Lao grammar error correction corpus construction method and system based on error distribution guidance large model, and electronic equipment

Abstract

The invention relates to a Lao grammar error correction corpus construction method and system based on an error distribution guidance large model, and electronic equipment. According to the invention, the existing speech recognition model is utilized to simulate common grammar errors which appear actually; then, a large language model is utilized to automatically generate data covering various error distributions according to rules and constraints; then, the data are cleaned and preprocessed; then, a small language model is used as an error correction model to carry out error correction on the fused corpus; and finally, screening out Lao grammar error correction corpora with a lower word error rate through model evaluation, thereby effectively solving the problem of lack of Lao grammar error correction corpora. Lao grammar error correction corpora are generated by effectively utilizing error distribution of Lao grammar to guide the large model, and a better experimental result is obtained on Lao grammar error correction tasks.

Description

[0001] Technical Language

[0002] The invention relates to a method, system and electronic device for constructing Laotian grammar error correction corpus based on an error distribution guidance large model, and belongs to the technical field of natural language processing. Background Art

[0003] With the deepening of globalization, the learning and use of Lao language has become increasingly important. However, learners and users of Lao language often face the problem of grammatical errors, which not only affects the fluency of communication, but also hinders the improvement of learners' language ability. Common grammatical errors include inconsistent verb tenses, inconsistent subject and predicate, improper use of prepositions, etc., which are common in the process of learning and using. At present, traditional methods for correcting grammatical errors mainly rely on rule-based grammar checking tools. These tools detect errors through preset grammatical rules. Although they can solve the problem to a certain extent, they are difficult to adapt to the complexity of language use. In addition, with the development of artificial intelligence technology, grammar correction methods based on machine learning have gradually emerged. However, when dealing with Lao language, existing models still have problems of lack of data and insufficient training, resulting in unsatisfactory accuracy and efficiency of grammar correction.

[0004] Speech recognition technology has also developed rapidly in recent years. By converting speech into text, a large-scale corpus can be effectively generated. However, the text generated based on speech recognition usually contains multiple types of errors and is not manually annotated. These unlabeled data are very challenging in the training process, but they also provide a rich source of information, providing a possible improvement path for subsequent error correction models. Large models such as MT5 and MBART have performed well in multiple natural language processing tasks. These models can be trained on large-scale corpora and demonstrate strong capabilities in tasks such as grammatical correction and text generation. However, for the specific needs of Lao, existing research has not yet explored in depth how to effectively combine the unlabeled data generated by speech recognition with the data generated by large models to construct high-quality grammatical correction corpora. In view of this, the construction of Lao grammatical correction corpora based on error distribution guidance is particularly necessary. By conducting in-depth analysis of the generated corpus, identifying common error types and integrating them, it is possible to improve not only the quality of the corpus, but also the training effect of the model. Such research not only helps to promote the education and popularization of Lao, but also provides new ideas and methods for the study of grammatical correction in other minority languages. Therefore, exploring how to effectively combine speech recognition models and corpora generated by large models will be of great significance to improving the grammatical correction effect of Lao. Summary of the invention

[0005] The present invention provides a method, system and electronic device for constructing Lao grammar correction corpus based on an error distribution guidance large model to solve the problem of insufficient Lao grammar correction corpus. The present invention obtains relatively high-quality Lao grammar correction corpus by fusing data obtained by speech recognition with data generated by a large model.

[0006] The technical solution of the present invention is: a method for constructing Laotian grammar correction corpus based on an error distribution guidance large model, the method comprising:

[0007] Step 1: Collection of unlabeled data. First, use the existing Lao speech recognition model to automatically generate large-scale Lao corpus; then, clean, filter and preprocess the corpus; finally, generate Lao unlabeled data with different error distributions;

[0008] Step 2: First, select relatively standard texts from the basic corpus as the original correct sentences; then, based on certain rules and constraints, use the large model to automatically generate sentences with various error distributions to construct pseudo data; finally, clean and deduplicate the generated pseudo data to form the initial grammatical error correction pseudo data;

[0009] Step 3: Further improve the performance of Lao grammar correction by fusing all error distribution data. First, analyze the error distribution of Lao unlabeled data and grammar correction pseudo data to identify common error types; fuse unlabeled data and pseudo data containing different error types to form a corpus containing different error distributions;

[0010] Step 4: Use the fused corpus to train the language model and build a Lao grammar correction model for Lao grammar correction; evaluate the correction prediction results to obtain the final Lao grammar correction corpus.

[0011] As a further solution of the present invention, the step 1 comprises the following:

[0012] Step 1.1: Generate unlabeled data with various error distributions using the existing Lao speech recognition model;

[0013] Step 1.2: Divide the unlabeled data into independent sentences with one sentence per line;

[0014] Step 1.3: Screen the sentences and remove those that are too short or too long;

[0015] Step 1.4: Preprocess the filtered sentences to remove sentences containing unrecognizable special symbols; the processed sentences are used as Lao unlabeled data with different error distributions.

