Occupational interest prediction method and device, computer equipment and storage medium

Abstract

The invention relates to the technical field of artificial intelligence, and discloses an occupational interest prediction method and device, computer equipment and a storage medium. Comprising the steps of obtaining a test text of a testee describing a test image; inputting the test image and the test text into a prediction model to enable the prediction model to determine prediction scores of the testee in different psychological dimensions based on the test text, and outputting occupational interest information matched with the prediction scores; wherein the prediction model is obtained by carrying out fine tuning training on the image-text large model, and the occupational interest information is used for representing the interest degree of the testee in different preset occupational categories. According to the method, the potential occupational interest of the user can be carefully captured, accurate intelligent prediction of the occupational interest is realized, and contusion caused by blind selection of the user is reduced.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method, apparatus, computer device and storage medium for predicting career interests. Background Art

[0002] With the rapid development of the Internet and information technology, a large amount of career information and personal job search data have been digitized. However, existing career assessments usually use fixed questions to evaluate users through answers. However, this way of answering questions may make it difficult for users to engage, and the assessment process becomes repetitive and boring, affecting the accuracy of the results. Summary of the invention

[0003] In view of this, the present application provides a career interest prediction method, apparatus, computer device and storage medium to solve the technical problem that the prediction of career interests is inaccurate and boring.

[0004] In a first aspect, according to one aspect of the present application, a method for predicting career interests is provided, comprising:

[0005] Acquire a test text describing a test image from a testee, wherein the test image is used to test the psychological characteristics of the testee;

[0006] The test image and the test text are input into a prediction model so that the prediction model determines the predicted scores of the subject in different psychological dimensions based on the test text, and outputs career interest information matching the predicted scores; wherein the prediction model is obtained by fine-tuning the large image and text model, and the career interest information is used to characterize the subject's interest level in different preset career categories.

[0007] Optionally, before inputting the test image and the test text into the prediction model, the career interest prediction method further includes:

[0008] Acquire training samples and psychological score labels and occupational category labels corresponding to the training samples in different psychological dimensions, wherein the training samples include sample images and sample texts describing the sample images;

[0009] Acquire training samples and psychological score labels and occupational category labels corresponding to the training samples in different psychological dimensions, wherein the training samples include sample images and sample texts describing the sample images;

[0010] Based on the sample image, the sample text, the psychological score label and the occupational category label, fine-tuning the image-text large model to obtain a prediction model;

[0011] Among them, the preset occupational categories include at least one of the following: scientific research, assembly and repair, education and training, consulting services, engineering construction, medical services, and finance.

[0012] Optionally, before inputting the test image and the test text into the prediction model, the career interest prediction method further includes:

[0013] Inputting the test text into the quality assessment model to obtain an assessment result, wherein the quality assessment model is obtained by fine-tuning the image-text model through training samples and quality scoring rules corresponding to the training samples;

[0014] If the evaluation result of the test text is unqualified, outputting a first prompt message based on the reason for the unqualified result;

[0015] Inputting the test image and the test text into a prediction model comprises:

[0016] If the evaluation result of the test text is qualified, it is determined that the test text meets the quality requirements, and the test image and the test text that meets the quality requirements are input into the prediction model.

[0017] Optionally, the training sample includes a sample image and a sample text describing the sample image, and before inputting the test text into the quality assessment model to obtain the assessment result, the career interest prediction method further includes:

[0018] Performing word segmentation processing on the sample text to determine multiple word segments of the sample text;

[0019] Extracting text features of the multiple word segments, wherein the text features include at least one of the following: text length, word repetition frequency, number of punctuation marks, character repetition, and word occurrence frequency;

[0020] Evaluate the text features of multiple word segments based on the preset scoring rules to determine the quality label of the sample text;

[0021] Based on the sample images, the sample texts and the quality labels, the image-text large model is fine-tuned and trained to obtain the quality assessment model.

[0022] Optionally, after the test image and the test text are input into the prediction model, the career interest prediction method further includes:

[0023] Visit a designated academic website to determine the educational information of the test subject;

[0024] If the educational background information matches the identity information of the subject, obtaining the historical employment evaluation data and user portrait data of the subject;

[0025] Based on the historical employment assessment data, the user portrait data and the career interest information of the subject, career recommendations are made for the subject.

[0026] Optionally, obtaining a test text in which the test subject describes the test image includes:

[0027] In response to an input operation, obtaining the test text; or,

[0028] The voice information is received, and recognition processing is performed on the voice information to obtain the test text.

[0029] Optionally, before obtaining the test text describing the test image from the testee, the career interest prediction method further includes:

[0030] Matching a first image and a second image based on the identity information of the subject, wherein the first image is used to test the psychological characteristics of the subject, and the second image is used to test the credibility of the subject;

[0031] combining the first image and the second image to generate the test image;

[0032] Inputting the test image and the test text into a prediction model comprises:

[0033] Inputting the test image and the test text into the prediction model, so that the prediction model determines the predicted scores of the test subject in different psychological dimensions and the credibility of the test subject based on the test text, and outputs the career interest information matching the predicted scores and the credibility;

[0034] After the test image and the test text are input into the prediction model, the career interest prediction method further includes:

[0035] If the credibility is less than or equal to the credibility threshold, a second prompt message is output.

