provides a teaching plan generation method and device, computing equipment and a storage medium. For each position of a teaching plan template, two alternative character sets and probabilities of alternative characters in the respective sets are provided; the target probability of the alternative characters is determined, the probability of the alternative characters in the first alternative character set and the probability of the alternative characters in the second alternative character set are comprehensively considered, and the comprehensiveness of the target probability is improved; through the target probability of the alternative characters, the target character at the position is determined in the set formed by the first alternative character set and the second alternative character set, and the search range of the target character is expanded; the target teaching plan is generated according to the target characters on all the positions of the teaching plan template, the comprehensiveness and comprehensiveness of teaching plan generation are improved, and the problem that an existing teaching plan generation method is single in consideration factor is solved.

Description

Technical Field

[0001] This specification relates to the field of artificial intelligence technology, and in particular to a teaching plan generation method, apparatus, computing device, and storage medium. Background Art

[0002] A lesson plan is a document that teachers use to plan and organize teaching activities for the purpose of carrying out a course. It contains teaching objectives, teaching content, teaching methods, learning activities, etc. It helps teachers to systematically arrange and organize teaching activities to ensure the effectiveness and orderliness of teaching.

[0003] In the field of educational informatization, research on computer-assisted lesson plan generation has made considerable progress. Among existing technologies, one type generates lesson plans based on textbook content. This technology typically extracts knowledge points from textbook text through methods such as text analysis and keyword extraction, and then generates lesson plans according to a preset template structure. However, this technology focuses solely on textbook content and lacks the incorporation of expert experience. Another type generates lesson plans based on example teaching materials. This method generates new lesson plans by matching and imitating the format and content of existing excellent lesson plan examples. However, this approach overly relies on the experience of historical lesson plans and is difficult to adapt to different teaching scenarios and personalized teaching needs. Both of these existing technologies suffer from the problem of considering relatively limited factors when generating lesson plans, resulting in a lack of comprehensiveness and integration. Summary of the Invention

[0004] To overcome the problems existing in the related art, this specification provides a teaching plan generation method, apparatus, computing device and storage medium.

[0005] According to a first aspect of an embodiment of this specification, a teaching plan generation method is provided, the method comprising:

[0006] For each position of the teaching plan template, obtaining a first candidate character set corresponding to the position and the probability of each candidate character in the first candidate character set;

[0007] Obtaining a second candidate character set corresponding to the position and the probability of each candidate character in the second candidate character set;

[0008] determining, based on the probability of each candidate character in the first candidate character set and the probability of each candidate character in the second candidate character set, a target probability for each candidate character in the first candidate character set and a target probability for each candidate character in the second candidate character set;

[0009] determining a target character for the position in a collection of the first candidate character set and the second candidate character set based on target probabilities of the candidate characters in the first candidate character set and target probabilities of the candidate characters in the second candidate character set;

[0010] Generate target lesson plans based on the target characters at various positions of the lesson plan template.

[0011] According to a second aspect of the embodiments of this specification, a teaching plan generating device is provided, the device comprising:

[0012] A first acquisition module is configured to acquire, for each position in the teaching plan template, a first candidate character set corresponding to the position and a probability of each candidate character in the first candidate character set;

[0013] A second acquisition module is configured to acquire a second candidate character set corresponding to the position and a probability of each candidate character in the second candidate character set;

[0014] a target probability determination module, configured to determine a target probability for each candidate character in the first candidate character set and a target probability for each candidate character in the second candidate character set based on the probability of each candidate character in the first candidate character set and the probability of each candidate character in the second candidate character set;

[0015] a target character determination module, configured to determine a target character for the position in a collection of the first candidate character set and the second candidate character set based on the target probabilities of the candidate characters in the first candidate character set and the target probabilities of the candidate characters in the second candidate character set;

[0016] The target lesson plan generation module is used to generate a target lesson plan based on the target characters at various positions of the lesson plan template.

[0017] According to a third aspect of the embodiments of this specification, a computing device is provided, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the operations performed by the teaching plan generation method described in the first aspect are implemented.

[0018] According to a fourth aspect of the embodiments of this specification, a computer-readable storage medium is provided, on which a program is stored, and the program is used by a processor to execute the operations performed by the teaching plan generation method described in the first aspect above.

[0019] According to a fifth aspect of the embodiments of this specification, a computer program product is provided, which includes a computer program, and when the computer program is executed by a processor, it implements the operations performed by the teaching plan generation method described in the first aspect above.

[0020] The technical solutions provided by the embodiments of this specification may have the following beneficial effects:

[0021] For each position of the lesson plan template, two alternative character sets and the probabilities of the alternative characters in their respective sets are provided; the determination of the target probability of the alternative characters comprehensively considers the probability of the alternative characters in the first alternative character set and the probability of the alternative characters in the second alternative character set, thereby improving the comprehensiveness of the target probability; through the target probability of the alternative characters, the target character of the position is determined in the collection consisting of the first alternative character set and the second alternative character set, thereby expanding the search range of the target character; according to the target characters at each position of the lesson plan template, a target lesson plan is generated, thereby improving the comprehensiveness and comprehensiveness of the lesson plan generation, and solving the problem of the existing lesson plan generation method considering only a single factor.

[0022] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the specification and, together with the description, serve to explain the principles of the specification.