[0016] As a further solution of the present invention, step 2 comprises the following:

[0017] Step 2.1, select sentences from the Lao unlabeled data base corpus as the original correct sentences;

[0018] Step 2.2, using the Meta-Llama model to generate corresponding erroneous sentences with different error distributions based on correct sentences based on certain rules and constraints;

[0019] Step 2.3: Clean and preprocess the formed erroneous sentences, remove sentences containing impurities, and form initial grammatical error correction pseudo data.

[0020] As a further solution of the present invention, step 3 comprises the following:

[0021] Step 3.1: Analyze the error distribution of Lao unlabeled data and grammar correction pseudo data, and identify common error types, including missing subject, improper word order, and repeated words;

[0022] Step 3.2: Fuse the Lao unlabeled data containing different error types and the grammatical correction pseudo data to form a fused corpus containing multiple error distributions.

[0023] As a further solution of the present invention, step 4 comprises the following:

[0024] Step 4.1, train the mbart language model with the fusion corpus containing multiple error distributions to build a Lao grammar error correction model;

[0025] Step 4.2: Use word error rate to evaluate the prediction effect of the Lao grammatical correction model to obtain high-quality Lao grammatical correction corpus.

[0026] The present invention also provides a Lao grammar correction corpus construction system based on an error distribution guidance large model, comprising: a module for executing the above-mentioned Lao grammar correction corpus construction method based on an error distribution guidance large model.

[0027] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for constructing a Laotian grammatical correction corpus based on an error distribution guidance large model is implemented.

[0028] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for constructing a Lao grammatical error correction corpus based on an error distribution guidance large model is implemented.

[0029] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the above-mentioned method for constructing Lao grammar correction corpus based on the error distribution guidance large model.

[0030] The beneficial effects of the present invention are:

[0031] 1. By combining the unlabeled data generated by the speech recognition model with the data generated by the large model, the present invention can effectively capture and analyze common grammatical errors in Lao and significantly improve the accuracy of the error correction model.

[0032] 2. The present invention reduces the reliance on manual labeling by effectively utilizing unlabeled data, reduces the time and cost of data processing, and improves research efficiency.

[0033] 3. The present invention improves the generalization ability of the model by integrating multiple data sources, so that the error correction model can show better stability and accuracy when facing different users and texts.

[0034] 4. The present invention can construct a high-quality Lao grammar correction corpus by deeply analyzing and cleaning the generated corpus. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a flow chart of the present invention; DETAILED DESCRIPTION

[0036] Example 1: Figure 1 As shown, a method for constructing a Laotian grammatical error correction corpus based on an error distribution guidance large model, the method comprising:

[0037] Step 1: Obtain Lao unlabeled data containing different error distributions; the specific steps are as follows:

[0038] Step 1.1: Generate unlabeled data with various error distributions using the existing Lao speech recognition model;

[0039] Step 1.2: Divide the unlabeled data into independent sentences with one sentence per line;

[0040] Step 1.3: Screen the sentences and remove those that are too short or too long;

[0041] Step 1.4: Preprocess the filtered sentences to remove sentences containing unrecognizable special symbols; the processed sentences are used as Lao unlabeled data with different error distributions.

[0042] Step 2: Select relatively standardized texts from the above unlabeled data as the original correct sentences, and based on certain rules and constraints, use the large model to automatically generate sentences with various error distributions to construct pseudo data, clean and remove duplicates from the generated pseudo data, and obtain the initial grammatical error correction pseudo data; Step 2 includes the following:

[0043] Step 2.1, select sentences from the Lao unlabeled data base corpus as the original correct sentences;

[0044] Step 2.2, using the Meta-Llama model to generate corresponding erroneous sentences with different error distributions based on correct sentences based on certain rules and constraints;

[0045] Step 2.3: Clean and preprocess the formed erroneous sentences, remove sentences containing impurities, and form initial grammatical error correction pseudo data.

[0046] Step 3: Fusing Lao unlabeled data and grammar correction pseudo data to further improve Lao grammar correction performance; Step 3 includes the following:

[0047] Step 3.1: Analyze the error distribution of Lao unlabeled data and grammar correction pseudo data, and identify common error types, including missing subject, improper word order, repeated words, etc.

[0048] Step 3.2: Fuse the Lao unlabeled data containing different error types and the grammatical correction pseudo data to form a fused corpus containing multiple error distributions.

[0049] Step 4: Use the fused corpus to train the language model and construct a Lao grammar correction model for Lao grammar correction; evaluate the correction prediction results to obtain the final Lao grammar correction corpus. Step 4 includes the following:

[0050] Step 4.1, train the mbart language model with the fusion corpus containing multiple error distributions to build a Lao grammar error correction model;

[0051] Step 4.2: Use word error rate to evaluate the prediction effect of the Lao grammatical correction model to obtain high-quality Lao grammatical correction corpus.

