Voice input evaluation method and device for occupational skill evaluation, computer equipment and readable storage medium

Abstract

The invention discloses a voice input evaluation method and device for occupational skill evaluation, computer equipment and a readable storage medium, and the method comprises the steps: firstly obtaining examination point configuration data of a target evaluation scene, including a standard answer text; and then, constructing a related sample data set, and training the enhanced SeACoParaformer architecture based on the data set to obtain a special speech recognition model. And then, inputting a to-be-evaluated voice stream into the model for identification, and generating a corresponding text. And finally, scoring the to-be-evaluated text according to the standard answer text, and comprehensively considering a keyword score and a semantic similarity score as a final occupational skill evaluation result. Through the design, the efficiency and accuracy of occupational skill assessment are effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a voice input evaluation method, device, computer equipment and readable storage medium for professional skill evaluation. Background Art

[0002] With the development of technology, professional skills assessment has become an important way to measure personal professional ability. Traditional assessment methods mostly rely on written tests or practical operations, but these methods have limitations in efficiency and objectivity. Summary of the invention

[0003] The purpose of the present invention is to provide a voice input evaluation method, device, computer equipment and readable storage medium for professional skill assessment.

[0004] In a first aspect, an embodiment of the present invention provides a voice input evaluation method for professional skill evaluation, comprising:

[0005] Acquire test point configuration data of a target professional skill assessment scenario, wherein the test point configuration data includes a standard answer text;

[0006] Constructing a sample data set for the target occupational skill assessment scenario, and training the enhanced SeACoParaformer architecture based on the sample data set to obtain a trained occupational skill assessment speech recognition model;

[0007] Inputting the speech stream to be evaluated into the occupational skill evaluation speech recognition model for recognition, and obtaining the text to be evaluated corresponding to the speech stream to be evaluated;

[0008] Scoring the text to be evaluated based on the standard answer text to obtain a keyword score and a semantic similarity score;

[0009] The keyword score and the semantic similarity score are used as the professional skill evaluation score corresponding to the speech stream to be evaluated.

[0010] In a possible implementation, the enhanced SeACoParaformer architecture includes a hot word component and a synonym enhancement component, and the inputting of the speech stream to be evaluated into the occupational skill evaluation speech recognition model for recognition to obtain the text to be evaluated corresponding to the speech stream to be evaluated includes:

[0011] The speech stream to be evaluated is optimized by the hot word component to obtain a text to be processed and input into the synonym enhancement component;

[0012] Performing synonym enhancement on the text to be processed based on the synonym enhancement component to obtain an enhanced text sequence;

[0013] The text sequence is used as the text to be evaluated corresponding to the speech stream to be evaluated.

[0014] In a possible implementation, the performing synonym enhancement on the to-be-processed text based on the synonym enhancement component to obtain an enhanced text sequence includes:

[0015] Performing word segmentation processing on the text to be processed to obtain multiple word segments, and using a synonym dictionary to determine a synonym set for each word segment; using a pre-trained embedding table to retrieve a word segmentation embedding vector and a synonym set embedding vector corresponding to each word segment and each synonym set, and determining the position codes corresponding to each of the word segmentation embedding vector and the synonym set embedding vector, so as to realize synonym embedding retrieval;

[0016] An attention mechanism is used to combine the word segmentation embedding vector, the synonym set embedding vector, and the position encodings corresponding to the word segmentation embedding vector and the synonym set embedding vector to determine an attention weight to achieve context fusion;

[0017] Based on the attention weights, weighted averages are performed on the synonym set embedding vectors to obtain fused synonym set embedding vectors, thereby achieving word embedding fusion;

[0018] According to the word segmentation embedding vector, the word weight pre-configured by the word segmentation embedding vector, and the synonym set embedding vector, an enhanced word segmentation embedding vector of each word segment is calculated to achieve enhanced vector generation;

[0019] Decoding the enhanced word segmentation embedding vector of each word segment in combination with the cross-attention mechanism to obtain a decoding state of the enhanced word segmentation embedding vector of each word segment, wherein the decoding state combines the self-attention mechanism and the cross-self-attention mechanism;

[0020] Mapping each of the decoding states to a vocabulary to obtain a score vector corresponding to the vocabulary for each of the word segments; transforming the score vector using a softmax function to obtain a probability distribution of each word segment corresponding to the vocabulary, and outputting an output word segment of each word segment included in the text to be processed based on the probability distribution to achieve vocabulary selection;

[0021] The output word segmentations are integrated to obtain the enhanced text sequence.

[0022] In a possible implementation, the test point configuration data includes a plurality of test question links, and the inputting of the voice stream to be evaluated into the professional skill evaluation speech recognition model for recognition to obtain the text to be evaluated corresponding to the voice stream to be evaluated includes:

[0023] In the target test question phase, the speech stream to be evaluated is input into the occupational skill evaluation speech recognition model for recognition, and at least one text to be evaluated corresponding to the speech stream to be evaluated is obtained;

[0024] The at least one text to be evaluated is integrated to obtain a text to be evaluated corresponding to the target test question segment.

[0025] In a possible implementation, scoring the text to be evaluated based on the standard answer text to obtain a keyword score and a semantic similarity score includes:

[0026] Based on the standard answer text, a sliding window of the text to be evaluated corresponding to the target test question is formed;

[0027] Based on the size of the sliding window, semantic segmentation is performed on the text to be evaluated corresponding to the target test question link to obtain a plurality of short sentences to be evaluated;

[0028] Based on the standard answer text, each of the short sentences to be evaluated is scored in turn to obtain a keyword score and a semantic similarity score.

[0029] In a possible implementation, the test point configuration further includes keywords, keyword weights, and scoring strategies, and scoring the standard answer text and each of the short sentences to be evaluated in turn to obtain a keyword score and a semantic similarity score, including:

[0030] Acquire a keyword scoring threshold and a semantic matching scoring threshold according to the scoring strategy;

[0031] Perform keyword matching on each of the short sentences to be evaluated according to the keywords and the keyword weights to obtain a pending keyword score;

[0032] In the case where the pending keyword score exceeds the keyword scoring threshold, taking the pending keyword score as the final keyword score;

[0033] Perform semantic matching on each of the short sentences to be evaluated according to the standard answer text to obtain a pending semantic similarity score;

[0034] In the case that the pending semantic similarity score exceeds the semantic matching score threshold, the pending semantic similarity score is used as the final semantic similarity score.

[0035] In a possible implementation, the constructing of a sample data set for the target occupational skill assessment scenario includes:

[0036] Filtering voice data and text data related to the target professional skill assessment scenario from a plurality of different professional skill assessment history records;

[0037] Performing at least one data enhancement operation of adding noise, adjusting speech speed, and adjusting intonation on the filtered speech data;

[0038] The speech data after the data enhancement operation is combined with the corresponding text data to form the sample data set.

[0039] In a second aspect, an embodiment of the present invention provides a speech input evaluation device for professional skill evaluation, comprising:

[0040] An acquisition module is used to acquire test site configuration data of a target occupational skill assessment scenario, wherein the test site configuration data includes a standard answer text; construct a sample data set for the target occupational skill assessment scenario, and train the enhanced SeACoParaformer architecture based on the sample data set to obtain a trained occupational skill assessment speech recognition model; input the to-be-assessed speech stream into the occupational skill assessment speech recognition model for recognition, and obtain the to-be-assessed text corresponding to the to-be-assessed speech stream;

[0041] An evaluation module is used to score the text to be evaluated based on the standard answer text to obtain a keyword score and a semantic similarity score; and use the keyword score and the semantic similarity score as the professional skill evaluation score corresponding to the voice stream to be evaluated.

[0042] In a third aspect, an embodiment of the present invention provides a computer device, comprising a processor and a non-volatile memory storing computer instructions, wherein when the computer instructions are executed by the processor, the computer device executes the method described in the first aspect.

[0043] In a fourth aspect, an embodiment of the present invention provides a readable storage medium, wherein the readable storage medium includes a computer program, and when the computer program is executed, the computer device where the readable storage medium is located is controlled to execute the method described in the first aspect.

[0044] Compared with the prior art, the beneficial effects provided by the present invention include: adopting a voice input evaluation method, device, computer equipment and readable storage medium for occupational skill evaluation disclosed in the present invention, by obtaining the test point configuration data of the target evaluation scenario, including the standard answer text. Then, a relevant sample data set is constructed, and the enhanced SeACoParaformer architecture is trained based on the data set to obtain a dedicated speech recognition model. Subsequently, the voice stream to be evaluated is input into the model for recognition to generate the corresponding text. Finally, the text to be evaluated is scored according to the standard answer text, and the keyword score and semantic similarity score are comprehensively considered as the final occupational skill evaluation result. Such a design effectively improves the efficiency and accuracy of occupational skill evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can also be obtained based on these drawings without creative work.

[0046] Figure 1 A schematic diagram of the steps of a voice input assessment method for vocational skill assessment provided by an embodiment of the present invention;

[0047] Figure 2 A schematic block diagram of the structure of a speech input evaluation device for professional skill evaluation provided by an embodiment of the present invention;

[0048] Figure 3 A schematic block diagram of the structure of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0050] The specific implementation modes of the present invention are described in detail below in conjunction with the accompanying drawings.

[0051] In order to solve the technical problems in the aforementioned background technology, Figure 1 A flow chart of a voice input assessment method for professional skill assessment provided in an embodiment of the present disclosure is provided, and the voice input assessment method for professional skill assessment is introduced in detail below.