[0024] Figure 1 This is a schematic diagram of an application scenario of a teaching plan generation method provided by the present invention;

[0025] Figure 2 It is a schematic diagram of a "compression-expansion" framework for lesson plan generation provided by the present invention;

[0026] Figure 3 It is a schematic diagram of a teaching plan generating device provided by the present invention;

[0027] Figure 4 It is a structural schematic diagram of a computing device provided by the present invention. DETAILED DESCRIPTION

[0028] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with this specification. Rather, they are merely examples of apparatuses and methods consistent with certain aspects of this specification, as detailed herein.

[0029] The terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit this specification. The singular forms "a," "the," and "the" used in this specification are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0030] It should be understood that although the terms first, second, third, etc. may be used in this specification to describe various information, such information should not be limited to these terms. These terms are merely used to distinguish information of the same type from one another. For example, first information may also be referred to as second information, and similarly, second information may also be referred to as first information without departing from the scope of this specification. Depending on the context, the term "if" as used herein may be interpreted as "when," "when," or "in response to determining."

[0031] See also Figure 1 , Figure 1 This is a flow chart of a teaching plan generation method provided by the present invention. Figure 1 As shown, the method includes the following steps:

[0032] S101. For each position of the teaching plan template, obtain a first candidate character set corresponding to the position and the probability of each candidate character in the first candidate character set.

[0033] The teaching plan template in the present invention includes one or more positions, and the characters at the positions are to be determined.

[0034] The first candidate character set and the probability of each candidate character in the first candidate character set can reflect a selection range and a selection strategy of a character selection method.

[0035] The first candidate character set includes one or more candidate characters, and the candidate characters are characters that may appear in the position.

[0036] The probability of each candidate character in the first candidate character set is the probability of selecting the candidate character as the character at the position when determining the character at the position based on the first candidate character set and the probabilities of each candidate character in the first candidate character set.

[0037] For example, assuming that the first alternative character set includes character A and character B, the probability of character A in the first alternative character set is 80%, and the probability of character B in the first alternative character set is 70%; then the character at the position is determined based on the first alternative character set and the probabilities of each alternative character in the first alternative character set. When selecting the character at the position, the probability of selecting character A as the character at the position is 80%, and the probability of selecting character B as the character at the position is 70%.

[0038] As a possible implementation method, S101 obtains the first alternative character set corresponding to the position and the probability of each alternative character in the first alternative character set, including: compressing at least two lesson plan samples that match the target course information to generate the first alternative character set corresponding to the position of the first lesson plan and the probability of each alternative character in the first alternative character set; the first lesson plan is a lesson plan obtained by compressing at least two lesson plan samples that match the target course information; the position of the first lesson plan corresponds one-to-one to the position of the lesson plan template.

[0039] The position of the first lesson plan in the present invention corresponds one-to-one to the position of the lesson plan template. The first alternative character set corresponding to this position of the first lesson plan and the probability of each alternative character in the first alternative character set are the first alternative character set corresponding to this position of the lesson plan target and the probability of each alternative character in the first alternative character set.

[0040] The target lesson plan to be generated may be a lesson plan corresponding to the target course information, where the target course information includes subject, grade, chapter, text, etc. For example, the target course information is "the first lesson of the first unit of the first volume of the seventh grade Chinese textbook for junior high school students".

[0041] The first candidate character set corresponding to the position in the first lesson plan includes one or more candidate characters. The candidate characters in the first candidate character set are the candidate characters that may appear at the position when the first lesson plan is obtained by compressing at least two lesson plan examples that match the target course information. The probability of each candidate character in the first candidate character set is the probability of selecting the candidate character as the character at the position when the first lesson plan is obtained by compressing at least two lesson plan examples that match the target course information.

[0042] In this way, the first candidate character set and the probability of each candidate character in the first candidate character set can reflect the scope and strategy of selecting characters in the lesson plan generation method of obtaining a lesson plan by matching at least two lesson plan samples with the target course information.

[0043] In some embodiments, the lesson plan samples that match the target course information can be lesson plans designed by teachers, experts, etc. for the target course information. In this way, compressing these lesson plan samples can absorb the teaching wisdom of human teachers and refine the teachers' existing experience. The first alternative character set and the probability of each alternative character in the first alternative character set can reflect the teachers' existing experience and the teaching wisdom of human teachers.

[0044] In some embodiments, expert lesson plan samples can be obtained in real time through RAG (Retrieval Augmented Generation) to form a teaching strategy reference pool, and lesson plan samples corresponding to the target course information can be determined in the teaching strategy reference pool to solve the problem that the static model cannot adapt to the latest teaching cases.

[0045] As a possible implementation method, at least two lesson plan samples that match the target course information are compressed to generate a first set of alternative characters corresponding to the position of the first lesson plan and the probability of each alternative character in the first set of alternative characters, including: inputting the lesson plan compression prompt word, the target course information and at least two lesson plan samples that match the target course information into a trained lesson plan compressor, so that the lesson plan compressor outputs the first set of alternative characters corresponding to the position of the first lesson plan and the probability of each alternative character in the first set of alternative characters.

