A cognitive state recognition method, device and terminal equipment

Abstract

The application provides a cognitive state recognition method and device and a terminal device, and is suitable for the technical field of data processing. The method comprises the following steps: generating a plurality of user voice text information and a plurality of user behavior text information according to a plurality of user voice information, a plurality of user behavior information, a preset user voice text conversion model and a preset user behavior text conversion model; generating initial cognitive state recognition information according to the plurality of user voice text information, the plurality of user behavior text information and a preset initial cognitive state recognition model; and generating target cognitive state recognition information according to the initial cognitive state recognition information, a plurality of historical cognitive state recognition information, a plurality of preset user cognitive state division threshold information and preset user cognitive state switching sequence information. The application realizes accurate recognition of the cognitive state of a user after receiving a design stimulus, and is used for adaptively evaluating the receiving effect of the user on the design stimulus.

Description

A cognitive state recognition method, device and terminal equipment Technical Field

[0001] This application belongs to the field of data processing technology, and in particular relates to cognitive state recognition methods, devices and terminal equipment. Background Technology

[0002] In the design process, "design stimuli" typically serve as external information carriers to activate creative thinking; the birth of creative ideas often requires the assistance of external stimuli. However, design stimuli are not directly absorbed by the creative individual and immediately transformed into new solutions; instead, they undergo a series of processing steps within the designer's brain. Research indicates that the internal processing of design stimuli typically involves five core cognitive states: Encounter, Judgment, Transformation, Generation, and Decision. Encounter refers to the designer's initial reception of the stimulus and preliminary reading, browsing, or understanding of its content; Judgment refers to the designer's evaluation or judgment of the stimulus content, including whether the stimulus is relevant to the current task, has value, or is worth further development; Transformation refers to the designer's reinterpretation of the stimulus content to connect it with the current task context; Generation refers to the designer's generation of new concepts or ideas based on the stimulus content; and Decision refers to the designer's selection of the processed stimulus and determination of subsequent steps. Stimulus processing is typically accomplished through these five cognitive states. The sequence, rhythm, and duration of these cognitive states determine the quality of stimulus processing, thereby affecting the degree of stimulus absorption and the final creative outcome.

[0003] However, existing design stimulus methods mainly focus on recommending more suitable stimulus content to users, but do not pay attention to the developmental process of how design stimuli are absorbed, processed and applied by users. As a result, the quality and efficiency of designers' creative output after providing design stimuli are still difficult to control. Summary of the Invention

[0004] In view of this, embodiments of this application provide a cognitive state recognition method, apparatus, and terminal device, aiming to solve the problem in the prior art that it is impossible to recognize the cognitive state of a user after receiving designed stimulus content.

[0005] The first aspect of this application provides a cognitive state recognition method, including:

[0006] Acquire multiple user voice information, multiple user behavior information, and multiple historical cognitive state recognition information;

[0007] Based on the multiple user voice information, multiple user behavior information, a preset user voice-to-text conversion model, and a preset user behavior-to-text conversion model, multiple user voice-to-text information and multiple user behavior-to-text information are generated.

[0008] Based on the multiple user voice text information, multiple user behavior text information, and the preset initial cognitive state recognition model, initial cognitive state recognition information is generated.

[0009] Based on the initial cognitive state identification information, multiple historical cognitive state identification information, multiple preset user cognitive state division threshold information, and preset user cognitive state switching order information, target cognitive state identification information is generated.

[0010] A second aspect of this application provides a cognitive state recognition device, comprising:

[0011] The information acquisition module is used to acquire multiple user voice information, multiple user behavior information, and multiple historical cognitive state recognition information;

[0012] The user voice text information and user behavior text information generation module is used to generate multiple user voice text information and multiple user behavior text information based on the multiple user voice information, multiple user behavior information, a preset user voice text conversion model and a preset user behavior text conversion model.

[0013] The initial cognitive state recognition information generation module is used to generate initial cognitive state recognition information based on the multiple user voice text information, multiple user behavior text information and the preset initial cognitive state recognition model.

[0014] The target cognitive state identification information generation module is used to generate target cognitive state identification information based on the initial cognitive state identification information, multiple historical cognitive state identification information, multiple preset user cognitive state division threshold information, and preset user cognitive state switching order information.

[0015] A third aspect of this application provides a terminal device, the terminal device including a memory and a processor, the memory storing a computer program executable on the processor, and the processor executing the computer program to implement the steps of the cognitive state recognition method described in the first aspect above.

[0016] A fourth aspect of this application provides a computer-readable storage medium, comprising: storing a computer program, wherein when executed by a processor, the computer program implements the steps of the cognitive state recognition method described in the first aspect above.

[0017] Compared with the prior art, the beneficial effects of this application are as follows: This application can accurately identify the cognitive state of users after receiving design stimuli, and can automatically evaluate the user's reception effect of the design stimuli. This helps optimize the way creators process and handle design stimuli, improve the creativity and performance of design results, and at the same time realize refined and intelligent control of the creative conception process of multiple users, ensuring the quality and efficiency of creative output. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 is a schematic diagram of the implementation process of the cognitive state recognition method provided in Embodiment 1 of this application;

[0020] Figure 2 is a schematic diagram of the implementation process of the cognitive state recognition method provided in Embodiment 2 of this application;

[0021] Figure 3 is a schematic diagram of the implementation process of the cognitive state recognition method provided in Embodiment 3 of this application;

[0022] Figure 4 is a schematic diagram of the implementation process of the cognitive state recognition method provided in Embodiment 4 of this application;

[0023] Figure 5 is a schematic diagram of the implementation process of the cognitive state recognition method provided in Embodiment 5 of this application;

[0024] Figure 6 is a schematic diagram of the implementation process of the cognitive state recognition method provided in Embodiment 6 of this application;

[0025] Figure 7 is a schematic diagram of the implementation process of the cognitive state recognition method provided in Embodiment 7 of this application;

[0026] Figure 8 is a schematic diagram of the cognitive state recognition device provided in an embodiment of this application;

[0027] Figure 9 is a schematic diagram of the terminal device provided in an embodiment of this application. Detailed Implementation

[0028] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0029] To illustrate the technical solution described in this application, specific embodiments are provided below.

[0030] Figure 1 shows a flowchart of the cognitive state recognition method provided in Embodiment 1 of this application, which is described in detail below:

[0031] Step S101: Obtain multiple user voice information, multiple user behavior information, and multiple historical cognitive state recognition information.

[0032] In this embodiment, the designed stimulus event can be retrieved from a manually set designed stimulus library and triggered when the designed stimulus content is presented on the user interface. A complete designed stimulus event can start from a start time and end at an end time. The start time of the designed stimulus event can be determined based on the time when the designed stimulus content is presented to the user, and the end time of the designed stimulus event can be determined based on the time when the user places the relevant stimulus content in the stimulus effect evaluation area. User voice information can refer to the voice data corresponding to multiple rounds of dialogue generated in a designed stimulus event. Specifically, it can be the voice dialogue information of multiple times in a designed stimulus event, which can be obtained by recording during the designed stimulus event. Specifically, it can be obtained by collecting voice within the complete time range from the start time to the end time of the designed stimulus event. User behavior information can be the data corresponding to various interactive actions generated by users in a designed stimulus event. Specifically, it can be the multiple sets of behavioral data corresponding to multiple rounds of dialogue generated in the designed stimulus event, which can include actions such as clicking, scrolling, copying, dragging, and placing the designed stimulus content into any stimulus effect evaluation area such as "useful", "useless", "observe", etc. It can be obtained by automatically detecting user interaction behavior during the designed stimulus event. Specifically, it can be obtained by monitoring various user operations on the interactive interface in real time within the complete time range of the designed stimulus event and classifying them into the behavioral data of each dialogue according to the time sequence of the dialogue. Historical cognitive state identification information can be a set of cognitive states identified by a computer system and arranged in chronological order throughout the entire process of a complete designed stimulus event, from the user's initial exposure to the designed stimulus content to the event's conclusion. This information can represent the user's processing of the designed stimulus content and can be stored in sequence. Understandably, historical cognitive state identification information can be included within the current designed stimulus event. Before calculating the user's current cognitive state, the computer system identifies the user's cognitive state information from different past time periods, which can include exposure, judgment, transformation, generation, and decision-making, indicating the user's cognitive states in different past time periods and stored in time series form.