[0036] In a second aspect, a career interest prediction device is provided, comprising:

[0037] An acquisition module, used for acquiring a test text describing a test image from a testee, wherein the test image is used for testing the psychological characteristics of the testee;

[0038] A prediction module is used to input the test image and the test text into a prediction model so that the prediction model determines the predicted scores of the subject in different psychological dimensions based on the test text, and outputs career interest information matching the predicted scores; wherein the prediction model is obtained by fine-tuning the large image and text model, and the career interest information is used to characterize the subject's interest level in different preset career categories.

[0039] Optionally, the acquisition module is further used to acquire training samples and psychological score labels and occupational category labels corresponding to the training samples in different psychological dimensions, wherein the training samples include sample images and sample texts describing the sample images;

[0040] The occupational interest prediction device also includes:

[0041] A first training module is used to fine-tune the image-text model based on the sample image, the sample text, the psychological score label and the occupation category label to obtain the prediction model;

[0042] Among them, the preset occupational categories include at least one of the following: scientific research, assembly and repair, education and training, consulting services, engineering construction, medical services, and finance.

[0043] Optionally, the occupational interest prediction device further includes:

[0044] A detection module, used for inputting the test text into the quality assessment model to obtain an assessment result, wherein the quality assessment model is obtained by fine-tuning the image-text model through training samples and quality scoring rules corresponding to the training samples;

[0045] A first prompt module, configured to output a first prompt message based on the reason for failure if the evaluation result of the test text is failure;

[0046] The detection module is further configured to determine that the test text meets the quality requirements if the evaluation result of the test text is qualified;

[0047] The prediction module is specifically used to determine the predicted scores of the subject in different psychological dimensions and the career interest information based on the test image and the test text that meets the quality requirements.

[0048] Optionally, the training sample includes a sample image and a sample text describing the sample image; and the occupation interest prediction device further includes:

[0049] A word segmentation module, used to perform word segmentation processing on the sample text and determine multiple word segments of the sample text;

[0050] A feature extraction module, used to extract text features of the plurality of segmented words, wherein the text features include at least one of the following: text length, word repetition frequency, number of punctuation marks, character repetition degree, and word occurrence frequency;

[0051] An evaluation module, used to evaluate the text features of multiple word segments based on the preset scoring rules to determine the quality label of the sample text;

[0052] The second training module is used to fine-tune the image-text large model based on the sample image, the sample text and the quality label to obtain the quality assessment model.

[0053] Optionally, the occupational interest prediction device further includes:

[0054] An identity verification module is used to access a designated academic website to determine the educational background information of the subject;

[0055] The acquisition module is further used to acquire the historical employment evaluation data and user portrait data of the subject if the educational background information matches the identity information of the subject;

[0056] The recommendation module is used to make career recommendations for the subject based on historical employment assessment data, the user portrait data and the subject's career interest information.

[0057] Optionally, the acquisition module is specifically used to acquire the test text in response to an input operation; or to receive voice information, and perform recognition processing on the voice information to obtain the test text.

[0058] Optionally, the occupational interest prediction device further includes:

[0059] A test configuration module, configured to match a first image and a second image based on the identity information of the subject, wherein the first image is used to test the psychological characteristics of the subject and the second image is used to test the credibility of the subject; and to combine the first image and the second image to generate the test image;

[0060] The prediction module is specifically used to input the test image and the test text into the prediction model, so that the prediction model determines the predicted scores of the test subject in different psychological dimensions and the credibility of the test subject based on the test text, and outputs the career interest information matching the predicted scores and the credibility;

[0061] The second prompt module is used to output second prompt information if the credibility is less than or equal to the credibility threshold.

[0062] In a third aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned career interest prediction method when executing the computer program.

[0063] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned career interest prediction method are implemented.

[0064] In the scheme implemented by the above-mentioned career interest prediction method, device, computer equipment and storage medium, the principle of psychological projection test is adopted. Through the intelligent analysis of test pictures and complex texts for testing psychological characteristics by the prediction model based on the large picture and text model, the test subject's situation in multiple specific psychological dimensions such as emotional stability, stress resistance, and concentration is understood in detail, and the career interest information is dynamically matched through the predicted scores of the test subject in different psychological dimensions. In this way, the user's potential career interests can be captured in detail, making the career interest assessment results more accurate and flexible, and realizing accurate intelligent prediction of career interests. In this way, users can more clearly understand which career fields match their own personality, preferences, and interests, thereby reducing the frustration caused by blind choices.