[0046] In some embodiments, the lesson plan sample is a natural language text, and the large language model is a deep learning model trained using a large amount of text data, which can understand the meaning of language text and generate natural language text; therefore, the large language model can be used as a lesson plan compressor to compress at least two lesson plan samples that match the target course information to obtain a first lesson plan compressed based on the above-mentioned lesson plan samples; when generating the lesson plan, the lesson plan compressor will calculate and output at least two alternative characters corresponding to each position on the lesson plan and the probability of the alternative characters, that is, output the first alternative character set corresponding to the position of the first lesson plan and the probability of each alternative character in the first alternative character set.

[0047] In some embodiments, the general large language model can be trained for domain adaptation through supervised fine-tuning based on a self-built professional lesson plan dataset, so that the general large language model can internalize educational principles such as constructivism and cognitive load theory to obtain a lesson plan compressor. By adjusting parameters, the lesson plans output by the lesson plan compressor can conform to the sentence structure and knowledge organization of the educational scenario.

[0048] S102: Obtain a second candidate character set corresponding to the position and the probability of each candidate character in the second candidate character set.

[0049] In the present invention, the first candidate character set and the probability of each candidate character in the first candidate character set, the second candidate character set and the probability of each candidate character in the second candidate character set correspond to two different character selection methods, that is, the above information corresponds to two different teaching plan generation methods.

[0050] For the description of the second candidate character set corresponding to the position and the probability of each candidate character in the second candidate character set, please refer to the description of the first candidate character set corresponding to the position and the probability of each candidate character in the first candidate character set in S101, which will not be repeated here.

[0051] In some embodiments, the second candidate character set corresponding to the position and the probability of each candidate character in the second candidate character set, and the probability of each candidate character in the first candidate character set and the second candidate character set corresponding to the position can be information obtained by different methods for the same target course information. The target course information includes information such as subject, grade, chapter, and text.

[0052] It should be noted that the candidate characters included in the first candidate character set and the candidate characters included in the second candidate character set may be completely identical, partially identical, or completely different, and the present invention does not impose any limitation on this.

[0053] As a possible implementation method, S102 obtains the second alternative character set corresponding to the position and the probability of each alternative character in the second alternative character set, including: expanding the teaching material content corresponding to the target course information to generate the second alternative character set corresponding to the position of the second lesson plan and the probability of each alternative character in the second alternative character set; the second lesson plan is a lesson plan obtained by expanding the teaching material content corresponding to the target course information; the position of the second lesson plan corresponds one-to-one to the position of the lesson plan template.

[0054] The position of the second lesson plan in the present invention corresponds one-to-one to the position of the lesson plan template. The second alternative character set corresponding to this position of the second lesson plan and the probability of each alternative character in the second alternative character set are the second alternative character set corresponding to this position of the lesson plan target and the probability of each alternative character in the second alternative character set.

[0055] The second candidate character set corresponding to the position in the second lesson plan includes one or more candidate characters. The candidate characters in the second candidate character set are candidate characters that may appear at the position when the second lesson plan is obtained by expanding the teaching material corresponding to the target course information. The probability of each candidate character in the second candidate character set is the probability of selecting the candidate character as the character at the position when the second lesson plan is obtained for the teaching material corresponding to the target course information.

[0056] In this way, the second candidate character set and the probability of each candidate character in the second candidate character set can reflect the scope and strategy of selecting characters in the lesson plan generation method of expanding the teaching material content corresponding to the target course information to obtain the lesson plan.

[0057] As a possible implementation method, the teaching material content corresponding to the target course information is expanded to generate a second set of alternative characters corresponding to the position of the second lesson plan and the probability of each alternative character in the second set of alternative characters, including: inputting the teaching material expansion prompt words, the target course information and the teaching material content corresponding to the target course information into a trained teaching material expander, so that the teaching material expander outputs the second set of alternative characters corresponding to the position of the second lesson plan and the probability of each alternative character in the second set of alternative characters.

[0058] In some embodiments, the teaching material content is natural language text, and the large language model is a deep learning model trained using a large amount of text data, which can understand the meaning of language text and generate natural language text; therefore, the large language model can be used as a teaching material expander to expand the course content corresponding to the target course information to obtain a second lesson plan that is expanded based on the above-mentioned teaching material content. When generating the lesson plan, the teaching material expander will calculate and output at least two alternative characters corresponding to each position on the lesson plan and the probability of the alternative characters, that is, output the second alternative character set corresponding to the position of the second lesson plan and the probability of each alternative character in the second alternative character set.

[0059] In some embodiments, the general large language model can be trained for domain adaptation through supervised fine-tuning based on a self-built professional lesson plan dataset, so that the general large language model can internalize educational principles such as constructivism and cognitive load theory to obtain a teaching material expander. Through parameter adjustment, the lesson plan output by the teaching material expander conforms to the sentence structure and knowledge organization of the educational scenario.

[0060] S103 : Determine target probabilities for each candidate character in the first candidate character set and target probabilities for each candidate character in the second candidate character set according to the probabilities of each candidate character in the first candidate character set and the probabilities of each candidate character in the second candidate character set.

[0061] In order to avoid the limitations of a single information source, the present invention comprehensively considers the probabilities of the first candidate characters and the second candidate characters when determining the target probabilities of the candidate characters in the first candidate characters set and the target probabilities of the candidate characters in the second candidate characters set for each position of the teaching plan template.