[0033] In this embodiment, in a designed stimulus event, the computer system may present only one designed stimulus to the user. The computer system will not present the next designed stimulus until the user has made a decision regarding the current stimulus.

[0034] In this embodiment, the design stimulus content received by the user can refer to external information carriers used to activate creative thinking in design tasks within the fields of artificial intelligence and computer-aided creativity. These stimuli can take various forms, including text stimuli, image stimuli, audio stimuli, and video stimuli. It is understood that the design stimulus content can be presented to the user during the creative brainstorming activity. A dedicated database can be built around the design stimuli, and the user's cognitive state regarding the design stimulus event can be identified based on voice and behavioral data during the interaction between the user and the design stimulus.

[0035] In this embodiment, preferably, the design stimulus content can be presented in the form of service design heuristic cards, which can include structured fields such as number, title, definition, application example, image information, keywords and key content.

[0036] Step S102: Generate multiple user voice text information and multiple user behavior text information based on the multiple user voice information, multiple user behavior information, preset user voice text conversion model and preset user behavior text conversion model.

[0037] In this embodiment, both the preset user speech-to-text conversion model and the preset user behavior text conversion model can be manually set. The preset user speech-to-text conversion model can be a Paraformer Chinese speech recognition pre-trained model; the preset user behavior text conversion model can be a Qwen3-Max model, which can be called through the Qwen3-Max API. The preset user speech-to-text conversion model can be used to convert multiple user speech information to generate corresponding multiple user speech-to-text information, and the preset user behavior text conversion model can be used to convert multiple user behavior information to generate corresponding multiple user behavior text information.

[0038] Step S103: Generate multiple initial cognitive state recognition information based on the multiple user voice text information, multiple user behavior text information, and the preset initial cognitive state recognition model.

[0039] In this embodiment, the preset initial cognitive state recognition model can be manually set. Before the preset initial cognitive state recognition model performs calculations, the identification of non-stimulus processing states and pause states can be performed first. It is understood that user voice information and user behavior information can be collected and analyzed to identify the degree of understanding and processing of the designed stimulus content after receiving it. However, after receiving the designed stimulus content, the user may not immediately enter the understanding and processing stage. If the analysis and calculation are performed directly through the preset initial cognitive state recognition model, the identification result of the cognitive state may be incorrect. Therefore, before the preset initial cognitive state recognition model performs calculations, it is necessary to first identify whether the user is in the understanding and processing stage of the designed stimulus content. Specifically, the non-stimulus processing state refers to the user's dialogue content and operational behavior not revolving around the currently presented designed stimulus content, but turning to other topics unrelated to the designed stimulus content. This can be determined by calculating the similarity between the acquired user voice information and the user's designed task information in the designed stimulus event. The pause state refers to the user not engaging in any dialogue for a continuous period of time during the presentation of the designed stimulus content, which can be judged by the interval between user voice messages. When it is determined that the user is neither in a non-stimulus processing state nor a paused state, the system proceeds to calculate a preset initial cognitive state recognition model to identify the user's cognitive state in understanding and processing the designed stimulus content. This can be done by first identifying the non-stimulus processing state and then the paused state, or vice versa.

[0040] In this embodiment, the preset initial cognitive state recognition model can be a model combining BERT and CNN, which can identify different cognitive states of the user during the understanding and processing of designed stimuli. The output content can include five core cognitive state information: contact, judgment, transformation, generation, and decision. The preset initial cognitive state recognition model can be trained using a manually constructed cognitive state dataset. Specifically, the training process can involve first using a BERT pre-trained model to acquire multiple cognitive state data collected manually. corresponding feature vector It can be written as: ;in, Representing cognitive state data The length of the data can then be determined, and a CNN model can be used to process multiple cognitive state data. eigenvectors The convolution operation can be denoted as: ,in, The convolution kernel represents the convolution operation. The width is , This represents the bias in the convolution operation. Represents a non-linear activation function. This represents the output of the convolution operation, the parameters of the BERT pre-trained model are frozen during training, and the loss function throughout the training process. It can be written as:

[0041] in, Representing cognitive state data One-hot encoding, This represents the probability of the cognitive state output by the cognitive state recognition model.

[0042] The initial cognitive state recognition model obtained after training is the preset initial cognitive state recognition model. Specifically, during the calculation process, multiple user speech-text information and multiple user behavior-text information corresponding to each dialogue in the designed stimulus event dialogue sequence can be merged. Then, the merged speech-behavior-text information is input into the preset initial cognitive state recognition model. The preset initial cognitive state recognition model outputs the predicted probabilities of multiple cognitive states corresponding to each dialogue. This can be calculated by associating the cognitive states corresponding to previous dialogues, thereby generating multiple initial cognitive state recognition information that reflects the current cognitive processing stage of the dialogue. Understandably, the initial cognitive state recognition information can include contact, judgment, transformation, generation, and decision-making.

[0043] Step S104: Generate target cognitive state identification information based on the initial cognitive state identification information, multiple historical cognitive state identification information, multiple preset user cognitive state division threshold information, and preset user cognitive state switching order information.

[0044] In this embodiment, the preset user cognitive state switching order information can be manually preset. The switching order of various cognitive states and the judgment criteria for four types of cognitive state changes—advance, jump, cycle, and regression—can be set according to the logical sequence of contact, judgment, transformation, generation, and decision-making. Multiple preset user cognitive state classification thresholds can also be manually preset, including thresholds for the number of non-stimulus processing segments, pause duration, jump span, number of cycle repetitions, regression span, cognitive state coverage, Shannon entropy of the cognitive state sequence, and cognitive state sequence length. A complete cognitive state sequence can be constructed by combining multiple historical cognitive state identification information to determine the position of the current initial cognitive state identification information in the sequence. Then, the initial cognitive state identification information is compared with historical cognitive state identification information. Based on the preset user cognitive state switching order information, the current cognitive state change type is determined. Then, according to the multiple preset user cognitive state classification thresholds, further identification and calculation are performed on non-stimulus processing, pause, and cognitive state change situations to determine the cognitive state that best matches the current cognitive processing stage, thereby generating the target cognitive state identification information. The target cognitive state recognition information includes normal, observational, and abnormal situations under four cognitive state changes: forward, jumping, looping, and backward. Specifically, it can include normal, observational, and abnormal situations under the forward cognitive state, the jumping cognitive state, the looping cognitive state, the backward cognitive state, the non-stimulus processing state, and the paused state. This allows it to accurately reflect the user's cognitive processing results of the designed stimulus content in the current dialogue.

[0045] The cognitive state recognition method provided in this application embodiment can accurately identify the cognitive state of a user after receiving design stimulus content. It can automatically evaluate the user's reception effect of the design stimulus content, help optimize the way and process of creatives to process design stimulus content, improve the creativity and performance of design results, and at the same time realize refined and intelligent control of the creative conception process of multiple users, ensuring the quality and efficiency of creative output.