[0052] Step S201, obtaining test point configuration data of a target professional skill assessment scenario, wherein the test point configuration data includes a standard answer text;

[0053] Step S202, constructing a sample data set for the target occupational skill assessment scenario, and training the enhanced SeACoParaformer architecture based on the sample data set to obtain a trained occupational skill assessment speech recognition model;

[0054] Step S203, inputting the speech stream to be evaluated into the occupational skill evaluation speech recognition model for recognition, and obtaining the text to be evaluated corresponding to the speech stream to be evaluated;

[0055] Step S204, scoring the text to be evaluated based on the standard answer text to obtain a keyword score and a semantic similarity score;

[0056] Step S205: Using the keyword score and the semantic similarity score as the professional skill evaluation score corresponding to the speech stream to be evaluated.

[0057] In the embodiment of the present invention, for example, in the context of professional skill evaluation, the server is responsible for the management and data processing of the entire evaluation process. Taking the chef professional skill evaluation as an example, when a specific chef skill evaluation is conducted, the target professional skill evaluation scenario is the chef skill evaluation scenario.

[0058] The server first needs to obtain the test point configuration data for this scenario. The test point configuration data contains many important information, the most important of which is the standard answer text. For the chef skill assessment scenario, there will be different assessment links, such as ingredient preparation, cooking techniques, and dish seasoning.

[0059] In the ingredient preparation phase, the standard answer text will describe in detail the correct ingredient selection criteria, for example, "For a classic French steak, grain-fed beef tenderloin should be used, weighing between 200-250 grams. The freshness of the meat should ensure that the surface is shiny, odorless, and elastic when pressed." In the cooking technique phase, the standard answer text will mention things like "When frying steak, first heat the pan over high heat, put the steak in, and fry each side for 2-3 minutes to ensure a golden brown crust forms on the surface, then turn to low heat and continue frying until the inside reaches the desired degree of doneness (such as the corresponding internal temperature requirements for medium-rare, medium-rare, etc.)." These standard answer texts are an important basis for subsequent evaluations. The server obtains these test point configuration data through a pre-configured database or from a dedicated evaluation standard management module.

[0060] The server screens data from numerous professional skill assessment history records. Taking the chef skill assessment as an example, the server will find voice data and text data related to the current target scenario (such as French chef skill assessment) from previous chef skill assessment history records. These history records may be voice answer records from different chefs in previous assessment processes and corresponding standard answer text records.

[0061] For example, when a chef describes the process of frying steak, the voice record mentions "I heat the pan first, then put the steak in, fry it for a while and then turn it over." The corresponding standard answer text is the precise process of frying steak mentioned above. The server will filter out these relevant voice and text data.

[0062] The server will perform data enhancement operations on the filtered voice data.

[0063] Noise addition: simulates noise interference in a real kitchen environment, such as adding the sound of a burning stove, the sound of a range hood, etc. For example, the originally clear voice of a chef describing the steps of frying a steak, "I heat the pan first, then put the steak in, fry it for a while and turn it over", will become a voice with some noisy background sound after adding noise, but the content remains unchanged.

[0064] Speed ​​adjustment: In the actual operation, the chef's speed of speech may vary due to nervousness or different levels of proficiency. The server will speed up or slow down the normal speech, such as adjusting the above speech to a fast "I will heat the pan first, then put the steak in and fry it for a while, then turn it over" or a slow "I will heat the pan slowly, then put the steak in carefully, fry it for a while, and then turn it over."

[0065] Adjusting the tone of voice: Different chefs have different intonations when expressing themselves. Servers will adjust the intonation of their voices, such as changing a plain tone to an emphatic tone to emphasize certain key steps.

[0066] Combine the voice data after data enhancement with the corresponding text data. For example, the voice data about the steps of frying steak after the above processing is combined with the corresponding standard answer text to form part of the sample data set. This sample data set covers the combination of voice and text data in multiple links and situations in chef skill assessment.

[0067] The server uses the constructed sample dataset to train the enhanced SeACoParaformer architecture. This architecture has some special components, such as hot word component and synonym enhancement component. During the training process, the speech data in the sample dataset is used as input and the corresponding text data is used as output target.

[0068] For example, for a large amount of speech data containing vocabulary related to chef skills (such as hot words such as fry, stir-fry, grill, and ingredient names), the hot word component can learn the characteristic patterns of these hot words in speech, so as to more accurately identify these key hot words in subsequent recognition. The synonym enhancement component helps to deal with the different expressions that chefs may use in the description process. For example, "fried steak" may also be said as "steak fried in oil". By learning the synonym relationships in the sample data set, the model can better understand the meaning of different expressions and accurately identify them. After multiple iterations of training until the model converges, a trained professional skill assessment speech recognition model is obtained.

[0069] Suppose there is a chef who is undergoing a skill assessment. He describes the process of making French steak into a microphone, and the server receives the voice stream to be assessed. The server inputs the voice stream to be assessed into the previously trained professional skill assessment speech recognition model. The model will recognize the voice based on the previously learned voice features and language patterns and convert the voice into corresponding text. For example, if the chef says, "I will heat up the pan first, then put the selected steak in, and turn it over when it turns golden brown when it is fried", the model will obtain the corresponding text after recognition, "I will heat up the pan first, then put the selected steak in, and turn it over when it turns golden brown when it is fried".

[0070] If the speech stream to be evaluated contains some hot words in the chef skill evaluation scenario, such as "steak" and "frying", the hot word component will first optimize these hot words. Suppose the chef says "steak" with a heavy accent, the hot word component can accurately identify it as "steak" based on the pre-learned hot word features, and then obtain the text to be processed after hot word optimization.

[0071] Next, the text to be processed is input into the synonym enhancement component.

[0072] Determination of word segmentation and synonym sets: For the text to be processed, "I first heat the pan, then put the selected steak in, and turn it over when it turns golden yellow when it is fried", word segmentation is performed to obtain multiple word segmentations, such as "I" "first" "heat" the "pan" "then" "put" the "selected" steak "in" "fry" "when" "see" the "color" "turn" "golden yellow" "then" "turn over". The synonym set of each word segmentation is determined using a synonym dictionary. For example, the synonym set of "heat" may include "heat to very hot" "burn" and so on; the synonym set of "turn over" may include "turn over" "turn the noodles over" and so on.

[0073] Synonym embedding retrieval: Use the pre-trained embedding table to retrieve the word segment embedding vector and synonym set embedding vector corresponding to each word segment and each synonym set, and determine their corresponding position codes. For example, for the word "burn hot" and its synonym set "heat to very hot", their embedding vectors and position codes are obtained respectively.

[0074] Context fusion: The attention mechanism is used to combine the word segmentation embedding vector, the synonym set embedding vector, and their corresponding position encoding to determine the attention weight. For example, in the context of the chef's skill description, the word "heat" is closely related to the subsequent "frying steak" step. The attention mechanism will give this word and its synonyms in this context a reasonable weight, achieving context fusion.

[0075] Word embedding fusion: The synonym set embedding vectors are weighted averaged based on the attention weights to obtain the fused synonym set embedding vectors. For example, for the embedding vectors of the synonym set of "burning", the weighted average is calculated based on the attention weights to obtain the fused vector.

[0076] Enhanced vector generation: Based on the word embedding vector, the pre-configured word weight of the word, and the embedding vector of the synonym set, the enhanced word embedding vector of each word is calculated. For example, for the word "burning", the enhanced embedding vector is calculated based on the previously obtained information.

[0077] Decoding operation: Combine the cross-attention mechanism to decode the enhanced word embedding vector of each word, and obtain the decoding state of the enhanced word embedding vector of each word. This decoding state combines the self-attention mechanism and the cross-self-attention mechanism. For example, for the word "burning", after decoding, a decoding state containing more semantic information is obtained.

[0078] Vocabulary selection: Map each decoding state to the vocabulary to obtain the score vector of each word segmentation corresponding to the vocabulary; use the softmax function to transform the score vector to obtain the probability distribution of each word segmentation corresponding to the vocabulary, and output the output word segmentation of each word segmentation included in the processed text based on the probability distribution. For example, for the word "burning", select the output word segmentation that best matches the semantics based on the probability distribution (it may still be "burning" or the most appropriate one of its synonyms).

[0079] Integration to obtain the text to be evaluated: Finally, all the output segmentations are integrated to obtain the enhanced text sequence as the text to be evaluated corresponding to the speech stream to be evaluated.

[0080] If the chef's voice stream is long or has intermittent descriptions in a certain part of the test, at least one text to be evaluated will be obtained. For example, when the chef describes the steak making process, he divides the voice description into two parts. The server recognizes and obtains two texts to be evaluated respectively: "I heat the pan first, then put the selected steak in" and "turn it over when it turns golden brown when it is fried." The server integrates this at least one text to be evaluated to obtain the text to be evaluated corresponding to this part of the test: "I heat the pan first, then put the selected steak in, and turn it over when it turns golden brown when it is fried."

[0081] Take the steak making process in the chef skill assessment as an example. Based on the standard answer text, the server forms a sliding window for the text to be evaluated in this process. Assume that the standard answer text is "I first heat the pan, then put in the grain-fed beef tenderloin, weighing 200-250 grams and of acceptable freshness. When frying, use high heat to form a golden brown crust on the surface (2-3 minutes per side), then turn to low heat and fry until the appropriate degree of doneness, and turn it over when the color turns golden yellow."