[0062] As a possible implementation method, the target probability of each alternative character in the first alternative character set and the target probability of each alternative character in the second alternative character set are determined according to the probability of each alternative character in the first alternative character set and the probability of each alternative character in the second alternative character set, including: determining a first weight corresponding to the probability of the alternative character in the first alternative character set; and a second weight corresponding to the probability of the alternative character in the second alternative character set; performing a weighted operation based on the probability of the alternative character in the first alternative character set, the probability of the alternative character in the second alternative character set, the first weight and the second weight, and determining the target probability of the alternative character according to the operation result.

[0063] In some embodiments, the first weight and the second weight may be hyperparameters, that is, the first weight and the second weight may be pre-set.

[0064] In some embodiments, the alternative character belongs to the first alternative character set but not the second alternative character set, or the alternative character belongs to the second alternative character set but not the first alternative character set. For these two situations, they can be handled as follows: if the alternative character does not belong to the first alternative character set, the probability of the alternative character in the first alternative character set can be determined as 0; if the alternative character does not belong to the second alternative character set, the probability of the alternative character in the second alternative character set can be determined as 0.

[0065] That is, when the alternative character belongs to the second alternative character set but not the first alternative character set, the target probability of the alternative character is determined based on the probability of the alternative character in the second alternative character set and the corresponding second weight; when the alternative character belongs to the first alternative character set but not the second alternative character set, the target probability of the alternative character is determined based on the probability of the alternative character in the first alternative character set and the corresponding first weight.

[0066] S104: Determine a target character at the position in the set consisting of the first candidate character set and the second candidate character set according to the target probability of each candidate character in the first candidate character set and the target probability of each candidate character in the second candidate character set.

[0067] The present invention targets each position of the teaching plan template, and the target character at the position is determined from the set consisting of the first candidate character set and the second candidate character set, thereby expanding the search range of the target character.

[0068] The target probability of each candidate character in the first candidate character set and the target probability of each candidate character in the second candidate character set comprehensively consider the probability of each candidate character in the first candidate character set and the probability of each candidate character in the second candidate character set, avoiding the limitations of a single information source.

[0069] In this way, by determining the target character in the above manner, the comprehensiveness and reliability of the target character selection are improved.

[0070] As a possible implementation method, S104 determines the target character at the position in the collection consisting of the first alternative character set and the second alternative character set based on the target probability of each alternative character in the first alternative character set and the target probability of each alternative character in the second alternative character set, including: determining the alternative character with the highest target probability in the collection consisting of the first alternative character set and the second alternative character set as the target character at the position.

[0071] In the present invention, the candidate character with the highest target probability at this position is the candidate character that is most likely to be at this position after comprehensively considering the probabilities of each candidate character in the first candidate character set and the probabilities of each candidate character in the second candidate character set. Therefore, this candidate character is determined as the target character at this position.

[0072] S105. Generate a target lesson plan based on the target characters at various positions of the lesson plan template.

[0073] The present invention comprehensively considers the probability of each alternative character in the first alternative character set corresponding to the position and the probability of each alternative character in the second alternative character set corresponding to the position for each position of the teaching plan template, obtains the target probability of each alternative character in the collection consisting of the first alternative character set and the second alternative character set, and then determines the target character at the position of the teaching plan template in the collection according to the target probability of each alternative character in the collection.

[0074] The target lesson plan generated based on the target characters at each position of the lesson plan template makes full use of the first candidate character set, the probability of each candidate character in the first candidate character set, the second candidate character set and the probability of each candidate character in the second candidate character set, thereby realizing the effective integration of two different lesson plan generation methods corresponding to the above information, avoiding the limitations of a single information source, and improving the comprehensiveness and integrated nature of lesson plan generation.

[0075] In some embodiments, the target characters may be set at corresponding positions to obtain a reference teaching plan, which may be used as a target teaching plan.

[0076] In some embodiments, the reference lesson plan can be further optimized to obtain a target lesson plan.

[0077] As a possible implementation method, S105 generates a target lesson plan according to target characters at various positions of the lesson plan template, including: inputting the target characters at various positions into a trained lesson plan generator, so that the lesson plan generator generates the target lesson plan.

[0078] In some embodiments, the lesson plan generator may set the target character at the position corresponding to the target character to obtain a reference lesson plan; and then generate a target lesson plan based on the reference lesson plan. The reference lesson plan composed of the target characters may directly represent the fusion result of the first candidate character set, the probability of each candidate character in the first candidate character set, the second candidate character set, and the probability of each candidate character in the second candidate character set. However, the format of the reference lesson plan (e.g., sentence structure, knowledge organization, etc.) may not fully match the format required by the lesson plan. After obtaining the reference lesson plan, the target lesson plan may be further generated based on the reference lesson plan, so that the target lesson plan meets the format requirements of the lesson plan while integrating the above information.

[0079] In some embodiments, the target characters and target lesson plans at various positions are natural language texts, and the large language model can be used to understand and generate natural language texts, so the large language model can be used as a lesson plan compressor.

[0080] In some embodiments, based on a self-built professional lesson plan dataset, a general large language model can be trained for domain adaptation through supervised fine-tuning, so that the general large language model can internalize educational principles such as constructivism and cognitive load theory to obtain a lesson plan generator; in this way, through parameter adjustment, the target lesson plan output by the lesson plan generator conforms to the sentence structure and knowledge organization of the educational scenario.

