Intelligent evaluating and teaching system for phrase making of English phrases

Abstract

The invention discloses an intelligent evaluation and teaching system for English phrase phrasing. In English teaching, phrase phrase making difficulty is large, and teachers are difficult to carry out personalized evaluation and timely feedback on phrase phrase making of students in school English teaching practices, so that the invention designs an intelligent evaluation and teaching method and system for English phrase phrase making. The system scores sentences made by learners according to phrases through specific cue words by a plurality of large language models to obtain three-dimensional scores of phrase use accuracy, sentence structure integrity and expression naturalness and personalized feedback; comprehensively considering the score and the feedback of each dimension to obtain the score and the total score of each dimension; designing a mastering degree model and a stability criterion, calculating and updating a word group mastering state of the student, pushing a new word group according to the mastering state, realizing a teaching closed loop of practice, evaluation and tutoring, and supporting a teacher to know mastering conditions of all the students and the word group; and personalized word group teaching in a large-scale class teaching environment can be realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent tutoring teaching of English phrase learning, and particularly relates to an English phrase sentence making intelligent evaluation and teaching support method and system based on large language model semantic analysis and learner state modeling. BACKGROUND

[0002] In the process of language learning, phrases are important units that connect vocabulary knowledge and sentence expression ability, and are the key link for learners to form language output ability. Through phrase sentence making exercises, learners can develop from formal memory to situational application.

[0003] In existing phrase teaching practice, phrase sentence making training mainly relies on manual correction by teachers, which has a long feedback cycle and a large workload, and is difficult to support large-scale and personalized evaluation and teaching in a class teaching environment. At the same time, existing automatic correction systems are mostly based on rule matching or keyword detection, and can only judge whether the surface structure of a sentence is correct, lack comprehensive understanding of semantic rationality and expression naturalness, and lack automatic evaluation and tutoring of learners' phrase sentence making ability.

[0004] The existing technology has the following disadvantages: the scoring result is greatly affected by the fluctuation of a single model; the scoring is not continuously associated with the learning history; the exercise task allocation is not dynamically adjusted based on the learning state; and the learning process lacks an automatic closed-loop mechanism.

[0005] Therefore, an intelligent evaluation and tutoring system capable of multi-model fusion semantic evaluation, learner state modeling, and adaptive task allocation is needed to support large-scale and personalized evaluation and teaching of English phrase sentence making. SUMMARY

[0006] To this end, we design an English phrase sentence making intelligent evaluation and teaching method and system, and a storage medium. The method can be applied in the design of an intelligent teaching system for the English subject, replacing the original one-on-one and personalized tutoring of students by teachers, greatly reducing the tutoring workload of teachers and the learning burden of students, improving teaching efficiency, and helping students to learn independently.

[0007] To solve the above technical problems, one of the purposes of the present application is to provide an English phrase sentence making intelligent evaluation and teaching support method, which comprises the following steps.

[0008] S1, a phrase resource library, a learning record database, and a model scoring interface module are established.

[0009] S2, according to the mastery state of the learner in the learning record database, an exercise phrase is selected from the phrase resource library.

[0010] S3, the phrase is presented to the learner and the learner's sentence making input is received.

[0011] S4. Send the sentence-making text to two large language models so that they can perform semantic analysis according to specific prompts.

[0012] S5. Each model outputs scores for three capability dimensions, as well as improvement feedback and prompts. The three dimensions include accuracy of phrase usage, completeness of sentence structure, and naturalness of language expression.

[0013] S6. Take into account the outputs of different models, give scores for each dimension and the total score, as well as detailed prompts and feedback.

[0014] S7. Update the mastery value of the corresponding phrase based on the phrase mastery model.

[0015] S8. Adjust the probability of selecting phrases in subsequent exercises based on the updated mastery level.

[0016] S9. Save the learner's status, including the rating results and feedback information, to the database and present it to the learner.

[0017] S10. Teachers aggregate and analyze all learners' status data to generate a status table and a visual dashboard showing their mastery of phrase construction.