[0082] If the sliding window size is set to 5 word segmentation lengths, the server will perform semantic segmentation on the text to be evaluated "I heat the pan first, then put the selected steak in, and turn it over when it turns golden brown" based on this size, and obtain multiple short sentences to be evaluated, such as "I heat the pan first", "then put the selected steak in", and "turn it over when it turns golden brown".

[0083] Keyword matching: The test point configuration also includes keywords, keyword weights, and scoring strategies. Assuming that the keywords include "steak", "pan", "fried", "golden", etc., the keyword weights are set to 0.2, 0.1, 0.3, 0.2, etc. according to the importance. For each short sentence to be evaluated, keyword matching is performed based on the keywords and keyword weights. For example, for the short sentence "I will heat up the pan first", which contains the keyword "pan", a pending keyword score (such as 0.1) is calculated based on the weight. If the pending keyword score exceeds the keyword scoring threshold (assuming it is 0.05), the pending keyword score is used as the final keyword score.

[0084] Semantic matching: semantically match each short sentence to be evaluated according to the standard answer text to obtain a pending semantic similarity score. For example, for the short sentence "I heat the pot first", the semantic similarity is calculated with the corresponding part in the standard answer text (some semantic similarity algorithms can be used, such as the cosine similarity algorithm, etc.), and a pending semantic similarity score (such as 0.8) is obtained. If this pending semantic similarity score exceeds the semantic matching score threshold (assuming it is 0.6), this pending semantic similarity score is used as the final semantic similarity score.

[0085] After the previous steps, the server obtained the keyword score and semantic similarity score. For the chef's skill evaluation, assume that the keyword score is 0.6 and the semantic similarity score is 0.7. The server uses these two scores as the professional skill evaluation score corresponding to the speech stream to be evaluated in this steak making link. This score can be used as an important basis for the comprehensive evaluation of the chef's skills. It can be summarized together with the scores of other links to finally get the chef's total score in the entire skill evaluation.

[0086] In an embodiment of the present invention, the enhanced SeACoParaformer architecture includes a hot word component and a synonym enhancement component. The voice stream to be evaluated is input into the professional skill assessment speech recognition model for recognition to obtain the text to be evaluated corresponding to the voice stream to be evaluated, which can be implemented through the following example.

[0087] The speech stream to be evaluated is optimized by the hot word component to obtain a text to be processed and input into the synonym enhancement component;

[0088] Performing synonym enhancement on the text to be processed based on the synonym enhancement component to obtain an enhanced text sequence;

[0089] The text sequence is used as the text to be evaluated corresponding to the speech stream to be evaluated.

[0090] In the embodiment of the present invention, for example, it is assumed that the current professional skill assessment scenario is a clinical skill assessment of a doctor. The voice stream to be assessed is a voice record of a doctor describing the process of diagnosing pneumonia on a patient.

[0091] In the medical scenario, hot words may include words closely related to pneumonia diagnosis, such as "pneumonia", "auscultation", "X-ray", "cough", and "fever". When the server receives the doctor's voice stream to be evaluated, the hot word component starts working.

[0092] For example, a doctor may say "Fei Yan" instead of "Pneumonia" due to his accent or fast speaking. The hotword component can recognize that this is an incorrect pronunciation of the hotword "Pneumonia" based on the hotword pattern learned during pre-training, and correct it to "Pneumonia". For another example, a doctor may say "I first performed auscultation", and the hotword component can correct "Auscultation" to "Auscultation" based on the phonetic features of the hotword "Auscultation". After the optimization of the hotword component, the text to be processed is obtained, such as "I diagnosed pneumonia in this patient with cough and fever symptoms, first performed auscultation, and then checked the X-ray".

[0093] The server inputs the processed text optimized by the hot word component into the synonym enhancement component. This process is like passing the preliminarily sorted information to the next more sophisticated processing unit, preparing for the subsequent deeper mining of the text's semantic information.

[0094] For the input text to be processed "I diagnosed pneumonia in this patient with cough and fever symptoms by first performing auscultation and then checking the X-ray film", the synonym enhancement component first performs word segmentation processing.

[0095] The obtained segmented words include "I", "to", "this", "have", "cough", "and", "fever", "symptoms", "of", "patient", "conducted", "diagnosis", "first", "conducted", "auscultation", "then", "checked", and "X-rays". Then, the synonym dictionary is used to determine the synonym set of each segmented word.

[0096] For example, the synonym set for "cough" may include "dry cough", "sputum", etc. (in some cases, doctors will describe the type of cough more specifically); the synonym set for "fever" may include "fever" and "high body temperature"; the synonym set for "diagnosis" may include "judgment" and "diagnosis"; the synonym set for "view" may include "observation" and "review", etc.

[0097] The pre-trained embedding table is used to retrieve the word segmentation embedding vector and the synonym set embedding vector corresponding to each word segment and each synonym set, and their respective corresponding position encodings are determined.

[0098] Taking the word "cough" as an example, the word embedding vector corresponding to "cough" is found in the pre-trained embedding table. This vector can represent the semantic features of the word "cough" in the vector space. At the same time, for words in the synonym set such as "dry cough" and "sputum", their corresponding synonym set embedding vectors are also found. And according to the position of these words in the text to be processed, their corresponding position codes are determined. Position codes can reflect the order information of words in a sentence, which is very important for understanding semantics.

[0099] The attention weight is determined by combining the word segmentation embedding vector, the synonym set embedding vector, and their corresponding position encoding.

[0100] In the text to be processed in this medical scenario, "cough" and "fever" are symptoms of patients and are closely related to "diagnosis of pneumonia". Through the attention mechanism, this semantic relationship is taken into account when calculating the attention weight. For example, the association weight between the word "cough" and "diagnosis of pneumonia" will be relatively high, because cough is one of the common symptoms of pneumonia. Similarly, "auscultation" is also strongly associated with "diagnosis of pneumonia", so when calculating the attention weight, these related participles and synonym sets will be given appropriate weights to achieve context fusion.

[0101] The synonym set embedding vectors are weighted averaged based on the attention weights to obtain the fused synonym set embedding vectors.

[0102] Continuing with the example of "cough", if the attention weights corresponding to the synonym sets of "cough" "dry cough" and "sputum" are 0.3 and 0.2 (hypothesis), then the synonym set embedding vectors of "dry cough" and "sputum" are weighted averaged according to these weights to obtain the fused synonym set embedding vector. This fused vector can comprehensively consider the importance of synonyms in a specific context and more accurately represent the semantic information of "cough" and its synonyms in this medical scenario.

[0103] The enhanced word segmentation embedding vector of each word is calculated based on the word segmentation embedding vector, the pre-configured word weight of the word segmentation, and the synonym set embedding vector.

[0104] For the word "pneumonia", it has a pre-configured word weight (assuming it is 0.5, because pneumonia is the core word in this diagnosis scenario). Combined with its word embedding vector and information related to synonyms (if any), the enhanced word embedding vector is calculated. This enhanced vector can more comprehensively represent the semantic information of the word "pneumonia" in the current medical scenario, taking into account the importance of the word itself and possible synonym information.

[0105] The enhanced word segmentation embedding vector of each word is decoded in combination with the cross-attention mechanism to obtain the decoding state of the enhanced word segmentation embedding vector of each word. This decoding state combines the self-attention mechanism and the cross-self-attention mechanism.

[0106] For example, for the word "auscultation", its enhanced word embedding vector interacts with other related words (such as "pneumonia" and "patients") through the cross-attention mechanism, and uses the self-attention mechanism to focus on its own internal semantic information, and further integrates the relationship with other word segments through the cross-self-attention mechanism, thus obtaining a decoding state containing rich semantic information. This decoding state can better reflect the semantic role of "auscultation" in the entire pneumonia diagnosis process.

[0107] Map each decoding state to the vocabulary to obtain the score vector of each word segmentation corresponding to the vocabulary; use the softmax function to transform the score vector to obtain the probability distribution of each word segmentation corresponding to the vocabulary, and output the output word segmentation of each word segmentation included in the text to be processed based on the probability distribution.

[0108] Take the word "view" as an example. After its decoding state is mapped to the vocabulary, a score vector is obtained. This score vector represents the similarity score between "view" and each word in the vocabulary. Then use the softmax function to convert this score vector into a probability distribution. Assume that in this probability distribution, "view" itself has the highest probability, but "observe" also has a certain probability. According to this probability distribution, the final output word may still be "view" (if its probability is high enough), but if the probability of "observe" meets the requirements under certain conditions, "observe" may also be output.

[0109] Finally, all the output segmented words are integrated to obtain the enhanced text sequence. For example, after the above processing, the output segmented words are "I", "to", "this person", "have", "dry cough" (if "dry cough" is selected in the vocabulary selection), "and", "fever" (if "fever" is selected), "symptoms", "of", "patients", "underwent", "pneumonia", "diagnosed" (if "diagnosed" is selected), "first", "underwent", "auscultation", "then", "observation" (if "observation" is selected), "X-ray". This enhanced text sequence is the text to be evaluated corresponding to the speech stream to be evaluated. Compared with the original text to be processed, it is more accurate and rich in semantic expression, and can provide better input for the subsequent scoring process.

[0110] In the embodiment of the present invention, the synonym enhancement component is used to perform synonym enhancement on the text to be processed to obtain an enhanced text sequence, which can be implemented through the following examples.