[0065] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0067] Figure 1 This is a schematic diagram of an application environment of a method for predicting occupational interests in an embodiment of the present application;

[0068] Figure 2 is a flow chart of a method for predicting occupational interests in one embodiment of the present application;

[0069] Figure 3 is a structural diagram of a career interest prediction device in one embodiment of the present application;

[0070] Figure 4 is a structural schematic diagram of a computer device in one embodiment of the present application;

[0071] Figure 5 It is another structural schematic diagram of a computer device in one embodiment of the present application. DETAILED DESCRIPTION

[0072] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other without conflict.

[0073] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be interpreted as limiting the present application.

[0074] Those skilled in the art will appreciate that, unless expressly stated, the singular forms "a", "an", "said" and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present application refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, or there may be intermediate elements. In addition, the "connection" or "connection" used herein may include wireless connection or wireless fusion. The term "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.

[0075] Now, exemplary embodiments according to the present application will be described in more detail with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in a variety of different forms and should not be interpreted as being limited to the embodiments set forth herein. It should be understood that these embodiments are provided to make the disclosure of the present application thorough and complete, and to fully convey the concepts of these exemplary embodiments to those of ordinary skill in the art.

[0076] The career interest prediction method provided in the embodiment of the present application can be applied to Figure 1 In an application environment, the client communicates with the server through a network. The server can obtain the test text of the test subject describing the test image through the client, and input the test image and the test text into the prediction model to obtain the predicted scores of the test subject in different psychological dimensions, as well as the career interest information matching the predicted scores; finally, the test subject is responded to and the career interest information is fed back to the client. In this application, the user's potential career interests can be captured in detail, making the career interest assessment results more accurate and flexible, and realizing accurate intelligent prediction of career interests. In this way, users can more clearly understand which career fields match their own personality, preferences, and interests, thereby reducing the frustration caused by blind choices.

[0077] The client may be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, and portable wearable devices. The server may be implemented by an independent server or a server cluster consisting of multiple servers. The present invention is described in detail below through specific embodiments.

[0078] In this embodiment, a method for predicting career interests is provided. Figure 2 As shown, the method includes:

[0079] S10, obtaining a test text in which the subject describes the test image.

[0080] Specifically, the test image is used to test the psychological characteristics of the subject, and the test image can be drawn by a professional assessment expert based on experience, and the embodiment of the present application does not make any specific limitation.

[0081] In this embodiment, a test image is randomly retrieved from the career interest assessment library and displayed to the subject, who is required to describe the content displayed by the test image so that the psychological characteristics of the subject can be analyzed through the test text described by the subject.

[0082] The career interest prediction method provided by the present invention can be applied to artificial intelligence engines in various application scenarios. The artificial intelligence engine is usually implemented through a server, which can receive user requests for career interest prediction in real time.

[0083] In an actual application scenario, step S10, i.e. obtaining a test text of the test subject describing the test image, may be performed in the following manner:

[0084] Method 1: Responding to an input operation, obtaining a test text.

[0085] In this embodiment, a content input module is displayed on the interface, and the subject can fill in the test text by inputting operations on the content input module.

[0086] Method 2: receiving voice information, performing recognition processing on the voice information, and obtaining a test text.

[0087] In this embodiment, the speech-to-text technology is used to provide users with a convenient text input method, which reduces the difficulty of obtaining test text when there may be writing disabilities, limited literacy or inconvenient hand operation, and helps to improve the operating efficiency of the entire test process.

[0088] For example, a career assessment expert selects multiple pictures to construct a career interest assessment gallery. When the subject starts the career assessment, a picture is randomly selected from the constructed career interest assessment gallery and displayed to the user. The user is also presented with a text, such as "Please describe the scene in the picture, what happened before this scene, and in what direction it will develop. Use your imagination and try to describe a complete story." The subject describes the picture and obtains the text content of the user's description of the picture. In the process of obtaining the user's description content, the user can directly enter the text content, or collect the user's voice description and then convert the voice into text to obtain the text content.

[0089] S20, inputting the test image and the test text into the prediction model, so that the prediction model determines the predicted scores of the testee in different psychological dimensions based on the test text, and outputs the career interest information matching the predicted scores.

[0090] Among them, the occupational interest information is used to characterize the subject's interest in different preset occupational categories. The psychological dimension can be set according to the psychological classification of psychology, for example, nervousness, rationality, depression, irritability, activity, respect for life, etc., which will not be listed one by one in the embodiments of this application. The preset occupational categories include at least one of the following: scientific research, assembly and repair, education and training, consulting services, engineering construction, medical services, and finance. This application does not specifically limit the preset occupational categories, which can be reasonably set according to the occupational classification.