[0046] Figure 2 shows a flowchart of the cognitive state recognition method provided in Embodiment 2 of this application. The difference between it and Embodiment 1 is that:

[0047] The plurality of user voice information includes first user voice information and second user voice information;

[0048] The first user voice information includes first user voice time sequence identifier information, first user voice start time information, and first user voice end time information;

[0049] The second user voice information includes second user voice time sequence identifier information, second user voice start time information, and second user voice end time information;

[0050] Wherein, the first user's voice time sequence identifier information is less than the second user's voice time sequence identifier information;

[0051] Step S102 specifically includes:

[0052] Step S201: Calculate the user voice interval information based on the first user voice end time information and the second user voice start time information.

[0053] In this embodiment, the first user voice information and the second user voice information can be two voice information entries within a single designed stimulus event, specifically two user voice information entries collected between the start and end times of the designed stimulus event. It is understood that there are at least two user voice information entries within a single designed stimulus event; this embodiment only lists two user voice information entries for illustrative purposes. It is understood that the first user voice time sequence identifier information can be used to indicate the collection time of the first user voice information within a designed stimulus event, and the second user voice time sequence identifier information can be used to indicate the collection time of the second user voice information within the same designed stimulus event. The fact that the first user voice time sequence identifier information is less than the second user voice time sequence identifier information can be used to indicate that the collection time of the first user voice information takes precedence over the collection time of the second user voice information. The user voice interval information refers to the time difference between the end time information of the first user voice information and the start time information of the second user voice information. This can be used to reflect the interval between two user voice information entries in the time dimension, and thus, this value can be used to determine whether the two voice entries belong to the same round of dialogue.

[0054] Step S202: Determine whether the user voice interval information is greater than the preset user voice interval threshold; if yes, proceed to step S203; if no, proceed to step S204.

[0055] In this embodiment, the preset user voice interval threshold can be preset manually. It can be set manually by referring to the threshold corresponding to the speech attribution determination rule of adjacent speech interval in the design stimulus event. It can be used to distinguish whether two user voice information belong to the same dialogue.

[0056] Step S203: Differentiate and process the first user voice information and the second user voice information to obtain multiple user voice information to be converted.

[0057] In this embodiment, if the user voice interval information is greater than the preset user voice interval threshold, it indicates that the first user voice information and the second user voice information belong to two different and independent dialogue contents. The specific processing refers to marking the first user voice information and the second user voice information as independent user voice information to be converted.

[0058] Step S204: Perform splicing processing on the first user voice information and the second user voice information to generate user voice information to be converted.

[0059] In this embodiment, if the user voice interval information is less than or equal to the preset user voice interval threshold, it indicates that the first user voice information and the second user voice information belong to the same dialogue content. The first user voice information and the second user voice information can be integrated according to the order of the first user voice time sequence identifier information and the second user voice time sequence identifier information, and merged into a complete voice information as the user voice information to be converted.

[0060] Step S205: Generate multiple user voice text information and multiple user behavior text information based on the multiple user voice information to be converted, multiple user behavior information, a preset user voice text conversion model, and a preset user behavior text conversion model.

[0061] In this embodiment, the preset user speech-to-text conversion model can be manually preset, and the preset user behavior text conversion model can also be manually preset. The preset user speech-to-text conversion model can be a Paraformer Chinese speech recognition pre-trained model, and the preset user behavior text conversion model can be a Qwen3-Max model that can be called via the Qwen3-Max API. Multiple user speech information to be converted can be input into the preset user speech-to-text conversion model for conversion, generating corresponding multiple user speech-to-text information. Then, multiple user behavior information can be simultaneously input into the preset user behavior text conversion model for conversion, generating corresponding multiple user behavior text information.

[0062] The cognitive state recognition method provided in this application embodiment enables refined classification and processing of multiple user voice information segments, optimizes the conversion basis of voice information to text information, and improves the accuracy of the generated initial cognitive state recognition information and target cognitive state recognition information. This strengthens the refined and intelligent control of user cognitive state recognition, assists in optimizing the way creators process and handle design stimuli, and ensures and enhances the creativity and performance of design results.

[0063] Figure 3 shows a flowchart of the cognitive state recognition method provided in Embodiment 3 of this application, which differs from Embodiment 2 above in that:

[0064] The user behavior information includes the time information of the user behavior and the user behavior action information; wherein, the time information of the user behavior and the user behavior action information correspond one-to-one.

[0065] Step S205 specifically includes:

[0066] Step S301: Extract the start time information and end time information of the multiple user voice information to be converted, and obtain the start time information and end time information of the multiple user voice information to be converted.

[0067] In this embodiment, each of the multiple user voice information to be converted corresponds to the voice content of different dialogue units in the designed stimulus event. By extracting the start time information and end time information corresponding to the user voice information to be converted, the time interval corresponding to each user voice information to be converted can be calculated, which can be used to provide accurate time anchors for subsequent matching of user behavior information within the same time interval.

[0068] Step S302: Based on the start time information and end time information of the voices of the multiple users to be converted, the occurrence time information of the multiple user behaviors is filtered to obtain the occurrence time information of the multiple user behaviors to be converted.

[0069] In this embodiment, multiple user behavior occurrence time information can be compared with multiple user voice start time information and multiple user voice end time information to be converted. Only user behavior occurrence time information within the voice time interval of each user to be converted is retained, so that the filtered user behavior occurrence time information can achieve accurate time dimension matching with the voice information of the user to be converted, thereby generating user behavior occurrence time information corresponding to the voice information of each dialogue unit.

[0070] Step S303: Determine multiple user behavior action information based on the occurrence time information of the multiple user behaviors to be converted.

[0071] In this embodiment, based on the multiple user behavior occurrence time information obtained through filtering, the corresponding time-related behavior action data can be retrieved from all user behavior action information. This ensures that each user behavior occurrence time information can be matched with a unique and corresponding user behavior action information, so that the behavior data and voice data can correspond in both time and content dimensions, thereby generating user behavior action information corresponding to each dialogue unit.

[0072] Step S304: Generate multiple user voice text information and multiple user behavior text information based on the multiple user voice information to be converted, multiple user behavior action information, preset user voice text conversion model and preset user behavior text conversion model.

[0073] In this embodiment, the preset user speech-to-text conversion model can be manually preset, and the preset user behavior text conversion model can also be manually preset. Specifically, the preset user speech-to-text conversion model can be a Paraformer Chinese speech recognition pre-trained model, and the preset user behavior text conversion model can be a Qwen3-Max model that can be called via the Qwen3-Max API. Multiple user speech information to be converted can be input into the preset user speech-to-text conversion model for conversion, generating corresponding multiple user speech-to-text information; similarly, multiple user behavior action information can be input into the preset user behavior text conversion model for conversion, generating corresponding multiple user behavior text information.

[0074] The cognitive state recognition method provided in this application embodiment enables accurate matching and filtering of user voice information and user behavior information in the time dimension, so that multiple generated user voice text information and multiple user behavior text information can achieve a high degree of correspondence in the spatiotemporal dimension, improve the accuracy of the generated initial cognitive state recognition information and target cognitive state recognition information, strengthen the refined control of user cognitive state recognition, thereby assisting in optimizing the way creators process and handle design stimuli, and ensuring and improving the creativity and performance of design results.

[0075] Figure 4 shows a flowchart of the cognitive state recognition method provided in Embodiment 4 of this application. The difference between this method and Embodiment 1 is that step S103 specifically includes:

[0076] Step S401: Merge the multiple user voice text information and multiple user behavior text information to generate multiple voice behavior text information to be recognized.

[0077] In this embodiment, the user voice text information and user behavior text information corresponding to each dialogue can be merged according to the chronological order of the dialogue sequence in the designed stimulus event, so that the merged text information can be used as multiple voice behavior text information to be identified.