[0081] The present invention relates to natural language processing and artificial intelligence education technology, and is mainly used in online education platforms, intelligent lesson preparation systems, teacher training tools, and other scenarios. Specifically, it includes:

[0082] 1. Intelligent lesson plan generation system: By integrating educational theory frameworks with textbook knowledge bases, it provides precise teaching plans for K12 (kindergarten through twelfth grade) education, vocational education, and other scenarios.

[0083] 2. Adaptive optimization of educational content: Dynamically adjust the lesson plan structure based on the subject knowledge point map, suitable for personalized learning systems.

[0084] 3. Teaching and research decision support: Provide educational institutions with teaching plan quality assessment tools based on data analysis.

[0085] Existing teaching plan generation technology based on large language models has the following core flaws:

[0086] 1. Lack of educational theory: Mainstream large-scale models (such as Tongyi Qianwen and Wenxin Yiyan) rely on general corpus pre-training and fail to systematically integrate educational frameworks such as curriculum standards and teaching syllabuses. As a result, the generated lesson plans lack educational theoretical support.

[0087] 2. Weak relevance to textbooks: Existing methods (such as Self-Critique) are not deeply related to specific chapters in textbooks, and the generated content often has misplaced or redundant knowledge points.

[0088] 3. Lack of practicality: The generated lesson plan has a complete structure but vague content, which requires teachers to spend a lot of time on secondary revisions.

[0089] As an example, a teaching plan generation method provided by the present invention is described by using a teaching plan automatic generation method based on a "compression-expansion" collaborative framework. This method solves the core pain points in the field of intelligent education through three technological breakthroughs:

[0090] 1. Deep Penetration of Educational Theory: Using supervised fine-tuning techniques, a general-purpose large model is trained on a teaching plan dataset for domain adaptation, guiding the internal parameters of the large language model to systematically incorporate core educational principles such as constructivism and cognitive load theory. A retrieval-enhanced generation architecture dynamically acquires real-world expert teaching plan examples, and a teaching plan compressor is constructed to condense retrieval results, transforming abstract educational theories into executable policy parameters.

[0091] 2. Dynamic Textbook Knowledge Expansion System: This system breaks through the limitations of traditional knowledge compression through the textbook expander module. It analyzes textbook knowledge nodes through prompt word engineering and, combined with the parameterized knowledge of large language models, generates expanded lesson plans that conform to cognitive principles.

[0092] 3. Double guarantee system for generation quality: Design a collaborative working mechanism for the lesson plan compressor and the teaching material expander, adopt integrated decoding technology, and dynamically integrate the dual-channel output weights of the compressor and expander, so that the generation plan has both the rigor of expert experience and the systematic knowledge of the teaching material.

[0093] In summary, this framework uses a three-step improvement mechanism of "refining existing experience - activating textbook knowledge - dynamic balanced output" to ensure that the generated lesson plans retain the teaching wisdom of human teachers and have the integrity of the knowledge system.

[0094] The core of this example is to propose a "compression-expansion" framework for lesson plan generation, which is achieved through the following innovative technologies:

[0095] (1) Supervised fine-tuning technology

[0096] Based on a self-built dataset of professional teaching plans, we used supervised fine-tuning techniques to adapt the general large model to its specific domain. This allowed the model to internalize educational principles such as constructivism and cognitive load theory. Parameter adjustments were then made to ensure that the model's output, after domain adaptation training, matched the sentence structure and knowledge organization of the educational context.

[0097] (2) Dynamic retrieval enhancement technology

[0098] Through the RAG architecture, we can obtain expert teaching case examples in real time and form a teaching strategy reference pool to solve the problem that static models cannot adapt to the latest teaching cases.

[0099] (3) “Compression-Expansion” Framework for Lesson Plan Generation

[0100] Figure 2 This is a schematic diagram of a "compression-expansion" framework for lesson plan generation provided by the present invention. Figure 2 As shown, the lesson plan compressor is a large language model that uses prompt word engineering to reconstruct the content of at least two lesson plan samples that match the target course information to obtain a first lesson plan, wherein the lesson plan sample is a lesson plan designed and written by an education expert.

[0101] The lesson plan compressor is based on the target course information x, at least two lesson plan samples D that match the target course information, and the lesson plan compression prompt word I comp Generate the first lesson plan Probability As shown in the following formula.

[0102]

[0103] in, For the first lesson plan The character at position i in The character before position i in the first lesson plan.

[0104] In the process of generating text, the large language model generates characters one by one based on the generated characters, that is, the lesson plan compressor generates The probability of You can as well as The probability of generating Determine the candidate character in the first candidate character set corresponding to position i and the probability of the candidate character.

[0105] like Figure 2 As shown, the textbook compressor is a large language model. Combined with the parameterized knowledge of the large language model, it parses the textbook knowledge nodes in the textbook content and generates a second lesson plan that conforms to cognitive laws.

[0106] The textbook expander is based on the target course information x, the textbook content e corresponding to the target course information, and the textbook expansion prompt word i tb The second lesson plan for success Probability As shown in the following formula.

[0107]

[0108] in, For the second lesson plan The character at position i.

[0109] in, For the second lesson plan The character at position i in It is the character before position i in the second lesson plan.

[0110] Similar to the first lesson plan information, the teaching material expander generates The probability of Can as well as The probability of generating Determine the candidate character in the second candidate character set corresponding to position i and the probability of the candidate character.