[0111] Performing word segmentation processing on the text to be processed to obtain multiple word segments, and using a synonym dictionary to determine a synonym set for each word segment; using a pre-trained embedding table to retrieve a word segmentation embedding vector and a synonym set embedding vector corresponding to each word segment and each synonym set, and determining the position codes corresponding to each of the word segmentation embedding vector and the synonym set embedding vector, so as to realize synonym embedding retrieval;

[0112] An attention mechanism is used to combine the word segmentation embedding vector, the synonym set embedding vector, and the position encodings corresponding to the word segmentation embedding vector and the synonym set embedding vector to determine an attention weight to achieve context fusion;

[0113] Based on the attention weights, weighted averages are performed on the synonym set embedding vectors to obtain fused synonym set embedding vectors, thereby achieving word embedding fusion;

[0114] According to the word segmentation embedding vector, the word weight pre-configured by the word segmentation embedding vector, and the synonym set embedding vector, an enhanced word segmentation embedding vector of each word segment is calculated to achieve enhanced vector generation;

[0115] Decode the enhanced token embedding vectors of each of the tokens by combining the cross-attention mechanism to obtain the decoding states of the enhanced token embedding vectors of each of the tokens, where the decoding states combine the self-attention mechanism and the cross-self-attention mechanism;

[0116] Map each of the decoding states to a vocabulary to obtain a score vector of each of the tokens corresponding to the vocabulary; use the softmax function to transform the score vector to obtain a probability distribution of each of the tokens corresponding to the vocabulary, and output an output token of each of the tokens included in the text to be processed based on the probability distribution, thereby implementing vocabulary selection;

[0117] Integrate the output tokens to obtain the enhanced text sequence.

[0118] In an embodiment of the present invention, by way of example, it is assumed that in a teacher teaching skill assessment scenario, the text to be processed is "When the teacher is telling a historical event, multimedia materials are used to assist teaching, and this way can enable students to better understand complex historical plots".

[0119] The server performs tokenization on this text to be processed, and the obtained tokens are "teacher", "when", "telling", "historical event", "using", "multimedia materials", "to", "assist teaching", "this way", "can", "enable", "students", "better", "understand", "complex", "historical plots".

[0120] Then, use a thesaurus to determine the synonym set of each token. For example, the synonym set of "telling" may include "lecturing", "expounding"; the synonym set of "multimedia materials" may be "multimedia materials", "multimedia resources"; the synonym set of "assist teaching" may include "assist teaching", "facilitate teaching"; the synonym set of "understand" may be "comprehend", "grasp", etc.

[0121] For the tokens and synonym sets in the above educational scenario, the server uses a pre-trained embedding table for retrieval.

[0122] Taking the token "telling" as an example, in the pre-trained embedding table, there is a specific vector to represent the semantic feature of "telling", which is the token embedding vector. For the words in the synonym sets such as "lecturing" and "expounding", there are also corresponding synonym set embedding vectors.

[0123] At the same time, determine the position code. In the text to be processed, "When the teacher narrated historical events, he used multimedia materials to assist teaching, which enabled students to better understand complex historical plots", "narration" is in the third position. According to the predefined coding rules, its position code is determined. Similarly, the position codes of synonyms such as "lecture" and "explanation" are also determined based on their possible text position relationships.

[0124] In this educational scenario, “narration” is semantically related to “historical events” and other participles such as “students” and “understanding”.

[0125] The attention weight is determined by combining the word segmentation embedding vector of "narrate", the synonym set embedding vector of its synonym set ("lecture", "explain"), and their respective position encodings through the attention mechanism. For example, since "narrate" is closely related to "historical events", this relationship will be taken into account when calculating the attention weight, giving "narrate" and its synonyms appropriate weights in this context. If the word "narrate" is more important in its association with "historical events", then its attention weight in this semantic relationship will be relatively high, achieving context fusion.

[0126] Taking the synonym set of "tell" "teach" and "explain" as an example, suppose that the attention weight of "teach" is determined to be 0.4 and the attention weight of "explain" is determined to be 0.6 through the attention mechanism.

[0127] The server performs a weighted average of the synonym set embedding vectors of "lecture" and "elaboration" based on these weights. If the synonym set embedding vector of "lecture" is vector A, and the synonym set embedding vector of "elaboration" is vector B, then the fused synonym set embedding vector = 0.4*vector A+0.6*vector B. In this way, a fused synonym set embedding vector that comprehensively considers the contextual relationship is obtained.

[0128] For the word "narrate", assume that its pre-configured word weight is 0.3 (because narration is a relatively important behavior in the teaching scenario).

[0129] Its word segmentation embedding vector is vector C, and the fused synonym set embedding vector is the vector calculated above (assuming it is vector D). Then the enhanced word segmentation embedding vector of "narrate" is calculated according to the formula = 0.3*vector C+(1-0.3)*vector D. This enhanced word segmentation embedding vector combines the importance of the word itself and the information of synonyms, and can more comprehensively represent the semantics of "narrate" in this educational scenario.

[0130] Taking the word "assisted teaching" as an example, its enhanced word segmentation embedding vector contains rich semantic information related to synonyms.

[0131] The server decodes it in combination with the cross-attention mechanism. The cross-attention mechanism allows the enhanced word segmentation embedding vector of "auxiliary teaching" to interact with the enhanced word segmentation embedding vectors of other related word segmentations (such as "teachers", "students", "multimedia materials", etc.). At the same time, the self-attention mechanism focuses on the internal semantic structure of "auxiliary teaching" itself, and the cross-self-attention mechanism further integrates its complex relationship with other word segmentations. Through the synergy of these mechanisms, the decoding state of the enhanced word segmentation embedding vector of the word "auxiliary teaching" is obtained, and this decoding state contains a deeper semantic understanding.

[0132] Taking the word "understanding" as an example, after its decoding state is mapped to the vocabulary (this vocabulary contains many words in the educational scenario), a score vector is obtained. This score vector represents the similarity score between "understanding" and each word in the vocabulary.

[0133] Then use the softmax function to convert this score vector into a probability distribution. Assume that in this probability distribution, "understanding" itself has the highest probability, but "comprehension" also has a certain probability. If this probability distribution meets certain conditions (for example, the probability of "comprehension" exceeds a certain set threshold or the probability of "comprehension" is lower than a certain lower limit), then "comprehension" may be output as the output participle of this participle; otherwise, "comprehension" is output.

[0134] After the above vocabulary selection, each word segmentation is output. For example, "narration" in the original text to be processed may be output as "teaching" (if selected in the vocabulary selection), "multimedia materials" may be output as "multimedia resources", etc.

[0135] The server integrates all the output word segments to obtain an enhanced text sequence, such as "When the teacher taught historical events, he used multimedia resources to assist teaching, which enabled students to better understand complex historical plots." Compared with the original text to be processed, this enhanced text sequence is more accurate and rich in semantic expression, and can better reflect the teacher's behavior and purpose in the teaching process, providing a better basis for subsequent evaluation.

[0136] In an embodiment of the present invention, the test point configuration data includes multiple test question links, and the voice stream to be evaluated is input into the professional skill assessment speech recognition model for recognition to obtain the text to be evaluated corresponding to the voice stream to be evaluated, which can be implemented through the following example.

[0137] In the target test question phase, the speech stream to be evaluated is input into the occupational skill evaluation speech recognition model for recognition, and at least one text to be evaluated corresponding to the speech stream to be evaluated is obtained;

[0138] The at least one text to be evaluated is integrated to obtain a text to be evaluated corresponding to the target test question segment.

[0139] In the embodiment of the present invention, for example, it is assumed that an automobile maintenance skill assessment is being conducted, and this assessment includes multiple test links, such as automobile fault diagnosis link, parts replacement link, maintenance tool use link, etc. Now, focus on the target test link of automobile fault diagnosis link.

[0140] When the maintenance technician describes his process of diagnosing the car fault into the microphone, the server receives the speech stream to be evaluated and inputs it into the pre-trained professional skill evaluation speech recognition model.

[0141] Since the maintenance technician may interrupt the description or explain the fault diagnosis ideas from different aspects, the model will get at least one text to be evaluated. For example, the maintenance technician first said, "This car is difficult to start, and the engine has abnormal vibration when running. I initially suspected that it was a problem with the spark plugs because the spark plugs have been used for a long time." After recognition, the model obtains a text to be evaluated, "This car is difficult to start, and the engine has abnormal vibration when running. I initially suspected that it was a problem with the spark plugs because the spark plugs have been used for a long time."

[0142] Then, the maintenance technician added, "And I checked the ignition coil and found that the connecting line of the ignition coil was not loose, so it is basically certain that the spark plug is faulty." The model again recognized another text to be evaluated, "Check the ignition coil, and its connecting line is not loose, so it is basically certain that the spark plug is faulty."

[0143] Taking software programming skills assessment as an example, one of the target test questions is algorithm design.

[0144] When the programmer explains the sorting algorithm he designed, the server receives the corresponding speech stream to be evaluated and inputs it into the speech recognition model. The programmer may first describe the basic idea of ​​the algorithm: "I designed an improved bubble sort algorithm. It is different from the traditional bubble sort in that each time when comparing adjacent elements, if the elements are already in order, the comparison is skipped to improve efficiency." The model recognizes a text to be evaluated: "Design an improved bubble sort algorithm. Unlike the traditional one, the comparison is skipped when the adjacent elements are already in order to improve efficiency."