[0078] In this embodiment, the first step may be to determine the first stimulus event in the design. Second dialogue time window User voice information It also calls the Tongyi Listening API to obtain user voice information in real time. Corresponding user voice text information Therefore, the first stimulus event in the design event is determined. Second dialogue time window User behavior information It also calls the Qwen3-Max API to transmit user behavior information in real time. Convert into corresponding user behavior text information Then user voice information can be merged. Corresponding user voice text information and user behavior information Corresponding user behavior text information , obtained the Second dialogue Text information of speech behavior to be recognized .

[0079] Step S402: Calculate the similarity between the multiple voice behavior text information to be identified and the preset user design task information to obtain the similarity information between the voice behavior text and the design task.

[0080] In this embodiment, the preset user design task information can be manually preset and can refer to the design content within the same scheme. For example, it can be the target audience, service area, service function, etc., and can be used to indicate the task theme that needs to be clarified before making specific designs. Similarity calculation can refer to calculating the similarity between the speech behavior text information to be recognized corresponding to the initial cognitive state recognition information and the preset user design task information, and the calculation result is used as the similarity information between the speech behavior text and the design task.

[0081] In this embodiment, the text information of the speech behavior to be recognized is calculated. and preset user design task information similarity between .

[0082] In this embodiment, it is understood that similarity calculations cannot be performed on paused states. Alternatively, paused states can be excluded before similarity calculations are performed to further identify non-stimulus processing states or other cognitive states.

[0083] Step S403: Determine whether the similarity information between the voice behavior text and the design task is less than the preset similarity threshold information between the voice behavior text and the design task; if yes, proceed to step S404; if no, proceed to step S405.

[0084] In this embodiment, the preset similarity threshold information between the voice behavior text and the design task can be preset by humans. By judging the numerical relationship between the similarity information between the voice behavior text and the design task and the preset similarity threshold information between the voice behavior text and the design task, it can be used to accurately determine whether the user's voice behavior text deviates from the main body of the design task. It can also be used to determine whether the voice behavior text information to be identified revolves around the preset user design task information, and thus determine whether the user is in the cognitive processing stage of the design stimulus content.

[0085] In this embodiment, the preset similarity threshold information between the speech behavior text and the design task can be manually set and can be used to represent the speech behavior text information to be recognized. and preset user design task information The similarity threshold between them can be denoted as: .like This allows for a reduction in the number of non-stimulatory treatment fragments. Add 1 to the numerical value. The number of non-stimulus treatment segments can be manually set. The initial value is 0 to record the number of times non-stimulus treatment segments are performed. The number of times the segment is performed can be... Second dialogue The duration of the corresponding initial cognitive state recognition information in the designed stimulus event is defined as follows: ,in .in, This indicates the start time information for recognizing the initial cognitive state. This indicates the end time information corresponding to the initial cognitive state recognition and matching indication information.

[0086] Step S404: Generate non-stimulus treatment state recognition information as initial cognitive state recognition information.

[0087] In this embodiment, when the similarity information between the voice behavior text and the design task is less than the preset similarity threshold information between the voice behavior text and the design task, it indicates that the voice behavior text information to be identified has not been developed around the preset user design task information, that is, the user has not yet entered the cognitive processing stage of the design stimulus content, thereby generating non-stimulus processing state identification information as the initial cognitive state identification information.

[0088] Step S405: Calculate the multiple user voice interval time information based on the start time information and end time information of the multiple user voice information.

[0089] In this embodiment, it can be understood that each user voice information corresponds to a unique start time information and end time information. The user voice interval time information corresponding to two user voice information can be obtained by subtracting the end time information corresponding to the previous user voice information from the start time information of each user voice information.

[0090] Step S406: Determine whether the multiple user voice interval time information is greater than the preset user voice interval time threshold information; if yes, proceed to step S407; if no, proceed to step S408.

[0091] In this embodiment, the preset user voice interval time threshold information can be set manually and can be used to distinguish whether the user voice interval time information corresponding to two voice dialogue contents exceeds the normal duration limit, thereby determining whether the user is in a pause state.

[0092] Step S407: Generate pause state recognition information as initial cognitive state recognition information.

[0093] In this embodiment, when the time interval between multiple user voice messages exceeds the preset threshold, it indicates that the user has been silent for too long during the conversation and has not effectively progressed in cognitive processing of the designed stimulus content, and is in a state of cognitive stagnation. Thus, pause state identification information is generated as initial cognitive state identification information, which can be used to mark the user's current stagnation in cognitive processing of the designed stimulus content.

[0094] Step S408: Generate initial cognitive state recognition information based on the multiple text information of the speech behaviors to be recognized and the preset initial cognitive state recognition model.

[0095] In this embodiment, if the similarity information between the voice behavior text and the design task is greater than or equal to a preset similarity threshold information between the voice behavior text and the design task, and the duration information of multiple user voices is less than or equal to a preset user voice pause time threshold information, it indicates that the user is in the effective cognitive processing stage of the design stimulus content. Then, multiple voice behavior text information to be identified can be input into a preset initial cognitive state recognition model to extract features of the voice behavior text information to be identified and classify cognitive states. Thus, the initial cognitive state recognition model outputs one of the five core cognitive state information categories of contact, judgment, transformation, generation, and decision as the user's current initial cognitive state recognition information.

[0096] In this embodiment, the preset initial cognitive state recognition model can be manually preset. Specifically, this model is a combination of BERT and CNN, which can be constructed by combining a BERT pre-trained model with a CNN model and trained using a manually collected cognitive state dataset. Multiple speech behavior text information to be recognized can be input into the preset initial cognitive state recognition model. Then, the BERT pre-trained model extracts the feature vectors of each speech behavior text information. A CNN model is then used to convolve the feature vectors, and the predicted probabilities of five cognitive states (contact, judgment, transformation, generation, and decision) corresponding to each speech behavior text information are output. Simultaneously, the cognitive states of previous dialogues in each dialogue unit are correlated for comprehensive judgment, thereby generating multiple initial cognitive state recognition information that reflects the cognitive processing stages of each dialogue.

[0097] In this embodiment, when k is not equal to 1, it can be based on the text information of the speech behavior to be recognized. The user's initial cognitive state is determined through a pre-set initial cognitive state recognition model in the first... Second dialogue Initial cognitive state recognition information This can be combined with obtaining user information in the first... Second dialogue Initial cognitive state recognition information Understandably, the k-1th dialogue here refers to the dialogue information corresponding to the last cognitive state information among multiple historical cognitive state recognition information, that is, the dialogue information corresponding to neither the pause state nor the non-stimulus processing state.

[0098] The cognitive state recognition method provided in this application accurately identifies non-stimulus processing states and pause states, effectively avoids misjudgment of cognitive state recognition, improves the accuracy of cognitive state recognition, helps optimize the way creators process and handle design stimuli, and enhances the creativity and performance of design results.

[0099] Figure 5 shows a flowchart of the cognitive state recognition method provided in Embodiment 5 of this application, which differs from Embodiment 4 above in that:

[0100] Multiple preset user cognitive state classification thresholds include multiple preset target non-stimulus processing state classification thresholds, multiple preset target pause state classification thresholds, and multiple preset target cognitive state classification thresholds.

[0101] Step S104 specifically includes:

[0102] Step S501: Determine whether the initial cognitive state identification information is non-stimulus processing state identification information; if yes, proceed to step S502; if no, proceed to step S505.

[0103] In this embodiment, by determining whether the initial cognitive state identification information is non-stimulus processing state identification information, it is ensured that the subsequently generated target cognitive state identification information can accurately match the user's current cognitive processing situation.

[0104] Step S502: Obtain information on the number of times the non-stimulation treatment state was identified.