[0091] It can be seen that in the above scheme, the principle of psychological projection test is adopted. Through the intelligent analysis of test pictures and complex texts of test psychological characteristics by the prediction model based on the large picture and text model, the test subject's situation in multiple specific psychological dimensions such as emotional stability, stress resistance, and concentration is understood in detail, and the career interest information is dynamically matched through the predicted scores of the test subject in different psychological dimensions. On the one hand, it can capture the user's potential career interests in detail, making the career interest assessment results more accurate and flexible, and realizing accurate intelligent prediction of career interests. On the other hand, by analyzing the user's psychological characteristics and predicting the user's career interests through the model, providing data-driven decision-making basis, psychological assessment and career interest screening can be completed in a short time, greatly improving the efficiency of career interest assessment. In this way, users can more conveniently and clearly understand which career fields match their personality, preferences, and interests, thereby reducing the frustration caused by blind selection. In addition, the evaluation form based on psychological projection is interesting. Users do not need to face traditional rigid questions, but express themselves easily and naturally through a more interactive test process, which improves the user's sense of participation and helps to enhance the evaluation experience.

[0092] For example, when the test text analysis shows "active thinking 9", "conservative 2", and "resistance to frustration 8", it means that the subject is more interested in jobs that may require high pressure and high responsiveness, and the career interest information can be determined as design interest level 10, finance interest level 9, etc. When the test text analysis shows "respect for life 7", "conservative 7", and "depression 3", it means that the subject is more interested in jobs that may require a sense of morality and responsibility, and the career interest information can be determined as medical service interest level 9, etc.

[0093] It is understandable that the system can match the subject's career interest information through the predicted scores and sort the preset career categories according to the degree of interest, so that the subject can intuitively perceive the predicted results.

[0094] Among them, the prediction model is obtained by fine-tuning the large picture and text model. In one embodiment, before step S20, the following steps are also included: obtaining training samples and psychological scoring labels and occupational category labels corresponding to the training samples in different psychological dimensions; based on sample images, sample texts, psychological scoring labels and occupational category labels, fine-tuning the large picture and text model to obtain the prediction model.

[0095] Among them, the training samples include sample images and sample texts describing the sample images. The sample text contains the subject's unique understanding and feedback of the test image. The training samples, psychological scoring labels and occupational category labels can be historical data. For example, the test image used in the user's previous career interest assessment is used as the sample image, and the user's description text of the sample image is used as the sample text. Professionals analyze the sample text and obtain different psychological dimension scores as the psychological scoring labels corresponding to the samples in different psychological dimensions. The final career interest results obtained after analysis are used as the occupational category labels. Specifically, the Vision-Language Models (VLMs) is a multimodal input model with visual perception and natural language understanding capabilities. The input of the Vision-Language Model is image and text, and the output is text. The Vision-Language Model includes an encoder for images, an encoder for text, encoding information that integrates images and text, and subsequent network structures for specific tasks. Exemplarily, the large image-text model can be a CLIP (Contrastive Language-Image Pretraining) model, a BLIP (Bootstrapping Language-Image Pre-training) model, a KOSMOS model, a LLaVA model, a Qwen-VL model, etc.

[0096] Taking the CLIP (Contrastive Language-Image Pretraining) model as the basic model, the sample text is cleaned to remove irrelevant information such as special characters, punctuation marks, and HTML tags (if any). For example, replace HTML space characters such as "" in the text with ordinary spaces to make the text purer. Load the weights of the pre-trained CLIP large model, freeze some parameters of the CLIP model according to the data characteristics of psychological assessment and interest matching tasks, and only fine-tune specific layers or modules related to the task. For psychological score labels, add one or more fully connected layers as output layers on the basis of the CLIP model to predict psychological scores. The number of neurons in the output layer should correspond to the dimension of the psychological score. The loss function can be the mean square error (MSE) loss function. By minimizing this loss function, the psychological score predicted by the model is as close to the true label as possible. For occupational category labels, regard occupational category classification as a multi-classification problem, and add a classifier to the output layer of the CLIP model, such as using the Softmax function to convert the model output into an occupational category probability distribution. The corresponding loss function can use the cross-entropy loss function (Cross-Entropy Loss) to encourage the model to correctly predict the occupational category. During the fine-tuning training process, the preprocessed sample images, sample texts, psychological score labels, and occupational category labels are combined into small batches of data. Each time a certain number of samples are input into the model for forward propagation, and the losses on the two tasks of psychological score prediction and occupational category classification are calculated. Based on the calculated loss, the gradient of each parameter is calculated through the back-propagation algorithm, and the optimization algorithm (such as Adam, Adagrad, etc.) is used to update the parameters of the model so that the model gradually converges to the optimal state.

[0097] In this embodiment, by introducing psychological scoring labels and occupational category labels related to the training samples to fine-tune the large picture and text model, the general language and image feature representation capabilities learned by the large picture and text model in the early stage can be used to further allow the large picture and text model to focus on learning more subtle and complex features and patterns between different training samples and key features and occupational interests in different psychological dimensions, and obtain a prediction model. So that the current test image and test text can be analyzed through the prediction model, and the user's predicted scores in different psychological dimensions can be output and possible occupational interest information can be output. Thus, while saving a lot of computing resources and training time costs, fine-tuning for specific psychological dimensions and occupational category-related tasks can allow the model to better focus on this specific application field, ensuring that the prediction results output by the model are more consistent with actual psychological preferences.