[0111] like Figure 2 As shown, integrated decoding technology can be used to dynamically integrate the dual-channel output weights of the compressor and expander to generate a reference teaching plan that retains the teaching wisdom of human teachers and has the integrity of the knowledge system.

[0112] Specifically, the candidate characters in the first candidate character set corresponding to position i are The probability of the candidate character in the second candidate character set corresponding to position i The weighted sum of the probabilities is determined as the candidate character at position i The target probability of the candidate character with the largest target probability is determined as the target character d at position i i , as shown below.

[0113]

[0114] Among them, α is the overall coefficient of the trade-off between the two probability distributions of the lesson plan compressor and the teaching material expander, V is the set of the first candidate character set corresponding to position i and the second candidate character set corresponding to position i, For the alternative characters in this collection.

[0115] The target character d i Set it at the corresponding position i and get the reference lesson plan d.

[0116] After obtaining the reference lesson plan d, the target course information x and the reference lesson plan d are input into the lesson plan generator G. The lesson plan generator G is a large language model. Relying on the contextual capabilities of the large language model, it generates a target lesson plan y that has both human teacher style and is related to the textbook knowledge based on the reference lesson plan d generated by integrated decoding and the target course information x, as shown in the following formula.

[0117] y~G(x,d)

[0118] The above-mentioned lesson plan compressor, teaching material expander and lesson plan generator can be any large model, including but not limited to: ChatGPT, LLaMA, Wenxin Yiyan, Tongyi Qianwen and other mainstream large models.

[0119] Corresponding to the aforementioned method embodiments, this specification also provides embodiments of an apparatus and a computing device to which the apparatus is applied.

[0120] See also Figure 3 , Figure 3 This is a block diagram of a teaching plan generating device provided by the present invention. Figure 3 As shown, the device includes:

[0121] The first acquisition module 301 is used to acquire, for each position of the teaching plan template, a first candidate character set corresponding to the position and the probability of each candidate character in the first candidate character set.

[0122] The second acquisition module 302 is configured to acquire a second candidate character set corresponding to the position and the probability of each candidate character in the second candidate character set.

[0123] The target probability determination module 303 is used to determine the target probability of each alternative character in the first alternative character set and the target probability of each alternative character in the second alternative character set based on the probability of each alternative character in the first alternative character set and the probability of each alternative character in the second alternative character set.

[0124] The target character determination module 304 is used to determine the target character of the position in the set consisting of the first candidate character set and the second candidate character set based on the target probability of each candidate character in the first candidate character set and the target probability of each candidate character in the second candidate character set.

[0125] The target lesson plan generating module 305 is used to generate a target lesson plan according to the target characters at various positions of the lesson plan template.

[0126] As a possible implementation method, the first acquisition module 301 is specifically used to compress at least two lesson plan samples that match the target course information to generate a first set of alternative characters corresponding to the position of the first lesson plan and the probability of each alternative character in the first set of alternative characters; the first lesson plan is a lesson plan obtained by compressing at least two lesson plan samples that match the target course information; the position of the first lesson plan corresponds one-to-one to the position of the lesson plan template.

[0127] As a possible implementation method, the first acquisition module 301 is specifically used to input the lesson plan compression prompt words, target course information, and at least two lesson plan samples that match the target course information into a trained lesson plan compressor, so that the lesson plan compressor outputs the first alternative character set corresponding to the position of the first lesson plan and the probability of each alternative character in the first alternative character set.

[0128] As a possible implementation method, the second acquisition module 302 is specifically used to expand the teaching material content corresponding to the target course information to generate a second set of alternative characters corresponding to the position of the second lesson plan and the probability of each alternative character in the second set of alternative characters; the second lesson plan is a lesson plan obtained by expanding the teaching material content corresponding to the target course information; the position of the second lesson plan corresponds one-to-one to the position of the lesson plan template.

[0129] As a possible implementation method, the second acquisition module 302 is specifically used to input the textbook extension prompt words, target course information and the textbook content corresponding to the target course information into the trained textbook expander, so that the textbook expander outputs the second alternative character set corresponding to the position of the second lesson plan and the probability of each alternative character in the second alternative character set.

[0130] As a possible implementation method, the target probability determination module 303 is specifically used to determine the first weight corresponding to the probability of the alternative character in the first alternative character set; and the second weight corresponding to the probability of the alternative character in the second alternative character set; perform a weighted operation based on the probability of the alternative character in the first alternative character set, the probability of the alternative character in the second alternative character set, the first weight and the second weight, and determine the target probability of the alternative character according to the operation result.

[0131] As a possible implementation manner, the target character determination module 304 is specifically configured to determine the candidate character with the highest target probability in the set of the first candidate character set and the second candidate character set as the target character at the position.

[0132] As a possible implementation method, the target lesson plan generation module 305 is specifically used to input the target characters at various positions into a trained lesson plan generator, so that the lesson plan generator generates a target lesson plan.

[0133] The implementation process of the functions and effects of each module in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0134] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the partial description of the method embodiments. The device embodiments described above are only schematic, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the scheme of this specification. Those of ordinary skill in the art can understand and implement it without paying any creative work.