[0105] In this embodiment, the number of times the non-stimulus treatment state is identified can be the number of times the computer system identifies the non-stimulus treatment state after initially identifying the user's previous cognitive state in the current designed stimulus event, and can be retrieved from the computer system.

[0106] Step S503: Calculate the duration of the non-stimulated user speech based on the start and end time information of the user speech information corresponding to the initial cognitive state recognition information.

[0107] In this embodiment, the initial cognitive state identification information is the non-stimulus processing state identification information. By extracting the start time and end time information of the user voice information corresponding to the initial cognitive state identification information, and subtracting the corresponding start time information from the end time information of the user voice information corresponding to the non-stimulus processing state, the current non-stimulus user voice duration information is obtained.

[0108] Step S504: Generate target cognitive state recognition information based on the non-stimulation processing state recognition count information, the non-stimulation user voice duration information, and multiple preset target non-stimulation processing state division threshold information.

[0109] In this embodiment, the multiple preset threshold information for dividing the target non-stimulation treatment state can be preset by humans and may include threshold information for the duration of the non-stimulation treatment state and threshold information for the number of times the non-stimulation treatment state occurs, which can be used to distinguish different situations of the non-stimulation treatment state.

[0110] In this embodiment, specifically, the duration of the observation can be... The threshold is defined as Duration of abnormal situations The threshold is defined as The number of times the non-stimulus treatment state was identified is defined as Among them, when the information on the number of times the non-stimulus treatment state is identified is simultaneously satisfied... and When the cognitive state under non-stimulus treatment is determined to be normal, it is considered normal if any of the following conditions are met. or When observing, the cognitive state under non-stimulus treatment is defined as such. This is done when any of the following conditions are met. or When the cognitive state under non-stimulus treatment is determined to be abnormal, different target cognitive state recognition information is generated.

[0111] Step S505: Determine whether the initial cognitive state recognition information is pause state recognition information; if yes, proceed to step S506; if no, proceed to step S507.

[0112] In this embodiment, after excluding the case where the initial cognitive state identification information is non-stimulus processing state identification information, the specific situation of the initial cognitive state identification information can be further determined.

[0113] Step S506: Generate target cognitive state recognition information based on the multiple user voice interval time information and multiple preset target pause state division threshold information.

[0114] In this embodiment, the multiple preset target pause state classification thresholds can be manually preset and may include normal thresholds, observation thresholds, and abnormal thresholds for the duration of the pause state. These thresholds can be used to define different specific pause states. Multiple user speech interval time information can be compared with these thresholds. If the user speech interval time information is within the normal threshold range, a normal pause state is generated as the target cognitive state identification information; if the user speech interval time information is between the normal threshold and the observation threshold, an observation pause state is generated as the target cognitive state identification information; if the user speech interval time information exceeds the abnormal threshold, an abnormal pause state is generated as the target cognitive state identification information, thereby achieving accurate determination of pause states.

[0115] In this embodiment, it can also be based on the first Second dialogue start time and the Second dialogue End time Calculate the first Second dialogue and the Second dialogue pause time Furthermore, based on the pause duration, the pause status is categorized into three types: normal, observation, and abnormal. When the following conditions are met... When the cognitive state is determined to be the normal state under pause, the threshold is defined as follows: This is a threshold used to distinguish whether user voice messages belong to different dialogues. When the time interval between two user voice messages is less than this threshold, the two user voice messages belong to the same dialogue; when the time interval between two user voice messages is greater than or equal to this threshold, the two user voice messages do not belong to the same dialogue. When the cognitive state is determined to be the observation state under pause; when the condition is met... When the cognitive state is determined to be an abnormal state under pause, different target cognitive state recognition information is generated.

[0116] Step S507: Generate target cognitive state identification information based on the initial cognitive state identification information, multiple historical cognitive state identification information, multiple preset target cognitive state division threshold information, and preset user cognitive state switching order information.

[0117] In this embodiment, the multiple preset target cognitive state division thresholds can be preset manually, and may include cognitive state change span thresholds, loop repetition number thresholds, backtracking span thresholds, cognitive state coverage thresholds, cognitive state sequence Shannon entropy thresholds, and cognitive state sequence length thresholds, etc.; the preset user cognitive state switching order information can be preset manually, and the judgment criteria for four types of cognitive state change types such as forward, jump, loop, and backtracking can be set according to the sequential logic of contact, judgment, transformation, generation, and decision. First, a complete cognitive state sequence can be constructed by combining initial cognitive state identification information with multiple historical cognitive state identification information. Then, the initial cognitive state identification information is compared with the historical cognitive state identification information, and the current cognitive state change type is determined based on the preset user cognitive state switching sequence information, namely one of the four cognitive state change types: forward, jump, loop, and back. Then, threshold information is divided according to multiple preset target cognitive states, which can quantify the cognitive state change situation. Based on the judgment results, the cognitive state that best matches the current cognitive processing stage can be analyzed, namely the normal situation, observation situation, and abnormal situation under the four cognitive state changes of forward, jump, loop, and back. Thus, target cognitive state identification information is generated, and user cognitive guidance strategy information can be matched based on the target cognitive state identification information.

[0118] In this embodiment, the user's pre-set cognitive state switching order information may include the location index of different cognitive states, i.e. , , , , These are the location indices for Encounter, Judgement, Transformation, Generation, and Decision, respectively. They can be determined based on the position index of the first... Second dialogue Initial cognitive state recognition information and the Second dialogue Initial cognitive state recognition information This involves identifying changes in cognitive states, specifically recognizing four types of cognitive state changes: forward movement, jumping, looping, and backward movement. If the following conditions are met... If the cognitive state change is marked as progress; if the condition is met... If the cognitive state change is satisfied, then the change is marked as a jump; Then, the change in cognitive state is marked as a cycle; if the following conditions are met... If so, the change in cognitive state is marked as regression.

[0119] In this embodiment, the progression can be categorized into three types based on duration: normal, observation, and abnormal. Understandably, the duration of the (k-1)th dialogue can be defined as... The duration under normal circumstances can be The threshold is defined as Duration under observation The threshold is defined as Then when the following conditions are met When it is normal, it is identified as such; when it meets the requirements... When, it is identified as an observed situation; when it meets the following conditions... When an abnormal situation is identified, different target cognitive state identification information is generated. It is understood that in this embodiment, the initial cognitive state identification information used to judge changes in cognitive state refers only to the five core cognitive states of contact, judgment, transformation, generation, and decision-making, and does not include situations where the initial cognitive state identification information is a paused state or a non-stimulus processing state.

[0120] In this embodiment, the jump occurrences can also be categorized into three types based on statistical results: normal, observed, and abnormal. The number of jumps can be defined as... Define the jump span as ,in The initial number of jumps can be set manually. Whenever a jump occurs, it makes Add 1 to the value. (The last part is incomplete and likely refers to a specific number or sequence.) Second dialogue The duration is defined as ,in Duration under observation The threshold is defined as Duration under abnormal circumstances The threshold is defined as Then, if the number of jumps is satisfied simultaneously... , jump span and When the number of jumps is met, it is identified as a normal situation; when the number of jumps is met simultaneously... , jump span and When the following conditions are met, the number of jumps is identified as the observed situation; , jump span or When this is identified as an abnormal situation, different target cognitive state recognition information is generated.

[0121] In this embodiment, the cyclical situation can also be divided into three types based on statistical results: normal, observation, and abnormal. This can be based on the... Second dialogue Text information of speech behavior to be recognized and the Second dialogue Text information of speech behavior to be recognized Calculate similarity Therefore, the number of times the loop repeats can be defined as Let the initial value of the number of repetitions be... .if This makes Add 1 to the value; if ,but Unchanged. Among them, Indicates similarity The threshold. Then when the condition is met... When it is normal, it is identified as such; when it meets the requirements... When, it is identified as an observed situation; when it meets the following conditions... When this occurs, it is identified as an abnormal situation, and different target cognitive state identification information is generated. Among them, cyclic repetition refers to a situation where the dialogue content before and after the identified cognitive state changes to a cyclic state is highly similar, indicating that the user has repeatedly performed the same processing in the same cognitive state without introducing new perspectives or information.