[0098] Further, as a refinement and expansion of the specific implementation of the above embodiment, in order to fully illustrate the specific implementation process of this embodiment, before step S20, the career interest prediction method also includes: inputting the test text into the quality assessment model to obtain an assessment result; if the assessment result of the test text is unqualified, outputting a first prompt message based on the reason for the unqualified. When the system performs a test text quality assessment, step S20 specifically includes: if the assessment result of the test text is qualified, determining that the test text meets the quality requirements, and inputting the test image and the test text that meets the quality requirements into the prediction model.

[0099] The quality assessment model is obtained by fine-tuning the large model of images and texts through training samples and the quality scoring rules corresponding to the training samples. The quality requirements can be reasonably set according to the test accuracy of the prediction model, for example, the text length needs to reach a specified number of words and the text repetition rate needs to be low.

[0100] In this embodiment, the test text provided by the testee is screened through the quality assessment model to avoid prediction bias caused by test texts that may have unclear semantics, are too brief in expression, or have a very low correlation with the test image, which helps to improve the quality of the career interest assessment. Furthermore, when the test text is judged to be unqualified, a first prompt message is output based on the specific reason for its unqualifiedness, thereby providing the testee with a clear and targeted improvement direction, allowing the testee to better understand the test requirements and reducing the difficulty for the testee to enter a test text that meets the quality requirements.

[0101] For example, considering that too little descriptive text may not be sufficient to support subsequent career interest analysis, the system sets a minimum number of words for input text, for example 100 words. If the user's input test text is less than 100 words, the user is prompted to enter more content until the minimum word limit is met before proceeding to the next evaluation process.

[0102] It is understandable that the quality assessment model can also be obtained by fine-tuning the large image and text model. Specifically, the sample text is segmented to determine multiple segmented words of the sample text; the text features of the multiple segmented words are extracted; the text features of the multiple segmented words are evaluated based on the preset scoring rules to determine the quality label of the sample text; based on the sample image, sample text and quality label, the large image and text model is fine-tuned to obtain the quality assessment model.

[0103] The text features include at least one of the following: text length, word repetition frequency, number of punctuation marks, character repetition, and word occurrence frequency. The preset scoring rules can be reasonably set according to the needs.

[0104] It is understandable that the training sample includes a sample image and a sample text describing the sample image. The training sample used for training the quality assessment model may be the same as or different from the training sample used for training the prediction model.

[0105] In this embodiment, the sample text of the image sample is first refined into a plurality of units with independent semantics. After extracting the text features of each word segment, these features are evaluated according to the quality scoring rules to determine the quality label. This helps to deeply analyze the structure and semantic information of the text, avoids making rough judgments only from the overall text level and missing key details, so as to more accurately reflect the actual quality level of the text. Finally, on the basis of the existing large image and text model, a large amount of text quality assessment data (sample images and their sample texts, as well as quality labels (qualified or unqualified) are constructed for fine-tuning training to obtain the final quality assessment model. The trained quality assessment model can learn the text quality judgment rules corresponding to different test images, which can not only maintain the effectiveness of text quality assessment, but also help to dynamically adapt to emerging text features and changing quality requirements.

[0106] It is understandable that the first text sample and sample text of the same test image can be the same data, and the first text sample and sample text need to be authentic, extensive and sufficient to ensure that the trained model is more practical, applicable to various types of users, and can obtain more accurate prediction results.

[0107] The LLaVA (Large Language and Vision Assistant) model is used as an example to illustrate the fine-tuning quality assessment model. The weights obtained by pre-training the LLaVA model on large-scale image and text data are initialized. For the part based on the Transformer architecture, the previous multi-head attention layer is frozen, and only the output layer or the part close to the output layer is fine-tuned. In this way, targeted training can be performed on the quality assessment task while utilizing the pre-training knowledge. On the basis of the LLaVA model, a quality assessment output layer is added to output different quality categories (such as qualified or unqualified). The loss calculated by the cross-entropy loss function (Cross-Entropy Loss) is used to calculate the gradient of each parameter through the back propagation algorithm. Starting from the output layer, the error gradient of each neuron is calculated step by step, and these gradients are propagated back to the previous layer. Then, the optimization algorithm (such as Adam, Adagrad, etc.) is used to update the parameters of the model according to the calculated gradients, so that the model gradually converges to the optimal state.

[0108] Furthermore, as a refinement and extension of the specific implementation methods of the above-mentioned embodiment, in order to fully illustrate the specific implementation process of this embodiment, before step S10, the career interest prediction method based on psychological projection also includes: matching the first image and the second image based on the identity information of the subject; combining the first image and the second image to generate a test image.

[0109] The first image may contain some elements that can cause emotional fluctuations, such as tense scenes, to test the psychological characteristics of the subject, which are used to represent the subject's mental state, emotions, motivations, and subconscious content. The second image may contain some elements that require the subject to make moral judgments, such as scenes involving conflicts of interest, to test the subject's credibility.