[0135] This specification also provides a computing device, see Figure 4 , Figure 4 This is a schematic diagram of the structure of a computing device provided by the present invention. Figure 4 As shown, the computing device includes a processor 410, a memory 420, and a network interface 430. The memory 420 is used to store computer instructions that can be executed on the processor 410. The processor 410 is used to implement a teaching plan generation method provided in any embodiment of this specification when executing the computer instructions. The network interface 430 is used to implement input and output functions. In more possible implementations, the computing device may also include other hardware, which is not limited in this specification.

[0136] This specification also provides a kind of computer-readable storage medium, and computer-readable storage medium can be various forms, such as, in different examples, computer-readable storage medium can be: RAM (Radom Access Memory, random access memory), volatile memory, non-volatile memory, flash memory, storage drive (such as hard disk drive), solid state drive, any type of storage disk (such as optical disk, DVD etc.), or similar storage medium, or their combination.Special, computer-readable medium can also be paper or other suitable medium that can print program.Computer-readable storage medium is stored with computer program, and computer program is implemented when being executed by processor and provides a kind of teaching plan generation method that any embodiment of this specification provides.

[0137] This specification also provides a computer program product, including a computer program, which, when executed by a processor, implements a teaching plan generation method provided in any embodiment of this specification.

[0138] It should be understood by those skilled in the art that one or more embodiments of this specification may be provided as a method, apparatus, computing device, computer-readable storage medium, or computer program product. Thus, one or more embodiments of this specification may take the form of a fully hardware embodiment, a fully software embodiment, or an embodiment combining software and hardware aspects. Furthermore, one or more embodiments of this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0139] The various embodiments in this specification are described in a progressive manner. Similar portions between the various embodiments can be referenced to each other. Each embodiment focuses on the differences between the other embodiments. In particular, the embodiments corresponding to the computing device are generally similar to the method embodiments, so the description is relatively simple. For relevant portions, refer to the description of the method embodiments.

[0140] The foregoing description is of specific embodiments of this specification. Other embodiments are within the scope of this specification. In some cases, the actions or steps described in this specification can be performed in an order different from that shown in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0141] The embodiments of the subject matter and functional operations described in this specification may be implemented in the following: digital electronic circuits, tangibly embodied computer software or firmware, computer hardware including the structures disclosed in this specification and their structural equivalents, or a combination of one or more of them. The embodiments of the subject matter described in this specification may be implemented as one or more computer programs, i.e., one or more modules of computer program instructions encoded on a tangible, non-transitory program carrier to be executed by a data processing device or to control the operation of a data processing device. Alternatively or additionally, the program instructions may be encoded on an artificially generated propagation signal, such as a machine-generated electrical, optical or electromagnetic signal, which is generated to encode information and transmit it to a suitable receiver device for execution by a teaching plan generation device. The computer storage medium may be a machine-readable storage device, a machine-readable storage substrate, a random or serial access memory device, or a combination of one or more of them.

[0142] The processes and logic flows described in this specification can be performed by one or more programmable computers executing one or more computer programs to perform the corresponding functions by operating on input data and generating output. The processes and logic flows can also be performed by, and apparatus can be implemented as, special purpose logic circuitry, such as an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit).

[0143] Computers suitable for executing computer programs include, for example, general-purpose and / or special-purpose microprocessors, or any other type of central processing unit. Typically, the central processing unit will receive instructions and data from a read-only memory and / or random access memory. The basic components of a computer include a central processing unit for implementing or executing instructions and one or more memory devices for storing instructions and data. Typically, a computer will also include one or more mass storage devices for storing data, such as magnetic disks, magneto-optical disks, or optical disks, or the computer will be operably coupled to such mass storage devices to receive data from them or to transmit data to them, or both. However, a computer does not necessarily have such devices. In addition, a computer can be embedded in another device, such as a mobile phone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a global positioning system (GPS) receiver, or a portable storage device such as a universal serial bus (USB) flash drive, to name a few.

[0144] Computer-readable media suitable for storing computer program instructions and data include all forms of non-volatile memory, media, and storage devices, including, for example, semiconductor memory devices (e.g., EPROM, EEPROM, and flash memory devices), magnetic disks (e.g., internal hard disks or removable disks), magneto-optical disks, and CD ROM and DVD-ROM disks. The processor and memory can be supplemented by, or incorporated in, special purpose logic circuitry.

[0145] Although this specification includes many specific implementation details, these should not be interpreted as limiting the scope of any invention or the scope of protection claimed, but are mainly used to describe the features of specific embodiments of specific inventions. Certain features described in multiple embodiments within this specification may also be implemented in combination in a single embodiment. On the other hand, the various features described in a single embodiment may also be implemented separately in multiple embodiments or in any suitable sub-combination. In addition, although features may work in certain combinations as described above and even initially claimed as such, one or more features from the claimed combination may be removed from the combination in some cases, and the claimed combination may point to a sub-combination or a variation of the sub-combination.

[0146] Similarly, although operations are depicted in a particular order in the accompanying drawings, this should not be understood as requiring that these operations be performed in the particular order shown or performed sequentially, or that all illustrated operations be performed to achieve the desired results. In some cases, multitasking and parallel processing may be advantageous. In addition, the separation of various system modules and components in the above-described embodiments should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product, or packaged into multiple software products.

[0147] Thus, specific embodiments of the subject matter have been described. Other embodiments are within the scope of this specification. In some cases, the actions described in this specification can be performed in a different order and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In some implementations, multitasking and parallel processing may be advantageous.