[0018] The second objective of this invention is to provide an intelligent evaluation and teaching support system for English phrase sentence construction, comprising: a phrase task acquisition module, an input processing module, a multi-model scoring module, an evaluation and prompt result fusion module, a learner status module, an adaptive task allocation module, and a teaching data analysis module, wherein each module works together to implement the above-mentioned method steps.

[0019] A third objective of this invention is to provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method steps.

[0020] As can be seen from the above technical solutions, compared with the existing technologies, the present invention improves the stability of scoring through a multi-model fusion mechanism; achieves continuous modeling of learning states through a mastery model; improves the relevance of practice through an adaptive scheduling mechanism; and forms a technical closed loop of "practice, assessment, and tutoring," thereby improving the automation and personalization of language assessment and teaching. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall process structure provided by the present invention. Detailed Implementation

[0022] The following combination Figure 1 The specific implementation of the present invention will be described below.

[0023] like Figure 1 As shown, the method of the present invention includes...

[0024] First, initialize the learners' mastery of phrases in the unit.

[0025] Based on the learner's mastery level, select the phrase with the lowest mastery level in this unit from the phrase resource library and present it to the learner.

[0026] Students create sentences and submit them to multiple large models for evaluation.

[0027] Based on feedback from various language models, scores are calculated across three dimensions and the total score to determine mastery level; assessments and prompts are then presented to students.

[0028] If the mastery level does not meet the requirements, continue practicing this phrase; if the mastery level meets the requirements, determine whether the learner has mastered all the phrases in this unit; if all the phrases in this unit have been mastered, prompt the learner to complete the learning of this unit and move on to the next unit; otherwise, continue practicing the phrases in this unit; if all units have been mastered, the process ends.

[0029] In the above process, score generation, mastery update, and question selection are implemented based on a multi-model semantic comprehensive evaluation mechanism and a mastery dynamic modeling algorithm, thus forming a closed-loop system of learning practice, evaluation, and tutoring tips.

[0030] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent English phrase sentence construction evaluation and teaching support method, characterized by the following steps: S1, establishing a phrase resource library, a learner state database and a large model scoring interface module; S2, selecting phrases that need to be practiced from the phrase resource library according to the learner's mastery state; S3, presenting the selected phrases to the learner, receiving their keyboard input based on phrase sentence construction, and prohibiting copy and paste; S4, sending the sentence text to two large language models to perform semantic analysis according to specific prompt sentences; S5, each model outputs three ability dimension scores, improvement feedback and prompt information, and the three dimensions include phrase usage accuracy, sentence structure integrity and language expression naturalness; S6, considering the outputs of different models comprehensively, giving each dimension score and total score, and detailed prompt feedback; S7, updating the mastery value of the corresponding phrase based on the phrase mastery model; S8, according to the updated mastery value, preferentially push the phrases that do not reach the mastery threshold to the student for practice; S9, saving the learner state including scoring results and feedback information to the database and presenting it to the learner; S10, the teacher aggregates and analyzes all learner state data to generate a phrase sentence mastery state table and a visual dashboard.

2. The method of claim 1, wherein: The phrase resource library includes phrase identification, phrase text, Chinese explanation and semantic label.

3. The method of claim 1, wherein: The learner state database includes learner identification, phrase identification, dimension score, overall mastery and time stamp.

4. The method of claim 1, wherein: The comprehensive consideration is based on the output results of each large model, and the average value of the scores of each model is taken as the score; the detailed evaluation feedback comes from the similarities and differences of the output results of each large model.

5. The method of claim 1, wherein: The mastery model is updated incrementally or decrementally according to multiple scoring results, and combined with a stability criterion to divide the mastery state.

6. The method of claim 1, wherein: The selection probability is realized by giving higher extraction weight to unmastered phrases.

7. The method of claim 1, wherein: The aggregate analysis includes learner ability dimension and mastery statistics, phrase mastery state statistics and unit progress trend analysis.

8. An intelligent English phrase sentence construction evaluation and teaching support system, comprising: Phrase task acquisition module, input processing module, multi-model scoring module, evaluation and prompt result fusion module, learner state module, adaptive task allocation module and teaching data analysis module, which cooperatively realize any method of claims 1-7.

9. A computer readable storage medium having a computer program stored thereon, wherein the program is executed by a processor to realize any method of claims 1-7.

Images