[0110] Specifically, the identity information of the subject includes at least one of the following: age, gender, occupation, education level, etc.

[0111] In this embodiment, on the one hand, by using the identity information of the subject to match the first image and the second image suitable for the subject, it is possible to fully consider the cognitive differences caused by the subject's different ages, genders, etc., and match the available test images suitable for the subject through the differences, so as to facilitate more accurate testing for individual differences and ensure the accuracy of subsequent tests. On the other hand, by combining the first image for testing psychological characteristics and the second image for testing credibility to generate a test image, a comprehensive evaluation of the subject can be achieved, and the subject can be prevented from disguising himself due to his understanding of the rules of psychological testing, further improving the accuracy of the test results.

[0112] It is understandable that the first image and the second image may be combined by carefully splicing the two images, or by fusing the contents of the first image and the second image through image processing technology to form a test image.

[0113] It should be noted that if there is no need to perform credibility evaluation, the first image can be directly used as a test image to evaluate the predicted scores of the subject in different psychological dimensions.

[0114] Furthermore, since the test image not only contains the content of the psychological characteristics and credibility of the test subject, in step S20, the test image and the test text are input into the prediction model, which specifically includes: inputting the test image and the test text into the prediction model, so that the prediction model determines the predicted scores of the test subject in different psychological dimensions and the credibility of the test subject based on the test text, and outputs the career interest information and credibility that match the predicted scores. After step S20, the system will also compare the credibility output by the model with the credibility threshold; if the credibility is less than or equal to the credibility threshold, the second prompt information is output.

[0115] In this embodiment, while performing the psychological projection test, credibility verification is also performed. When the credibility obtained by the test reaches a certain level (greater than the credibility threshold), the career interest information is output. On the contrary, when the credibility is less than or equal to the credibility threshold, the second prompt information is output to remind the user that the test has failed and needs to be retested. This reduces the probability of false or inaccurate career interest information and improves the accuracy of career interest matching.

[0116] It is understandable that the credibility test is also predicted by the prediction model. As mentioned above, when fine-tuning the large picture and text model, the credibility label of the text sample is added as the model input sample. An output layer for analyzing the credibility of the text sample is added to the large picture and text model. Through continuous iterative fine-tuning of the training samples, the final prediction model can simultaneously realize the psychological feature evaluation and credibility evaluation of the test text.

[0117] Furthermore, as a refinement and expansion of the specific implementation methods of the above-mentioned embodiment, in order to fully illustrate the specific implementation process of this embodiment, after step S20, the career interest prediction method also includes: accessing a designated academic website to determine the academic information of the subject; if the academic information is consistent with the identity information of the subject, obtaining the subject's historical employment assessment data and user portrait data; based on the historical employment assessment data, user portrait data and the subject's career interest information, making career recommendations for the subject.

[0118] Among them, historical employment assessment data is used to indicate the actual work ability and experience accumulation of the test subject, and user portrait data is used to indicate that the test subject presents unique personality, behavior patterns, etc.

[0119] Specifically, the designated academic website is a trustworthy academic qualification inquiry website, such as China Higher Education Student Information & Career Center. It is worth mentioning that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and provide corresponding operation portals for users to choose to authorize or refuse.

[0120] In this embodiment, the designated academic website is opened, and the academic qualifications provided by the testee are verified through the official standard information provided on the designated academic website, which can make the whole process more standardized and orderly, ensuring that each testee receives academic qualifications verification in a fair and just environment. After the official verification is passed, the testee's past work performance (historical employment evaluation data reflects work results, ability display, etc.), personal personality characteristics and behavioral preferences (content reflected in user portrait data) and current career interests are comprehensively considered to recommend suitable careers for the testee. The occupations recommended by the system are both in line with the interests and can give full play to their professional and ability advantages, avoiding the deviations that may occur when relying on a single dimension for recommendation, improving the quality and success rate of recommendations, and better helping the testee make more rational career choice decisions.

[0121] It should be noted that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0122] In one embodiment, if Figure 3 As shown, as a specific implementation of the above-mentioned career interest prediction method, an embodiment of the present application provides a career interest prediction device 300, and the career interest prediction device 300 includes: an acquisition module 301 and a prediction module 302.

[0123] The acquisition module 301 is used to acquire a test text describing a test image by the testee, wherein the test image is used to test the psychological characteristics of the testee;

[0124] The prediction module 302 is used to input the test image and the test text into the prediction model so that the prediction model determines the predicted scores of the subject in different psychological dimensions based on the test text, and outputs the career interest information matching the predicted scores; wherein the prediction model is obtained by fine-tuning the large image and text model, and the career interest information is used to characterize the subject's interest level in different preset career categories.