[0122] In this embodiment, the rollback situation can also be divided into three types based on statistical results: normal, observed, and abnormal. The rollback span can be defined as... ,in The first Second dialogue The duration is defined as ,in , will last The threshold is defined as Then when both conditions are met... and When the condition is met, it is identified as a normal situation; when both conditions are met simultaneously... and When, it is identified as an observed situation; when both conditions are met simultaneously. and When this is identified as an abnormal situation, different target cognitive state recognition information is generated.

[0123] The cognitive state recognition method provided in this application can accurately distinguish and identify non-stimulus processing states, pause states, and the core cognitive states of users during the stimulus design process, thereby improving the accuracy and relevance of target cognitive state recognition information. It can be used to provide reliable data support for the accurate matching of dynamic stimulus guidance strategies, so as to help optimize the way creators process and handle design stimuli, and ensure and improve the creativity and performance of design results.

[0124] Figure 6 shows a flowchart of the cognitive state recognition method provided in Embodiment Six of this application. The difference between this method and Embodiment Five is that, after step S507, the method further includes:

[0125] Step S601: Calculate the length information of the cognitive state identification sequence based on the initial cognitive state identification information and multiple historical cognitive state identification information.

[0126] In this embodiment, multiple historical cognitive state identification information can be a set of cognitive states arranged in chronological order. Initial cognitive state identification information is then added to the end of the multiple historical cognitive state identification information to form a complete cognitive state identification sequence. The number of cognitive states contained in this sequence is then counted to generate the length information of the cognitive state identification sequence. It is understood that the five core cognitive states that can be used to supplement historical cognitive state identification information are contact, judgment, transformation, generation, and decision-making.

[0127] Step S602: When the length of the cognitive state recognition sequence is less than the preset threshold information for the length of the cognitive state recognition sequence, proceed to step S603.

[0128] In this embodiment, the preset cognitive state recognition sequence length threshold information can be preset by humans and can be set to 4. It can be used to determine whether the current cognitive state recognition sequence has reached the length standard for effective analysis.

[0129] In this embodiment, when the length of the cognitive state recognition sequence is greater than or equal to the preset threshold information for the length of the cognitive state recognition sequence, the composition structure of the cognitive state recognition sequence and the length information of the cognitive state recognition sequence can be analyzed and calculated. Subsequently, user cognitive guidance strategy information can be matched based on the analysis and calculation results of the composition structure and the length information of the cognitive state recognition sequence.

[0130] In this embodiment, specifically, the method for analyzing and calculating the composition structure of the cognitive state recognition sequence can be to first determine the first... Second dialogue Cognitive state recognition sequence When the length of the cognitive state recognition sequence is greater than 3, the sequence's composition structure can be analyzed. This analysis can be based on coverage and Shannon entropy. Coverage refers to the occurrence of different cognitive states within the same stimulus event, used to measure the completeness of cognitive states during stimulus processing. The initial value can be manually set to 0. Sequences can be recognized in a cognitive state. In this process, whenever a new cognitive state that has not appeared before is encountered, the coverage is updated. The update method can be to increment the coverage value by 1. This allows for the statistical analysis of all cognitive recognition information, including contact, judgment, transformation, generation, and decision-making, within the cognitive state recognition sequence. The proportion of occurrences in ,in This allows for the calculation of cognitive state recognition sequences. Shannon entropy It can be written as: It can convert the cognitive state sequence The compositional structure is divided into three types: normal, observed, and abnormal. Then, when simultaneously satisfying... and When the condition is met, it is identified as a normal situation; when both conditions are met simultaneously... and When, it is identified as an observed situation; when any of the following conditions are met. or When this occurs, it is identified as an abnormal situation. Understandably, the ranges for coverage and Shannon entropy are both derived experimentally.

[0131] In this embodiment, specifically, the determination of the cognitive state recognition sequence length information can be as follows: based on the first... Second dialogue length of cognitive state sequence ,in The length of the cognitive state sequence can be... The situation is divided into four types: normal sequences, deep sequences, and ultra-long sequences. When the cognitive state sequence length is satisfied... When the sequence is in normal condition, it is considered a normal sequence; when the cognitive state sequence length is satisfied... When the cognitive state sequence length is satisfied, it is determined to be a deep sequence; when the cognitive state sequence length is satisfied, it is determined to be a deep sequence. hour, The completeness and depth of cognitive processing can also be determined based on the sequence length. For example, a deep sequence indicates that the user has thought deeply about the stimulus, while an extremely long sequence indicates that the cognitive process is redundant and may have excessive loops or stagnation, requiring timely termination.

[0132] In this embodiment, after analyzing and calculating the composition structure and length information of the cognitive state recognition sequence, the initial cognitive state recognition information is integrated with multiple historical cognitive state recognition information, and then reordered in chronological order to generate updated historical cognitive state recognition information. By updating the historical cognitive state recognition information and returning to the steps of acquiring user voice information and user behavior information, multiple user voice information and multiple user behavior information in the design stimulus event can be continuously collected, thereby realizing continuous recognition and processing of user cognitive state.

[0133] Step S603: Generate multiple updated historical cognitive state identification information based on the initial cognitive state identification information and multiple historical cognitive state identification information.

[0134] In this embodiment, when the length of the cognitive state recognition sequence is less than the preset threshold for the length of the cognitive state recognition sequence, the initial cognitive state recognition information is integrated with multiple historical cognitive state recognition information, and then reordered in chronological order to generate multiple updated historical cognitive state recognition information. This provides a complete historical data foundation for subsequent analysis and calculation of user voice information and user behavior information, thereby ensuring the accuracy of continuous recognition of the user's cognitive state.

[0135] Step S604: The multiple updated historical cognitive state identification information is used as multiple historical cognitive state identification information, and the process returns to step S101.

[0136] In this embodiment, by updating the historical cognitive state recognition information and returning to the steps of acquiring user voice information and user behavior information, the continuous collection of multiple user voice information and multiple user behavior information in the design stimulus event is achieved, thereby improving the accuracy and continuity of recognizing the user's cognitive state.

[0137] The cognitive state recognition method provided in this application improves the accuracy of target cognitive state recognition information, can provide reliable data support for the accurate matching of dynamic stimulus guidance strategies, help optimize the way creators process and handle design stimuli, and ensure and improve the creativity and performance of design results.

[0138] Figure 7 shows a flowchart of the cognitive state recognition method provided in Embodiment 7 of this application. The difference between this method and Embodiment 1 is that, after step S104, the method further includes:

[0139] Step S701: Obtain stimulus information from multiple user designs.

[0140] In this embodiment, the multiple user design stimulus information can be the design stimulus content that the user encounters in the current design stimulus event. It can include structured field information such as the number, title, definition, application example, image information, keywords and key content corresponding to the service design heuristic card. It can also include design stimulus content information in different forms such as text stimuli, image stimuli, audio stimuli, and video stimuli presented in the design stimulus event. The design stimulus content information can be directly retrieved from a human-built design stimulus database.

[0141] Step S702: Generate multiple user cognitive guidance information based on the multiple target cognitive state identification information, multiple user-designed stimulus information, and multiple preset user cognitive guidance strategy information.