[0145] Afterwards, the programmer further explained the time complexity of the algorithm: "In the best case, the time complexity of this algorithm can reach O(n), because when the elements are basically ordered, most comparison operations can be skipped." The model then identified a text to be evaluated: "The best case time complexity of the algorithm is O(n), and most comparison operations can be skipped when the elements are basically ordered."

[0146] In the automobile repair fault diagnosis phase, the server needs to integrate at least one previously obtained text to be evaluated.

[0147] For the two previously mentioned texts to be evaluated regarding automobile fault diagnosis, "This car is difficult to start, and there is abnormal shaking when the engine is running. It is initially suspected that there is a problem with the spark plug because the spark plug has been used for a long time" and "Check the ignition coil, and its connecting line is not loose, which basically confirms that the spark plug is faulty", the server integrates them.

[0148] The server will integrate according to semantic logic. For example, it will arrange them in the order of the normal fault diagnosis process to obtain the integrated text to be evaluated "This car is difficult to start and there is abnormal shaking when the engine is running. It was initially suspected to be a problem with the spark plugs because they have been used for a long time. After checking the ignition coil, it was found that the connection line of the ignition coil was not loose, so it was basically determined to be a spark plug failure."

[0149] In the algorithm design phase of software programming, the server integrates the two previously obtained texts to be evaluated.

[0150] Integrate "designing an improved bubble sort algorithm, which is different from the traditional one and skips comparison when adjacent elements are already in order to improve efficiency" and "the best case time complexity of the algorithm is O(n), and most comparison operations can be skipped when the elements are basically in order."

[0151] According to the logical order described in the programming algorithm, the server integrates the text to be evaluated as "I designed an improved bubble sort algorithm. Unlike the traditional bubble sort, each time it compares adjacent elements, if the elements are already in order, the comparison is skipped, which can improve efficiency. In the best case, the time complexity of this algorithm can reach O(n), because when the elements are basically in order, most comparison operations can be skipped."

[0152] Through this integration process, a complete and semantically coherent text to be evaluated can be obtained in each target question link, so that it can be scored based on the standard answer text and other operations. This integration can more comprehensively and accurately reflect the performance of candidates in specific question links, avoiding information fragmentation caused by speech interruptions or explanations from different angles.

[0153] In the embodiment of the present invention, the text to be evaluated is scored based on the standard answer text to obtain a keyword score and a semantic similarity score, which can be implemented through the following examples.

[0154] Based on the standard answer text, a sliding window of the text to be evaluated corresponding to the target test question is formed;

[0155] Based on the size of the sliding window, semantic segmentation is performed on the text to be evaluated corresponding to the target test question link to obtain a plurality of short sentences to be evaluated;

[0156] Based on the standard answer text, each of the short sentences to be evaluated is scored in turn to obtain a keyword score and a semantic similarity score.

[0157] In the embodiment of the present invention, for example, it is assumed that in the construction skills assessment of a building project, the target test question link is the concrete pouring link.

[0158] The standard answer text is "Before pouring concrete, check the formwork and steel bars to ensure that the formwork is firmly installed and without deformation, and that the specifications, quantity and position of the steel bars meet the design requirements. Then, use appropriate pouring equipment, such as a concrete pump truck or hopper, to pour the concrete evenly into the formwork. During the pouring process, pay attention to vibration to make the concrete dense and avoid defects such as honeycombed surfaces. The pouring height should not be too high, and it should be poured in layers with the thickness of each layer controlled between 30-50 cm."

[0159] Based on the standard answer text, the server forms a sliding window for the text to be evaluated in this link. Assume that the size of the sliding window is set to 15 Chinese characters (this size can be set according to actual needs and language expression habits). This sliding window is like a specific perspective for observing the text to be evaluated, which is used for subsequent semantic segmentation and scoring operations.

[0160] For the text to be evaluated in which construction workers describe the operation process during the concrete pouring stage, it is assumed to be "I checked the formwork and steel bars first. The formwork was installed firmly, and the specifications, quantity and position of the steel bars were correct. Then I used a concrete pump truck to pour the concrete into the formwork. During pouring, I vibrated the concrete to make it dense. I poured it in layers, and each layer was about 40 cm thick."

[0161] The server performs semantic segmentation on the text to be evaluated according to the sliding window size (15 Chinese characters). The first short sentence to be evaluated is "I checked the formwork and steel bars first, and the formwork was installed firmly." This short sentence is just within the sliding window range and fully expresses an operation step. Then the short sentence corresponding to the next sliding window is "The specifications, quantity and location of the steel bars are correct. Then use a concrete pump truck to pour the concrete into the formwork." Although this spans part of the content of two sentences, a relatively complete short sentence is obtained according to the sliding window division. By analogy, continue to segment and obtain the short sentences waiting for evaluation, "Vibrate during pouring to make the concrete dense and pour in layers" and "The thickness of each layer is about 40 cm."

[0162] In the configuration of test points for the concrete pouring link of construction projects, the keywords include "formwork", "rebar", "concrete pump truck", "vibration", "layered pouring", etc. The keyword weights are set according to their importance. For example, the weights of "formwork" and "rebar" are 0.15 (because the preliminary inspection of formwork and rebar is an important prerequisite for pouring), the weight of "concrete pump truck" is 0.1 (as a pouring equipment, it is part of the pouring operation), the weight of "vibration" is 0.2 (vibration has a great influence on the density of concrete), and the weight of "layered pouring" is 0.2. The scoring strategy is set to get a score of the corresponding weight if the keyword is fully matched, a score of half the weight if it is partially matched, and 0 points if it is not matched.

[0163] The short sentence to be evaluated, "I checked the formwork and steel bars first, and the formwork was installed firmly," contains the keywords "formwork" and "steel bars," which are a perfect match, so the scores of these two keywords are 0.15 respectively. The score of the pending keywords of this short sentence is 0.15+0.15=0.3.

[0164] For the short sentence "The specifications, quantity and position of the steel bars are correct. Then use a concrete pump truck to pour the concrete into the formwork", the keywords "steel bars", "concrete pump truck" and "formwork" are all completely matched. The score of "steel bars" is 0.15, the score of "concrete pump truck" is 0.1, and the score of "formwork" is 0.15. The score of the pending keyword is 0.15+0.1+0.15=0.4.

[0165] The server uses a predefined semantic similarity algorithm, such as a cosine similarity algorithm, to calculate the semantic similarity between each short sentence to be evaluated and the corresponding part of the standard answer text.

[0166] For the short sentence to be evaluated, "I checked the formwork and steel bars first, and the formwork was installed firmly", the semantic similarity is calculated between it and the corresponding part of the standard answer text "Before pouring concrete, the formwork and steel bars should be checked to ensure that the formwork is installed firmly and without deformation, and the specifications, quantity and position of the steel bars meet the design requirements". Assume that the semantic similarity calculated by the cosine similarity algorithm is 0.85. If the semantic matching score threshold is set to 0.7, since 0.85>0.7, the semantic similarity score of this short sentence is 0.85.

[0167] For the short sentence "The thickness of each layer is about 40 cm", the semantic similarity calculation is performed with the corresponding part of the standard answer text "The pouring height should not be too high, it should be poured in layers, and the thickness of each layer should be controlled between 30-50 cm", and the semantic similarity is 0.9. Because 0.9>0.7 (semantic matching score threshold), the semantic similarity score of this short sentence is 0.9.

[0168] Through the above steps, the server scores each short sentence to be evaluated in turn based on the standard answer text, and obtains the keyword score and semantic similarity score of each short sentence. Then, the server can make a comprehensive evaluation of the performance of the entire text to be evaluated in the target test link, providing accurate data support for the final professional skills assessment.

[0169] In an embodiment of the present invention, the test point configuration also includes keywords, keyword weights and scoring strategies. The standard answer text is scored in turn with each of the short sentences to be evaluated to obtain a keyword score and a semantic similarity score, which can be implemented through the following examples.

[0170] Acquire a keyword scoring threshold and a semantic matching scoring threshold according to the scoring strategy;

[0171] Perform keyword matching on each of the short sentences to be evaluated according to the keywords and the keyword weights to obtain a pending keyword score;

[0172] In the case where the pending keyword score exceeds the keyword scoring threshold, taking the pending keyword score as the final keyword score;

[0173] Perform semantic matching on each of the short sentences to be evaluated according to the standard answer text to obtain a pending semantic similarity score;

[0174] In the case that the pending semantic similarity score exceeds the semantic matching score threshold, the pending semantic similarity score is used as the final semantic similarity score.

[0175] In an embodiment of the present invention, exemplarily, in the electrical engineering installation skill assessment, the target test link is the laying of electrical lines.

[0176] The server obtains information related to the scoring strategy from the test site configuration. For example, the scoring strategy specifies the basis for setting the keyword scoring threshold and the semantic matching scoring threshold. Assuming that in this scenario, the keyword scoring threshold is set to 0.3, which means that if the keyword matching score in a short sentence to be evaluated is lower than 0.3, the keyword matching degree is considered insufficient; the semantic matching scoring threshold is set to 0.6, that is, when the semantic similarity between the short sentence to be evaluated and the standard answer text is lower than 0.6, it indicates that the semantic matching degree is poor. These thresholds are determined based on multiple factors such as the requirements of the electrical engineering installation skills assessment, previous assessment experience, and expectations for the degree of skill mastery.