[0125] In this embodiment, the principle of psychological projection test is adopted. Through the intelligent analysis of test pictures and complex texts for testing psychological characteristics by the prediction model based on the large picture and text model, the testee's situation in multiple specific psychological dimensions such as emotional stability, stress resistance, and concentration is understood in detail, and the career interest information is dynamically matched through the predicted scores of the testee in different psychological dimensions. Thus, the user's potential career interests can be captured in detail, making the career interest assessment results more accurate and flexible, and realizing accurate intelligent prediction of career interests. In this way, users can more clearly understand which career fields match their own personality, preferences, and interests, thereby reducing the frustration caused by blind selection.

[0126] In one embodiment, the career interest prediction device 300 further includes: a first training module (not shown in the figure);

[0127] The acquisition module 301 is further used to acquire training samples and psychological scoring labels and occupational category labels corresponding to the training samples in different psychological dimensions, wherein the training samples include sample images and sample texts describing the sample images;

[0128] The first training module is used to fine-tune the image-text model based on sample images, sample texts, psychological rating labels, and occupational category labels to obtain a prediction model;

[0129] Among them, the preset occupational categories include at least one of the following: scientific research, assembly and repair, education and training, consulting services, engineering construction, medical services, and finance.

[0130] In one embodiment, the occupational interest prediction device 300 further includes: a detection module (not shown in the figure), a first prompt module (not shown in the figure);

[0131] The detection module is used to input the test text into the quality assessment model to obtain the assessment result, wherein the quality assessment model is obtained by fine-tuning the image-text model through the training samples and the quality scoring rules corresponding to the training samples;

[0132] A first prompt module, configured to output a first prompt message based on the reason for failure if the evaluation result of the test text is failure;

[0133] The detection module is further used to determine that the test text meets the quality requirements if the evaluation result of the test text is qualified;

[0134] The prediction module 302 is specifically used to determine the predicted scores and career interest information of the test subject in different psychological dimensions based on the test image and the test text that meets the quality requirements.

[0135] In one embodiment, the occupational interest prediction device further includes: a word segmentation module (not shown in the figure), a feature extraction module (not shown in the figure), an evaluation module (not shown in the figure) and a second training module (not shown in the figure);

[0136] A word segmentation module is used to perform word segmentation processing on the sample text and determine multiple word segments of the sample text;

[0137] A feature extraction module, used to extract text features of multiple word segments, wherein the text features include at least one of the following: text length, word repetition frequency, number of punctuation marks, word repetition degree, and word occurrence frequency;

[0138] An evaluation module is used to evaluate the text features of multiple word segments based on preset scoring rules to determine the quality label of the sample text;

[0139] The second training module is used to fine-tune the large image and text model based on sample images, sample texts and quality labels to obtain a quality assessment model.

[0140] In one embodiment, the occupational interest prediction device 300 further includes: an identity verification module (not shown in the figure), a recommendation module (not shown in the figure);

[0141] The identity verification module is used to access the designated academic website and determine the educational information of the testee;

[0142] The acquisition module 301 is also used to acquire the historical employment evaluation data and user portrait data of the subject if the educational background information matches the identity information of the subject;

[0143] The recommendation module is used to make career recommendations for the subjects based on historical employment assessment data, user portrait data and the subjects' career interest information.

[0144] In one embodiment, the acquisition module 301 is specifically configured to acquire a test text in response to an input operation.

[0145] In one embodiment, the acquisition module 301 is specifically used to receive voice information, perform recognition processing on the voice information, and obtain a test text.

[0146] In one embodiment, the occupational interest prediction device 300 further includes: a test configuration module (not shown in the figure) and a second prompt module (not shown in the figure);

[0147] A test configuration module, configured to match the first image with the second image based on the identity information of the subject,

[0148] The first image is used to test the psychological characteristics of the subject, and the second image is used to test the credibility of the subject; the first image and the second image are combined to generate a test image;

[0149] Prediction module 302, specifically used to input the test image and the test text into the prediction model, so that the prediction model determines the predicted scores of the testee in different psychological dimensions and the credibility of the testee based on the test text, and outputs the career interest information and credibility matching the predicted scores;

[0150] The second prompt module is used to output second prompt information if the credibility is less than or equal to the credibility threshold.

[0151] For the specific definition of the career interest prediction device, please refer to the definition of the career interest prediction method above, which will not be repeated here. Each module in the above-mentioned career interest prediction device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0152] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client through a network connection. When the computer program is executed by the processor, it realizes the functions or steps on the server side of an intelligent question-answering processing method based on artificial intelligence.

[0153] In one embodiment, a computer device is provided. The computer device may be a client, and its internal structure diagram may be as follows: Figure 5 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server through a network connection. When the computer program is executed by the processor, it realizes the functions or steps of the client side of an intelligent question-answering processing method based on artificial intelligence

[0154] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program: obtaining a test text in which a subject describes a test image, wherein the test image is used to test the psychological characteristics of the subject; inputting the test image and the test text into a prediction model, so that the prediction model determines the predicted scores of the subject in different psychological dimensions based on the test text, and outputs career interest information matching the predicted scores; wherein the prediction model is obtained by fine-tuning a large picture and text model, and the career interest information is used to characterize the subject's interest in different preset career categories.