[0142] In this embodiment, multiple preset user cognitive guidance strategy information can be manually set and have a one-to-one matching relationship with multiple target cognitive state identification information. Each target cognitive state identification information can be a non-stimulus processing state, a paused state, or a normal, observational, or abnormal situation under four cognitive state changes: forward, jump, loop, and backward. It can also be the length information of the cognitive state identification sequence and the composition structure of the cognitive state sequence. All of these can be matched with exclusive preset user cognitive guidance strategy information. Multiple target cognitive state identification information can be matched with multiple preset user cognitive guidance strategy information, and different preset user cognitive guidance strategy information can correspond to different guidance methods, that is, they can be combined with user-designed stimulus information to generate different user cognitive guidance information.

[0143] In this embodiment, it can first be determined whether the target cognitive state recognition information belongs to a state change under the categories of non-stimulus, pause, forward, jump, loop, or regression. Then, it can be determined whether the state change is normal, observation, or abnormal. This information is then used to accurately match multiple preset user cognitive guidance strategy information. At the same time, the guidance method is adapted in detail by combining the specific content in multiple user-designed stimulus information, such as the interpretation and application examples of service design heuristic cards. This allows the most suitable guidance method to be selected from multiple preset user cognitive guidance strategy information to generate multiple user cognitive guidance information.

[0144] In this embodiment, multiple preset user cognitive guidance strategies may include: a mild non-stimulus intervention strategy, with corresponding guidance methods including "the system highlights the current service design heuristic title" and "the system issues a voice prompt suggesting that designers return to the stimulus discussion"; a high non-stimulus intervention strategy, with corresponding guidance methods including "automatically reloading the current stimulus card, highlighting the title, enlarging and centering it," "the system issues a voice warning that the design team has left the design discussion and emphasizes returning to the discussion promptly," and "the system interface displays the duration of the current non-stimulus treatment and provides a text reminder of when the team has left the stimulus discussion"; and a pause mild intervention strategy, corresponding to... The guidance methods include "highlighting keywords and key phrases in the current service design heuristic card's explanation" and "generating corresponding guiding questions based on the cognitive state before the pause, such as 'Do we need to break down the stimulus content into several elements that can be used for the task?'"; it can include a high-intervention strategy for pauses, with corresponding guidance methods including "showing 1-2 pieces of discussion dialogue content from other users regarding this heuristic card in the system's stored historical data" and "adding another image and application example of this service design heuristic for auxiliary stimulation"; it can also include a low-intervention strategy for moving forward, with corresponding guidance methods including "displaying the current cognitive state above the card, The system may use arrows to indicate the next cognitive state to enter and provide text prompts to accelerate the transition to the next state, such as prompting the user to move from transformation to generation, 'Try to generate innovative ideas from the transformed content'; it may include a forward-height intervention strategy, with corresponding guidance methods including "System text prompts that the progress is too fast and the stimulus has not been processed deeply" and "The system requires the user to maintain the current cognitive state, such as 'Continue to try the transformation for 1 minute'"; it may include a jump-light intervention strategy, with corresponding guidance methods including "System prompts that a cognitive jump has occurred, displaying the jump details, including the number of jumps, the span, the previous cognitive state, and the current cognitive state, and explaining which cognitive states were omitted" and "System text..." Warnings about excessive leaps and suggestions for gradual reflection; this could include leap height intervention strategies, with corresponding guidance methods including "the system displays the omitted cognitive states and asks the user to complete them, such as 'Please take 1 minute to complete the transformation of the stimulus'" and "the system provides guidance on the steps to be completed, such as guiding the transformation 'You can consider transforming the application scenario of the service design heuristic card'"; this could also include cycle-based mild intervention strategies, with corresponding guidance methods including "the system warns that the current cognitive state has entered an excessive cycle" and "the system displays a text guiding the current cognitive state forward, such as prompting the user in the cycle judgment state 'Now try to extract the actionable points based on the judgment result'";This can include: a highly cyclical intervention strategy, with guidance methods including "the system prompts the user that they have lingered in the current cognitive state for too long and need to immediately move to the next cognitive state, and explains the task of the next state" and "the system displays two guiding texts to help them break out of the cycle"; a lightly cyclical intervention strategy, with guidance methods including "the system provides text prompts to guide the advancement of the cognitive state after the regression, such as prompting 'When hesitating, you can judge the value of the stimulus from the aspects of necessity, applicability, and novelty'"; a highly cyclical intervention strategy, with guidance methods including "the system prompts the user that they have over-regressed and points out the current risks" and "the system suggests that the user stop regression, continue to advance the previous cognitive steps, and provides two text instructions to move forward, such as 'It is recommended to continue the transformation, switch to other users' perspectives, and try to come up with solution ideas'"; and a lightly cyclical intervention strategy, with guidance methods including "the system provides language prompts about supplementing missing cognitive states, such as 'It is recommended to supplement service design heuristics'." The system can include: "judgment / transformation / generation" and "adding a thinking prompt for the missing cognitive state below the current stimulus card, such as prompting 'Is this service design card really suitable for our solution?' when a judgment is missing"; it can include a high-level intervention strategy based on the composition structure, with corresponding guidance methods including "displaying five types of cognitive states and their corresponding definitions, highlighting the currently missing cognitive state" and "providing guidance content for the currently missing cognitive state, such as providing different dimensions to consider when making a reasonable judgment on service design heuristics"; it can include a deep sequence intervention strategy, with corresponding guidance methods including "hiding the case text of the heuristic card, retaining the title and explanation" and "the system providing language prompts suggesting focus, such as 'The current thinking is quite sufficient, you can try to focus on the core direction'"; it can include an ultra-long sequence intervention strategy, with corresponding guidance methods including "forcibly requiring the end of the discussion on the service design heuristic card and placing it in the corresponding stimulus effect evaluation area" and "the system providing language prompts to accelerate termination, such as 'Please quickly end the stimulus discussion and make a decision'".

[0145] In this embodiment, when the target cognitive state identification information is the normal situation under the non-stimulus treatment state identification information, no intervention is required; when the target cognitive state identification information is the observation situation under the non-stimulus treatment state identification information, a mild non-stimulus treatment intervention strategy can be adopted; when the target cognitive state identification information is the abnormal situation under the non-stimulus treatment state identification information, a high non-stimulus treatment intervention strategy can be adopted.

[0146] When the target cognitive state recognition information is the normal situation under the pause state recognition information, no intervention is required; when the target cognitive state recognition information is the observation situation under the pause state recognition information, a mild pause intervention strategy can be adopted; when the target cognitive state recognition information is the abnormal situation under the pause state recognition information, a high pause intervention strategy can be adopted.

[0147] When the target cognitive state recognition information is the normal situation under the forward cognitive state recognition information, no intervention is required; when the target cognitive state recognition information is the observation situation under the forward cognitive state recognition information, a mild forward intervention strategy can be adopted; when the target cognitive state recognition information is the abnormal situation under the forward cognitive state recognition information, a high forward intervention strategy can be adopted.

[0148] When the target cognitive state recognition information is the normal situation under the jump cognitive state recognition information, no intervention is required; when the target cognitive state recognition information is the observation situation under the jump cognitive state recognition information, a jump mild intervention strategy can be adopted; when the target cognitive state recognition information is the abnormal situation under the jump cognitive state recognition information, a jump high intervention strategy can be adopted.

[0149] When the target cognitive state identification information is the normal situation under the cyclic cognitive state identification information, no intervention is required; when the target cognitive state identification information is the observation situation under the cyclic cognitive state identification information, a cyclic mild intervention strategy can be adopted; when the target cognitive state identification information is the abnormal situation under the cyclic cognitive state identification information, a cyclic high intervention strategy can be adopted.