[0148] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the present invention and practicing the invention claimed herein. This specification is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of this specification and include common knowledge or customary techniques in the art that are not claimed herein. That is, this specification is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and variations may be made without departing from the scope thereof.

[0149] The above description is only an optional embodiment of this specification and is not intended to limit this specification. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of this specification should be included in the scope of protection of this specification.

Claims

1. A teaching plan generation method, characterized in that: The method comprises: For each position of the teaching plan template, obtaining a first candidate character set corresponding to the position and the probability of each candidate character in the first candidate character set; Obtaining a second candidate character set corresponding to the position and the probability of each candidate character in the second candidate character set; determining, based on the probability of each candidate character in the first candidate character set and the probability of each candidate character in the second candidate character set, a target probability for each candidate character in the first candidate character set and a target probability for each candidate character in the second candidate character set; determining a target character for the position in a collection of the first candidate character set and the second candidate character set based on target probabilities of the candidate characters in the first candidate character set and target probabilities of the candidate characters in the second candidate character set; Generate target lesson plans based on the target characters at various positions of the lesson plan template.

2. The method according to claim 1, characterized in that The obtaining of a first candidate character set corresponding to the position and the probability of each candidate character in the first candidate character set includes: At least two lesson plan samples that match the target course information are compressed to generate a first set of candidate characters corresponding to the position of the first lesson plan and the probability of each candidate character in the first set of candidate characters; the first lesson plan is a lesson plan obtained by compressing at least two lesson plan samples that match the target course information; the position of the first lesson plan corresponds one-to-one to the position of the lesson plan template.

3. The method according to claim 2, characterized in that The compressing of at least two teaching plan samples matching the target course information to generate a first candidate character set corresponding to the position of the first teaching plan and the probability of each candidate character in the first candidate character set includes: The lesson plan compression prompt words, target course information and at least two lesson plan samples that match the target course information are input into a trained lesson plan compressor, so that the lesson plan compressor outputs a first set of alternative characters corresponding to the position of the first lesson plan and the probability of each alternative character in the first set of alternative characters.

4. The method according to claim 1, wherein Obtaining a second candidate character set corresponding to the position and the probability of each candidate character in the second candidate character set includes: The teaching material content corresponding to the target course information is expanded to generate a second set of alternative characters corresponding to the position of the second lesson plan and the probability of each alternative character in the second set of alternative characters; the second lesson plan is a lesson plan obtained by expanding the teaching material content corresponding to the target course information; the position of the second lesson plan corresponds one-to-one to the position of the lesson plan template.

5. The method according to claim 4, characterized in that The teaching material content corresponding to the target course information is extended to generate a second candidate character set corresponding to the position of the second teaching plan and the probability of each candidate character in the second candidate character set, including: The teaching material extension prompt word, target course information and teaching material content corresponding to the target course information are input into the trained teaching material expander, so that the teaching material expander outputs the second alternative character set corresponding to the position of the second lesson plan and the probability of each alternative character in the second alternative character set.

6. The method according to claim 1, characterized in that Determining, based on the probability of each candidate character in the first candidate character set and the probability of each candidate character in the second candidate character set, a target probability of each candidate character in the first candidate character set and a target probability of each candidate character in the second candidate character set, includes: Determining a first weight corresponding to the probability of the candidate character in the first set of candidate characters; and a second weight corresponding to the probability of the candidate character in the second set of candidate characters; A weighted operation is performed based on the probability of the candidate character in the first candidate character set, the probability of the candidate character in the second candidate character set, the first weight, and the second weight, and a target probability of the candidate character is determined according to the operation result.

7. The method according to claim 1, characterized in that Determining a target character at the position in a collection of the first candidate character set and the second candidate character set based on target probabilities of the candidate characters in the first candidate character set and target probabilities of the candidate characters in the second candidate character set includes: The candidate character with the highest target probability in the set consisting of the first candidate character set and the second candidate character set is determined as the target character at the position.

8. The method according to claim 1, characterized in that Generating a target lesson plan according to target characters at various positions of the lesson plan template includes: The target characters at each position are input into a trained lesson plan generator, so that the lesson plan generator generates a target lesson plan.

9. A teaching plan generating device, characterized in that: The device comprises: A first acquisition module is configured to acquire, for each position in the teaching plan template, a first candidate character set corresponding to the position and a probability of each candidate character in the first candidate character set; A second acquisition module is configured to acquire a second candidate character set corresponding to the position and a probability of each candidate character in the second candidate character set; a target probability determination module, configured to determine a target probability for each candidate character in the first candidate character set and a target probability for each candidate character in the second candidate character set based on the probability of each candidate character in the first candidate character set and the probability of each candidate character in the second candidate character set; a target character determination module, configured to determine a target character for the position in a collection of the first candidate character set and the second candidate character set based on the target probabilities of the candidate characters in the first candidate character set and the target probabilities of the candidate characters in the second candidate character set; The target lesson plan generation module is used to generate a target lesson plan based on the target characters at various positions of the lesson plan template.

10. A computing device, characterized in that The computing device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the operations performed by a teaching plan generation method as described in any one of claims 1 to 8 when executing the computer program.

11. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program, and the program is used by a processor to execute the operations performed by the teaching plan generation method as described in any one of claims 1 to 8.

Images