[0155] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: obtaining a test text in which a subject describes a test image, wherein the test image is used to test the psychological characteristics of the subject; inputting the test image and the test text into a prediction model, so that the prediction model determines the predicted scores of the subject in different psychological dimensions based on the test text, and outputs career interest information matching the predicted scores; wherein the prediction model is obtained by fine-tuning a large picture and text model, and the career interest information is used to characterize the subject's interest in different preset career categories.

[0156] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant descriptions on the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0157] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0158] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0159] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for predicting career interests, characterized in that: The method comprises: Acquire a test text describing a test image from a testee, wherein the test image is used to test the psychological characteristics of the testee; The test image and the test text are input into a prediction model so that the prediction model determines the predicted scores of the subject in different psychological dimensions based on the test text, and outputs career interest information matching the predicted scores; wherein the prediction model is obtained by fine-tuning the large image and text model, and the career interest information is used to characterize the subject's interest level in different preset career categories.

2. The method for predicting occupational interests according to claim 1, characterized in that: Before inputting the test image and the test text into the prediction model, the method further includes: Acquire training samples and psychological score labels and occupational category labels corresponding to the training samples in different psychological dimensions, wherein the training samples include sample images and sample texts describing the sample images; Based on the sample image, the sample text, the psychological score label and the occupational category label, fine-tuning the image-text large model to obtain a prediction model; Among them, the preset occupational categories include at least one of the following: scientific research, assembly and repair, education and training, consulting services, engineering construction, medical services, and finance.

3. The method for predicting occupational interests according to claim 1, characterized in that: Before inputting the test image and the test text into the prediction model, the method further includes: Inputting the test text into a quality assessment model to obtain an assessment result, wherein the quality assessment model is obtained by fine-tuning the image-text model through training samples and quality scoring rules corresponding to the training samples; If the evaluation result of the test text is unqualified, outputting a first prompt message based on the reason for the unqualified result; Inputting the test image and the test text into a prediction model comprises: If the evaluation result of the test text is qualified, it is determined that the test text meets the quality requirements, and the test image and the test text that meets the quality requirements are input into the prediction model.

4. The method for predicting occupational interests according to claim 3, characterized in that: The training sample includes a sample image and a sample text describing the sample image. Before inputting the test text into the quality assessment model to obtain the assessment result, the method further includes: Performing word segmentation processing on the sample text to determine multiple word segments of the sample text; Extracting text features of the multiple word segments, wherein the text features include at least one of the following: text length, word repetition frequency, number of punctuation marks, character repetition, and word occurrence frequency; Evaluate the text features of the multiple word segments based on a preset scoring rule to determine a quality label of the sample text; Based on the sample images, the sample texts and the quality labels, the image-text large model is fine-tuned and trained to obtain the quality assessment model.

5. The method for predicting occupational interests according to claim 1, characterized in that: After inputting the test image and the test text into the prediction model, the method further includes: Visit a designated academic website to determine the educational information of the test subject; If the educational background information matches the identity information of the subject, obtaining the historical employment evaluation data and user portrait data of the subject; Based on the historical employment assessment data, the user portrait data and the career interest information of the subject, career recommendations are made for the subject.

6. The method for predicting occupational interests according to any one of claims 1 to 5, characterized in that: The step of obtaining a test text in which the test subject describes the test image includes: In response to an input operation, obtaining the test text; or, The voice information is received, and recognition processing is performed on the voice information to obtain the test text.

7. The method for predicting occupational interests according to any one of claims 1 to 5, characterized in that: Before obtaining the test text describing the test image from the testee, the method further includes: Matching a first image and a second image based on the identity information of the subject, wherein the first image is used to test the psychological characteristics of the subject, and the second image is used to test the credibility of the subject; combining the first image and the second image to generate the test image; Inputting the test image and the test text into a prediction model comprises: Inputting the test image and the test text into the prediction model, so that the prediction model determines the predicted scores of the test subject in different psychological dimensions and the credibility of the test subject based on the test text, and outputs the career interest information matching the predicted scores and the credibility; After the test image and the test text are input into the prediction model, the career interest prediction method further includes: If the credibility is less than or equal to the credibility threshold, a second prompt message is output.

8. A career interest prediction device, characterized in that: The device comprises: An acquisition module, used for acquiring a test text describing a test image from a testee, wherein the test image is used for testing the psychological characteristics of the testee; A prediction module is used to input the test image and the test text into a prediction model so that the prediction model determines the predicted scores of the subject in different psychological dimensions based on the test text, and outputs career interest information matching the predicted scores; wherein the prediction model is obtained by fine-tuning the large image and text model, and the career interest information is used to characterize the subject's interest level in different preset career categories.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the career interest prediction method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the career interest prediction method according to any one of claims 1 to 7 is implemented.

Images