[0177] For the electrical line laying process, the keywords in the test point configuration include "electrical wires", "wire tubes", "junction boxes", "bridges", "grounding", etc. The keyword weights are set according to their importance in the laying of electrical lines. For example, the weight of "electrical wires" is 0.25 (because electric wires are the core element of line laying), the weight of "wire tubes" is 0.2 (wire tubes are used to protect electric wires), the weight of "junction boxes" is 0.15 (junction boxes play an important role in circuit connection and branching), the weight of "bridges" is 0.2 (bridges are used to carry and protect a large number of wires), and the weight of "grounding" is 0.2 (grounding is the key to electrical safety).

[0178] Suppose an electrician is describing the process of laying electrical lines in a short sentence to be evaluated: "I first pass the wires through the conduit, and then fix the conduit on the bridge."

[0179] For this short sentence, the server performs keyword matching. It contains the keywords "wire", "wire tube" and "bridge". The weight of "wire" is 0.25, which is a perfect match and gets 0.25 points; the weight of "wire tube" is 0.2, which is a perfect match and gets 0.2 points; the weight of "bridge" is 0.2, which is a perfect match and gets 0.2 points. Therefore, the pending keyword score of this short sentence is 0.25+0.2+0.2=0.65.

[0180] In the above example of electrical engineering installation, the pending keyword score is 0.65, and the keyword score threshold is 0.3. Because 0.65>0.3, the server uses 0.65 as the final keyword score of this short sentence to be evaluated.

[0181] The server uses a suitable semantic similarity algorithm, such as a cosine similarity algorithm based on word vectors, to perform semantic matching between each short sentence to be evaluated and the standard answer text.

[0182] For example, the standard answer text is "When laying electrical lines, first select wires of appropriate specifications based on design requirements, and pass the wires through pre-installed wire tubes. The wire tubes must be firmly fixed on the bridge, and ensure that the installation of the bridge complies with safety regulations. At the same time, pay attention to installing junction boxes where necessary to achieve circuit connection and branching. Good grounding must be ensured during the entire line laying process."

[0183] For the short sentence to be evaluated, "I first put the wires into the conduit, and then fixed the conduit on the bridge", the server calculates the semantic similarity between it and the corresponding part of the standard answer text. Through word vector conversion and cosine similarity calculation, the semantic similarity score to be determined is 0.8.

[0184] In this example of electrical engineering installation, the pending semantic similarity score is 0.8, and the semantic matching score threshold is 0.6. Because 0.8>0.6, the server uses 0.8 as the final semantic similarity score of this short sentence to be evaluated.

[0185] Through this process, the server accurately calculates the keyword score and semantic similarity score for each short sentence to be evaluated based on the keywords, keyword weights and scoring strategies in the test point configuration. This helps to comprehensively and objectively evaluate the candidate's skill level in the electrical line laying process and provides a reliable data basis for the entire electrical engineering installation skills assessment. In different professional skill assessment scenarios, you can follow these steps to accurately perform scoring operations based on the respective test point configurations and assessment requirements.

[0186] In the embodiment of the present invention, the construction of a sample data set for the target occupational skill assessment scenario may be implemented through the following examples.

[0187] Filtering voice data and text data related to the target professional skill assessment scenario from a plurality of different professional skill assessment history records;

[0188] Performing at least one data enhancement operation of adding noise, adjusting speech speed, and adjusting intonation on the filtered speech data;

[0189] The speech data after the data enhancement operation is combined with the corresponding text data to form the sample data set.

[0190] In the embodiment of the present invention, for example, it is assumed that the server is constructing a sample data set for an accounting professional skill assessment scenario.

[0191] The server stores a large number of historical records of different professional skill assessments, including assessment data from different industries and positions. Among these data, for the accounting professional skill assessment scenario, the server first needs to filter out relevant voice data and text data.

[0192] For example, among the numerous historical evaluation records, some are about engineer skill evaluation, and some are about chef skill evaluation. These data that are not related to accountant professional skill evaluation will be excluded. As for the accounting-related evaluation records, the server will carefully screen them.

[0193] In terms of voice data, it may be the voice records of accountants explaining financial statements, describing the accounting voucher processing process, and explaining tax calculations. For example, an accountant explains how to prepare a balance sheet: "First, determine the asset items, such as cash, accounts receivable, and inventory, then determine the liability items, such as short-term loans and accounts payable, and finally prepare the balance sheet according to the accounting equation."

[0194] The text data corresponds to these voices, accurately recording the accounting operation steps, principles, calculation methods, etc. For example: "Balance sheet preparation steps: first determine asset items, including cash, accounts receivable, inventory, etc.; then determine liability items, including short-term loans, accounts payable, etc.; prepare the balance sheet according to the accounting equation."

[0195] The server screens out nursing-related voice and text data from a large number of historical assessment records. Voice data may be a nurse's description of the patient care process, such as "When I was caring for this postoperative patient, I first checked his vital signs, including body temperature, blood pressure, heart rate, etc., and then checked whether the wound dressing was bleeding."

[0196] The corresponding text data is: "Postoperative patient care: first check vital signs (body temperature, blood pressure, heart rate, etc.), then check the bleeding of the wound dressing."

[0197] For the previously selected voice data of accountants preparing balance sheets, the server will add office environmental noise to simulate real office scenes. For example, add the sound of printers printing documents, the sound of colleagues discussing in low voices, etc. In this way, the originally clear voice "First, determine the asset items, such as cash, accounts receivable, and inventory, and then determine the liability items, such as short-term loans, accounts payable, etc., and finally prepare the balance sheet according to the accounting equation." becomes a voice with noisy background noise in the office, but the voice content remains unchanged.

[0198] The server can also adjust the speed of the accountant's voice data. It can speed up the normal speech to "First, determine the asset items such as cash, accounts receivable, and inventory, then determine the liability items such as short-term loans, accounts payable, etc. Finally, prepare the balance sheet according to the accounting equation." Or slow it down to "First, determine, um, asset items such as cash, accounts receivable, and inventory. Then, determine the liability items such as short-term loans, accounts payable, etc. Finally, prepare the balance sheet according to the accounting equation."

[0199] At the same time, adjust the tone of voice. For example, change the tone of plain narration to the tone of emphasizing key steps. When mentioning "prepare the balance sheet according to the accounting equation", raise the tone to highlight this key step.

[0200] For the voice data of nurses describing the patient care process, the server adds noises in the hospital environment, such as the beeping sound of medical instruments, footsteps outside the ward, etc. For example, "When I was caring for this postoperative patient, I first checked his vital signs, including body temperature, blood pressure, heart rate, etc., and then checked whether the wound dressing was bleeding." This voice will contain the noisy sounds of the hospital environment.

[0201] The normal speech speed of the nurse can be accelerated, such as "When I take care of the postoperative patient, I first check the vital signs, body temperature, blood pressure, heart rate, etc., and then check the wound dressing for bleeding." Or slow down to "When I take care of this postoperative patient, first, I check his vital signs, including body temperature, blood pressure, heart rate, etc. Then, I check the wound dressing to see if there is any bleeding."

[0202] When referring to “check the wound dressing for bleeding,” adjust your tone to a more cautious, serious one to reflect the importance of this step.

[0203] The server combines the voice data of the accountant preparing the balance sheet after adding noise, adjusting the speaking speed and adjusting the voice intonation with the corresponding text data "Balance sheet preparation steps: first determine the asset items, including cash, accounts receivable, inventory, etc.; then determine the liability items, including short-term loans, accounts payable, etc.; prepare the balance sheet according to the accounting equation."

[0204] Each set of voice data after data enhancement is combined with accurate text data one by one to form part of the sample data set. For example, the voice data of preparing balance sheets with office noise, fast speech speed, and emphasized intonation is combined with the corresponding text data. The voice-text data pairs after different enhancement methods are included in the sample data set, which will be used for subsequent model training, such as training the voice recognition model for accounting professional skills assessment.

[0205] Similarly, the server combines the voice data of nurses caring for patients, to which hospital environmental noise has been added, and to which different speech speeds and tones have been adjusted, with the corresponding text data "Postoperative patient care: first check vital signs (body temperature, blood pressure, heart rate, etc.), then check for bleeding from wound dressings."

[0206] Each enhanced voice data is paired with the corresponding text data and put into the sample data set. This sample data set will provide rich data resources for the training of the speech recognition model for nursing professional skills assessment, so that the model can better adapt to different speech situations and improve the accuracy of speech recognition in the nursing professional skills assessment scenario.

[0207] In order to more clearly describe the solution provided by the embodiment of the present invention, a more detailed implementation is provided below.

[0208] The present invention provides a voice input assessment method (device, equipment and medium) applied to occupational skill assessment scenarios. The method uses voice recognition technology and natural language semantic understanding technology to solve complex scenarios in occupational skill training and evaluation that are difficult to assess visually, such as the recognition of complex actions, whether tools are qualified and effective, and the description of operation processes. The implementation method of the present invention involves the fields of end-to-end voice recognition, natural language understanding, etc. It is divided into the following four stages:

[0209] (1) Test question configuration: that is, for the specific test points in the training and evaluation scenarios to be identified, the candidates should verbally express the content, organize the standard text, and clearly describe the content to be tested.

[0210] (2) Constructing a data set: Collecting voice data from different test takers in real assessment scenarios, and using standard text to find different people to record data, and augmenting these data to varying degrees. This part of the data is annotated and used as training data for the speech recognition model and the semantic understanding model.