[0052] The present invention also provides a Laotian grammar error correction corpus construction system based on an error distribution guidance large model, comprising:

[0053] The first acquisition module is used to acquire Lao unlabeled data containing different error distributions;

[0054] The second acquisition module is used to acquire grammatical error correction pseudo data;

[0055] The fusion module is used to fuse Lao unlabeled data and grammatical correction pseudo data;

[0056] The training and evaluation module is used to train the language model using the fused corpus, build a Lao grammar correction model for Lao grammar correction; and evaluate the correction prediction results to obtain the final Lao grammar correction corpus.

[0057] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for constructing a Laotian grammatical correction corpus based on an error distribution guidance large model is implemented.

[0058] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for constructing a Lao grammatical error correction corpus based on an error distribution guidance large model is implemented.

[0059] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the above-mentioned method for constructing Lao grammar correction corpus based on the error distribution guidance large model.

[0060] 1) The present invention uses a fusion corpus containing multiple error distributions as the experimental data of the present invention, and uses the word error rate (WER) as an evaluation index of the prediction effect of the Lao grammar error correction model. WER is a common indicator of the performance of text error correction systems. The lower the WER, the better the error correction effect of the model. The specific calculation method is as follows:

[0061]

[0062] in:

[0063] S (Substitutions): The number of incorrectly recognized words.

[0064] D (Deletions): The number of omitted words.

[0065] I (Insertions): The number of extra words.

[0066] N (Number): The total number of words in the reference text.

[0067] 2) In order to verify the effect of the method proposed in the present invention on improving the performance of the Lao grammar error correction model, the present invention selects the official mbart language model as the benchmark model. The corpus sources and corpus scales involved are shown in Table 1:

[0068] Table 1: Sources and size of the corpus involved in the verification part

[0069]

[0070] The final experimental results of the model fusion corpus of the present invention are shown in Table 2. First, it can be seen that the increase in the original corpus can enhance the error correction performance of the baseline model and reduce the word error rate; secondly, it is found that the generalization performance of the error correction model can be effectively improved by fusing different error corpora, proving that high-quality error correction corpora help the model capture more common knowledge between languages, and can also make full use of language model knowledge, further verifying the effectiveness of this method.

[0071]

[0072]

[0073] The specific implementation modes of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above implementation modes, and various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present invention.

Claims

1. A method for constructing Lao grammar correction corpus based on error distribution guidance large model, characterized by: The method comprises: Step 1: Obtain Lao unlabeled data containing different error distributions; Step 2: Obtain grammatical error correction pseudo data; Step 3: Fusion of Lao unlabeled data and grammatical correction pseudo data; Step 4: Use the fused corpus to train the language model and build a Lao grammar correction model for Lao grammar correction; evaluate the correction prediction results to obtain the final Lao grammar correction corpus.

2. A method for constructing a Laotian grammar correction corpus based on an error distribution guidance large model according to claim 1, characterized in that: The step 1 comprises the following: Step 1.1: Generate unlabeled data with various error distributions using the existing Lao speech recognition model; Step 1.2: Divide the unlabeled data into independent sentences with one sentence per line; Step 1.3: Screen the sentences and remove those that are too short or too long; Step 1.4: Preprocess the filtered sentences to remove sentences containing unrecognizable special symbols; the processed sentences are used as Lao unlabeled data with different error distributions.

3. A method for constructing a Lao grammar correction corpus based on an error distribution guidance large model according to claim 1, characterized in that: The step 2 includes the following: Step 2.1, select sentences from the Lao unlabeled data base corpus as the original correct sentences; Step 2.2, using the Meta-Llama model to generate corresponding erroneous sentences with different error distributions based on correct sentences based on certain rules and constraints; Step 2.3: Clean and preprocess the formed erroneous sentences, remove sentences containing impurities, and form initial grammatical error correction pseudo data.

4. A method for constructing a Lao grammar correction corpus based on an error distribution guidance large model according to claim 1, characterized in that: The step 3 includes the following: Step 3.1: Analyze the error distribution of Lao unlabeled data and grammar correction pseudo data, and identify common error types, including missing subject, improper word order, and repeated words; Step 3.2: Fuse the Lao unlabeled data containing different error types and the grammatical correction pseudo data to form a fused corpus containing multiple error distributions.

5. A method for constructing a Lao grammar correction corpus based on an error distribution guidance large model according to claim 1, characterized in that: The step 4 includes the following: Step 4.1, train the mbart language model with the fusion corpus containing multiple error distributions to build a Lao grammar error correction model; Step 4.2: Use word error rate to evaluate the prediction effect of the Lao grammatical correction model to obtain high-quality Lao grammatical correction corpus.

6. A Laotian grammar correction corpus construction system based on an error distribution guidance large model, characterized in that: include: A module for executing a method for constructing a Laotian grammatical error correction corpus based on an error distribution guidance large model as described in any one of claims 1 to 5.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, a method for constructing Lao grammatical error correction corpus based on an error distribution guidance large model as described in any one of claims 1 to 5 is implemented.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for constructing Lao grammatical error correction corpus based on an error distribution guidance large model as described in any one of claims 1 to 5 is implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for constructing Lao grammatical error correction corpus based on an error distribution guidance large model as described in any one of claims 1 to 5 is implemented.

Images