[0211] (3) Speech recognition: SeACoParaformer, a customized non-autoregressive speech recognition model based on hot words, is adopted. On this basis, in order to improve the robustness and accuracy of the model, this solution adds a synonym processing module. Through deep optimization of the specific corpus for vocational skills training and evaluation scenarios, the model can more accurately capture the language features of the candidates' oral expressions and convert the language signals of the candidates' oral expressions into text information.

[0212] (4) Scoring stage: In order to make the assessment results more truly reflect the actual situation of vocational skills training and evaluation, two schemes, keyword matching and semantic similarity calculation, are implemented in parallel. This aims to accurately assess the candidates' mastery of the key points of the operational skills assessment, while effectively preventing candidates from responding to the assessment through mechanical memorization or reciting answers, thereby ensuring the fairness, flexibility and authenticity of the evaluation.

[0213] (5) Calculate the comprehensive score: The total score is obtained by calculating the configured speech recognition and semantic understanding weights, thresholds and other parameters and their respective scores.

[0214] Phase 1: Test point configuration. This phase is to configure the test questions and test points of the assessment scenario in a procedural manner, and configure the standard text of the answers that candidates need to state, as well as keywords and their weights, scoring method thresholds, etc. for the voice test points.

[0215] A test question is divided into multiple links in the process. The links are connected in series through conditional components. Multiple business components (vision and voice) within a link can be executed in parallel and multiple times within the link.

[0216] Configure the score weights for keyword matching and semantic understanding matching for each voice test point, denoted as W key and W simu , so that W key +W simu =1.

[0217] For the keyword matching assessment option, you also need to configure the keywords that match the text of each test point, as well as their proportion in the test point keyword and its synonym list.

[0218] Phase 2: Constructing the data set. In order to ensure the richness and balance of the data, three schemes are used for data collection: first, voice data from candidates from different regions and age groups in real test scenarios are collected; second, volunteers with different voice characteristics are invited to record voice data according to the test intent and standard text, and volunteers are allowed to add appropriate personal style and understanding to simulate the diversity in real tests; third, data enhancement is performed using data obtained by technical means, including but not limited to speech speed adjustment, volume change, noise addition, and echo simulation. After manually annotating the above data, the training data set and the test data set are divided into a ratio of 7:3, and these data will not be used later.

[0219] The third stage: speech recognition. The professional skills assessment scenario requires scoring the real-time voice stream of the examinee. Given the diversity of the examinee group in the assessment scenario, including a wide age span, various accent characteristics, and different speaking speeds, this greatly increases the difficulty of the model covering all situations. This solution selected the SeACoParaformer with hot word speech recognition model, which has excellent performance for Chinese ASR, but encountered problems such as misrecognition of homophones and loss of speech features due to truncation during use. In the future, it will be necessary to perform tasks such as semantic understanding. Therefore, based on the original architecture, a synonym enhancement module was added:

[0220] The hot word module in SeACoParaformer can capture the correlation between the Predictor output and the hot words in the Paraformer, and predict the hot word output synchronized with the ASR result. Hot word excitation is achieved through the fusion of posterior probabilities. This can reduce transcription errors with similar pronunciations. For example, in this scenario, "insulated shoes" is likely to be transcribed as "military shoes" in the case of unclear speech.

[0221] In order to improve the stability and reliability of the model under different environments, different speakers, different speech speeds and intonations, and enhance the model's adaptability to speech diversity, this solution adds a synonym enhancement module to the SeACoParaformer architecture. This module is mainly composed of Encoder and Decoder. The input of the module is the synonym dictionary and the text optimized by the hot word module, and the output is the text sequence after synonym enhancement. The specific processing process is as follows:

[0222] Synonym embedding retrieval: For each word w after the input text is segmented i , find its synonym set S through the synonym dictionary i . i and each synonym s in the set j ∈S i , use the pre-trained embedding table to retrieve its corresponding embedding vector e i and e sj , and calculate their respective position codes p i and p sj .

[0223] Context fusion: To select and fuse synonym embeddings, an attention mechanism is used to realize the influence of context, where the attention weight α ij It is calculated dynamically based on the context. First, the weight score is calculated:

[0224]

[0225] where d k is the dimension of the embedding vector, which is then converted into a probability distribution through the softmax function to obtain the attention weight:

[0226]

[0227] Word embedding fusion: Use the calculated attention weights to perform weighted averaging on synonym embeddings to obtain the fused synonym embedding representation:

[0228]

[0229] Enhanced vector generation: embedding the original word into i (Its weight α i ) is combined with the synonym embedding representation obtained through the above context fusion process (or called fused synonym embedding) to generate the final enhanced embedding vector e' i .

[0230] e′ i =α i ·e i +(1-α i)·fused_embed i

[0231] So far, each word w i There is a corresponding enhancement vector e' i , this vector not only contains the information of the word itself, but also integrates the semantic information of related synonyms.

[0232] When decoding, the Decoder uses the cross-attension mechanism to combine the enhanced embedding vector of the Encoder, which is used to calculate the similarity between the Decoder state and the Encoder enhanced embedding vector. Suppose the current state of the Decoder is s t Converted into query vector q t , the cross-attention output is expressed as:

[0233]

[0234] Where K', V' are the encoder enhanced embedding vectors e' i The key and value matrix representation of .

[0235] This cross-attention output is combined with the self-attention output to generate the final decoder state representation, which contains both the self-attention information within the Decoder and the enhanced embedding vector information of the Encoder.

[0236] Vocabulary selection: Based on the current decoding state (combining self-attention and cross-attention information), the Decoder selects the most appropriate word or synonym from the vocabulary for output. This layer maps the decoder state to a score vector of the vocabulary size. Then, the score is converted to a probability distribution using the softmax function, and the word with the highest probability is selected as the output:

[0237] p(y t |y 1:t-1 )=softmax(Ws t +b)

[0238] Where W is the weight matrix of the output layer, b is the bias term, and s t is the current decoder state.

[0239] In the process of real-time speech transcription, incorrect truncation problems are often encountered due to the complexity of the environment. For example, when there is silence, due to the high noise, the silence period is not detected, and forced truncation will occur after a certain length of time. This will cause a complete sentence to be segmented into multiple speech slices, thus affecting the recognition accuracy. As shown in the example below, "The insulation ladder has a good appearance and no cracks within the qualified validity period" is a complete test point, but it is segmented into two segments during recognition.

[0240] In Phase 1, it is clear that a set of assessment questions will be divided into multiple assessment links. At the same time, in multiple test points within a link, the order of the candidates' operations can be disrupted, and parallel computing processes need to be designed to improve reasoning efficiency. In order to solve these problems, a new customized solution called link paragraph is provided. The two innovative points of the solution are as follows:

[0241] In one session, the user's answers to multiple test points are accumulated and spliced ​​into a paragraph. The paragraph consists of many short sentences, which are separated by punctuation marks transcribed by voice recognition to avoid semantic segmentation anomalies caused by unreasonable truncation. As shown in the following figure, multiple short sentences in a row represent the slices transcribed by voice recognition, and short sentences of the same color represent the text transcribed from a logically coherent test point.

[0242] The previous scoring scheme was to compare the real-time transcription of the speech with the standard text of the test site to get the score. The new scheme reverses this process and uses the standard text of the current test site to match the best answer in the link paragraph.

[0243] The fourth stage: the scoring stage. The speech transcription text output from the third stage is accumulated in a large link to form a paragraph, and then the standard text answer of the test point is used to select a suitable sliding window win to find the most matching answer from the paragraph. The size of the sliding window is the length of the list after semantic segmentation of the answer; starting from the latest answer in the paragraph, 1 to win short sentences are spliced ​​forward, and they are matched and scored with the answers in turn. The specific scoring rules are as follows: First, configure the test point keyword score and the semantic matching score threshold t key and t simu , the threshold is used to determine whether the score is valid.

[0244] Starting from the latest short sentence in the paragraph, loop forward to select 1 to win short sentence combinations, and perform subsequent steps on the selected short sentence combinations.

[0245] Keyword matching score: The keyword KW configured in the test point i , respectively match the selected phrase combination, and calculate the total score of keyword matching:

[0246]

[0247] Where s is the total score of the changed test center configuration, w i is the weight of a keyword. key <t key , then s key =0, indicating that the test point fails to match the keyword and no points are awarded.

[0248] Semantic similarity matching score: match the standard text in the test point with the selected short sentence combination to obtain the semantic matching score s simu , if s simu <t simu , then s simu =0, indicating that the test point fails to match the semantics and no points are awarded.

[0249] Phase 5: Calculate the comprehensive score. Calculate the overall score of the test point based on the configured keyword matching score and semantic understanding score. You can also configure corresponding weight coefficients for the keyword matching score and semantic understanding score respectively.

[0250] s total =s key +s simu

[0251] Please refer to Figure 2 , Figure 2 A speech input evaluation device 110 for professional skill evaluation provided in an embodiment of the present invention includes:

[0252] The acquisition module 1101 is used to obtain the test site configuration data of the target professional skill assessment scenario, wherein the test site configuration data includes the standard answer text; construct a sample data set for the target professional skill assessment scenario, and train the enhanced SeACoParaformer architecture based on the sample data set to obtain a trained professional skill assessment speech recognition model; input the to-be-assessed speech stream into the professional skill assessment speech recognition model for recognition, and obtain the to-be-assessed text corresponding to the to-be-assessed speech stream;

[0253] The evaluation module 1102 is used to score the text to be evaluated based on the standard answer text to obtain a keyword score and a semantic similarity score; and use the keyword score and the semantic similarity score as the professional skill evaluation score corresponding to the voice stream to be evaluated.

[0254] It should be noted that the implementation principle of the aforementioned voice input evaluation device 110 for vocational skill assessment can refer to the implementation principle of the aforementioned voice input evaluation method for vocational skill assessment, which will not be repeated here. It should be understood that the division of the various modules of the above device is only a division of logical functions. In actual implementation, they can be fully or partially integrated into a physical entity, or physically separated. And these modules can all be implemented in the form of software called by processing elements; they can also be all implemented in the form of hardware; some modules can also be implemented in the form of software called by processing elements, and some modules can be implemented in the form of hardware. For example, the voice input evaluation device 110 for vocational skill assessment can be a separately established processing element, or it can be integrated in a chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and called and executed by a processing element of the above device. The function of the voice input evaluation device 110 for vocational skill assessment. The implementation of other modules is similar. In addition, all or part of these modules can be integrated together, or they can be implemented independently. The processing element described here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in a processor element or an instruction in software form.

[0255] For example, the above modules may be one or more integrated circuits configured to implement the above methods, such as one or more application specific integrated circuits (ASIC), or one or more microprocessors (digital signal processors, DSP), or one or more field programmable gate arrays (FPGA), etc. For another example, when a module above is implemented in the form of a processing element scheduling program code, the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processor that can call program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0256] The embodiment of the present invention provides a computer device 100, which includes a processor and a non-volatile memory storing computer instructions. When the computer instructions are executed by the processor, the computer device 100 executes the aforementioned voice input evaluation device 110 for professional skill evaluation. Figure 3 As shown, Figure 3The computer device 100 is a block diagram of a structure of a computer device 100 provided in an embodiment of the present invention. The computer device 100 includes a speech input evaluation device 110 for professional skill evaluation, a memory 111, a processor 112 and a communication unit 113.

[0257] To achieve data transmission or interaction, the memory 111, the processor 112 and the communication unit 113 are electrically connected to each other directly or indirectly. For example, these components can be electrically connected to each other through one or more communication buses or signal lines. The voice input evaluation device 110 for professional skill assessment includes at least one software function module that can be stored in the memory 111 in the form of software or firmware or solidified in the operating system (OS) of the computer device 100. The processor 112 is used to execute the voice input evaluation device 110 for professional skill assessment stored in the memory 111, such as the software function modules and computer programs included in the voice input evaluation device 110 for professional skill assessment.

[0258] An embodiment of the present invention provides a readable storage medium, which includes a computer program. When the computer program is running, it controls the computer device where the readable storage medium is located to execute the aforementioned voice input evaluation device 110 for professional skill evaluation.

[0259] For illustrative purposes, the foregoing description is made with reference to specific embodiments. However, the above illustrative discussion is not intended to be exhaustive or to limit the present disclosure to the precise form disclosed. Numerous modifications and variations are possible in accordance with the above teachings. These embodiments are selected and described in order to best illustrate the principles of the present disclosure and its practical application, so that those skilled in the art can best utilize the present disclosure and utilize various embodiments with different modifications to suit the intended specific application.

Claims

1. A voice input evaluation method for professional skill assessment, characterized in that: include: Acquire test point configuration data of a target professional skill assessment scenario, wherein the test point configuration data includes a standard answer text; Constructing a sample data set for the target occupational skill assessment scenario, and training the enhanced SeACoParaformer architecture based on the sample data set to obtain a trained occupational skill assessment speech recognition model; Inputting the speech stream to be evaluated into the occupational skill evaluation speech recognition model for recognition, and obtaining the text to be evaluated corresponding to the speech stream to be evaluated; Scoring the text to be evaluated based on the standard answer text to obtain a keyword score and a semantic similarity score; The keyword score and the semantic similarity score are used as the professional skill evaluation score corresponding to the speech stream to be evaluated.

2. The method according to claim 1, characterized in that The enhanced SeACoParaformer architecture includes a hot word component and a synonym enhancement component. The inputting of the speech stream to be evaluated into the occupational skill evaluation speech recognition model for recognition to obtain the text to be evaluated corresponding to the speech stream to be evaluated includes: The speech stream to be evaluated is optimized by the hot word component to obtain a text to be processed and input into the synonym enhancement component; Performing synonym enhancement on the text to be processed based on the synonym enhancement component to obtain an enhanced text sequence; The text sequence is used as the text to be evaluated corresponding to the speech stream to be evaluated.

3. The method according to claim 2, characterized in that The process of performing synonym enhancement on the text to be processed based on the synonym enhancement component to obtain an enhanced text sequence includes: Performing word segmentation processing on the text to be processed to obtain multiple word segments, and using a synonym dictionary to determine a synonym set for each word segment; using a pre-trained embedding table to retrieve a word segmentation embedding vector and a synonym set embedding vector corresponding to each word segment and each synonym set, and determining the position codes corresponding to each of the word segmentation embedding vector and the synonym set embedding vector, so as to realize synonym embedding retrieval; An attention mechanism is used to combine the word segmentation embedding vector, the synonym set embedding vector, and the position encodings corresponding to the word segmentation embedding vector and the synonym set embedding vector to determine an attention weight to achieve context fusion; Based on the attention weights, weighted average is performed on the synonym set embedding vectors to obtain a fused synonym set embedding vector, thereby achieving word embedding fusion; According to the word segmentation embedding vector, the word weight pre-configured by the word segmentation embedding vector, and the synonym set embedding vector, an enhanced word segmentation embedding vector of each word segment is calculated to achieve enhanced vector generation; Decoding the enhanced word segmentation embedding vector of each word segment in combination with the cross-attention mechanism to obtain a decoding state of the enhanced word segmentation embedding vector of each word segment, wherein the decoding state combines the self-attention mechanism and the cross-self-attention mechanism; Mapping each of the decoding states to a vocabulary to obtain a score vector corresponding to the vocabulary for each of the word segments; transforming the score vector using a softmax function to obtain a probability distribution of each word segment corresponding to the vocabulary, and outputting an output word segment of each word segment included in the text to be processed based on the probability distribution to achieve vocabulary selection; The output word segmentations are integrated to obtain the enhanced text sequence.

4. The method according to claim 1, characterized in that: The test point configuration data includes a plurality of test question links, and the inputting of the to-be-evaluated voice stream into the occupational skill assessment voice recognition model for recognition to obtain the to-be-evaluated text corresponding to the to-be-evaluated voice stream includes: In the target test question phase, the speech stream to be evaluated is input into the occupational skill evaluation speech recognition model for recognition, and at least one text to be evaluated corresponding to the speech stream to be evaluated is obtained; The at least one text to be evaluated is integrated to obtain a text to be evaluated corresponding to the target test question segment.

5. The method according to claim 4, characterized in that Scoring the text to be evaluated based on the standard answer text to obtain a keyword score and a semantic similarity score, including: Based on the standard answer text, a sliding window of the text to be evaluated corresponding to the target test question is formed; Based on the size of the sliding window, semantic segmentation is performed on the text to be evaluated corresponding to the target test question link to obtain a plurality of short sentences to be evaluated; Based on the standard answer text, each of the short sentences to be evaluated is scored in turn to obtain a keyword score and a semantic similarity score.

6. The method according to claim 5, characterized in that The test point configuration also includes keywords, keyword weights and scoring strategies. The standard answer text is scored with each of the short sentences to be evaluated in turn to obtain a keyword score and a semantic similarity score, including: Acquire a keyword scoring threshold and a semantic matching scoring threshold according to the scoring strategy; Perform keyword matching on each of the short sentences to be evaluated according to the keywords and the keyword weights to obtain a pending keyword score; In the case where the pending keyword score exceeds the keyword scoring threshold, taking the pending keyword score as the final keyword score; Perform semantic matching on each of the short sentences to be evaluated according to the standard answer text to obtain a pending semantic similarity score; In the case that the pending semantic similarity score exceeds the semantic matching score threshold, the pending semantic similarity score is used as the final semantic similarity score.

7. The method according to claim 1, characterized in that The constructing of a sample data set for the target occupational skill assessment scenario includes: Filtering voice data and text data related to the target professional skill assessment scenario from a plurality of different professional skill assessment history records; Performing at least one data enhancement operation of adding noise, adjusting speech speed, and adjusting intonation on the filtered speech data; The speech data after the data enhancement operation is combined with the corresponding text data to form the sample data set.

8. A speech input evaluation device for vocational skill evaluation, characterized in that: include: An acquisition module is used to acquire test site configuration data of a target occupational skill assessment scenario, wherein the test site configuration data includes a standard answer text; construct a sample data set for the target occupational skill assessment scenario, and train the enhanced SeACoParaformer architecture based on the sample data set to obtain a trained occupational skill assessment speech recognition model; input the to-be-assessed speech stream into the occupational skill assessment speech recognition model for recognition, and obtain the to-be-assessed text corresponding to the to-be-assessed speech stream; An evaluation module is used to score the text to be evaluated based on the standard answer text to obtain a keyword score and a semantic similarity score; and use the keyword score and the semantic similarity score as the professional skill evaluation score corresponding to the voice stream to be evaluated.

9. A computer device, characterized in that: The computer device comprises a processor and a non-volatile memory storing computer instructions. When the computer instructions are executed by the processor, the computer device executes the method according to any one of claims 1 to 7.

10. A readable storage medium, characterized in that: The readable storage medium includes a computer program, and when the computer program is executed, the computer device where the readable storage medium is located is controlled to execute the method according to any one of claims 1 to 7.

Images