[0150] When the target cognitive state identification information is the normal situation under the regression cognitive state identification information, no intervention is required; when the target cognitive state identification information is the observation situation under the regression cognitive state identification information, a mild regression intervention strategy can be adopted; when the target cognitive state identification information is the abnormal situation under the regression cognitive state identification information, a high regression intervention strategy can be adopted.

[0151] In this embodiment, multiple preset user cognitive guidance strategies can be matched for the cognitive state recognition sequence length information. For example, normal sequences do not require additional intervention and the current cognitive rhythm is maintained; deep sequences are subject to deep sequence intervention strategies; and very long sequences are subject to very long sequence intervention strategies, thereby generating corresponding user cognitive guidance information.

[0152] In this embodiment, the cognitive state sequence can be... The compositional structure is divided into three types: normal, observed, and abnormal. Then, when simultaneously satisfying... and If the condition is identified as normal, no intervention is required; if both conditions are met simultaneously... and If the situation is identified as an observational case, a mild intervention strategy based on the compositional structure can be adopted; if any of the following conditions are met... or If an anomaly is identified, a structurally-driven intervention strategy can be employed. Understandably, the ranges for coverage and Shannon entropy are derived experimentally.

[0153] In this embodiment, the content of the design stimulus in the current design stimulus event can be replaced, supplemented or adjusted based on the target cognitive state identification information to serve as user cognitive guidance information. However, such operations are all used as guidance strategies for the same design stimulus event and may not be presented as new design stimulus content.

[0154] The cognitive state recognition method provided in this application improves the auxiliary intervention effect on the creative process of designing stimuli, effectively realizes the application path of artificial intelligence and computer-aided creative technology, thereby efficiently optimizing the creative process of designing stimuli and significantly ensuring and improving the creativity and performance of the design results.

[0155] Corresponding to the methods in the embodiments described above, Figure 8 shows a structural block diagram of the cognitive state recognition device provided in this application embodiment. For ease of explanation, only the parts related to the embodiments of this application are shown. The cognitive state recognition device illustrated in Figure 8 can be the execution subject of the cognitive state recognition method provided in the aforementioned embodiment one.

[0156] Referring to Figure 8, the cognitive state recognition device includes:

[0157] The information acquisition module 810 is used to acquire multiple user voice information, multiple user behavior information, and multiple historical cognitive state recognition information;

[0158] The user voice text information and user behavior text information generation module 820 is used to generate multiple user voice text information and multiple user behavior text information based on the multiple user voice information, multiple user behavior information, a preset user voice text conversion model and a preset user behavior text conversion model.

[0159] The initial cognitive state recognition information generation module 830 is used to generate multiple initial cognitive state recognition information based on the multiple user voice text information, multiple user behavior text information and the preset initial cognitive state recognition model.

[0160] The target cognitive state identification information generation module 840 is used to generate target cognitive state identification information based on the initial cognitive state identification information, multiple historical cognitive state identification information, multiple preset user cognitive state division threshold information, and preset user cognitive state switching order information.

[0161] The process by which each module in the cognitive state recognition device provided in this application realizes its respective function can be specifically referred to in the description of Embodiment 1 shown in Figure 1 above, and will not be repeated here.

[0162] It should be understood that the sequence number of each step in the above embodiments does not imply 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 this application.

[0163] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0164] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0165] The cognitive state recognition method provided in this application can be applied to terminal devices such as mobile phones, tablets, wearable devices, in-vehicle devices, and augmented reality / virtual reality devices. This application does not impose any restrictions on the specific type of terminal device.

[0166] Figure 9 is a schematic diagram of the structure of a terminal device provided in an embodiment of this application. As shown in Figure 9, the terminal device 9 of this embodiment includes: at least one processor 90 (only one is shown in Figure 9), and a memory 91, wherein the memory 91 stores a computer program 92 that can run on the processor 90. When the processor 90 executes the computer program 92, it implements the steps in the above-described embodiments of the cognitive state recognition methods, such as steps S101 to S104 shown in Figure 1. Alternatively, when the processor 90 executes the computer program 92, it implements the functions of each module / unit in the above-described device embodiments, such as the functions of modules 810 to 840 shown in Figure 8.

[0167] The terminal device 9 can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. The terminal device may include, but is not limited to, a processor 90 and a memory 91. Those skilled in the art will understand that Figure 9 is merely an example of the terminal device 9 and does not constitute a limitation on the terminal device 9. It may include more or fewer components than shown, or combine certain components, or use different components. For example, the terminal device may also include input / transmission devices, network access devices, buses, etc.

[0168] The processor 90 may be a central processing unit, or it may be other general-purpose processors, digital signal processors, application-specific integrated circuits, off-the-shelf programmable gate arrays or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0169] In some embodiments, the memory 91 may be an internal storage unit of the terminal device 9, such as a hard disk or memory of the terminal device 9. The memory 91 may also be an external storage device of the terminal device 9, such as a plug-in hard disk, smart memory card, secure digital card, flash memory card, etc., equipped on the terminal device 9. Furthermore, the memory 91 may include both internal and external storage units of the terminal device 9. The memory 91 is used to store operating systems, applications, bootloaders, data, and other programs, such as the program code of computer programs. The memory 91 can also be used to temporarily store data that has been sent or will be sent.

[0170] This application also provides a terminal device, which includes at least one memory, at least one processor, and a computer program stored in the at least one memory and executable on the at least one processor. When the processor executes the computer program, it causes the terminal device to implement the steps in any of the above method embodiments.

[0171] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.

[0172] This application provides a computer program product that, when run on a terminal device, enables the terminal device to implement the steps described in the various method embodiments above.

[0173] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the content and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for recognizing cognitive states, characterized in that, include: Acquire multiple user voice information, multiple user behavior information, and multiple historical cognitive state recognition information; Based on the multiple user voice information, multiple user behavior information, a preset user voice-to-text conversion model, and a preset user behavior-to-text conversion model, multiple user voice-to-text information and multiple user behavior-to-text information are generated; based on the multiple user voice-to-text information, multiple user behavior-to-text information, and a preset initial cognitive state recognition model, initial cognitive state recognition information is generated; based on the initial cognitive state recognition information, multiple historical cognitive state recognition information, multiple preset user cognitive state segmentation thresholds, and preset user cognitive state switching order information, target cognitive state recognition information is generated; the process of generating target cognitive state recognition information based on the multiple user voice-to-text information, multiple user behavior-to-text information, and the preset initial cognitive state recognition model is described. The state recognition model generates initial cognitive state recognition information through the following steps: merging multiple user speech text information and multiple user behavior text information to generate multiple speech behavior text information to be recognized; calculating the similarity between the multiple speech behavior text information to be recognized and preset user design task information to obtain similarity information between the speech behavior text and the design task; determining whether the similarity information between the speech behavior text and the design task is less than a preset similarity threshold; if so, generating non-stimulus processing state recognition information as initial cognitive state recognition information; if not, calculating based on the start time and end time information of the multiple user speech information... Multiple user speech interval time information is obtained; when the multiple user speech interval time information is greater than a preset user speech interval time threshold, pause state recognition information is generated as initial cognitive state recognition information; when the multiple user speech interval time information is less than or equal to the preset user speech interval time threshold, initial cognitive state recognition information is generated based on the multiple speech behavior text information to be recognized and the preset initial cognitive state recognition model; the multiple preset user cognitive state classification thresholds include multiple preset target non-stimulus processing state classification thresholds, multiple preset target pause state classification thresholds, and multiple preset target cognitive state classification thresholds; the initial cognitive state classification is generated based on the initial cognitive state classification thresholds. The steps of generating target cognitive state identification information, based on state identification information, multiple historical cognitive state identification information, multiple preset user cognitive state division threshold information, and preset user cognitive state switching order information, specifically include: when the initial cognitive state identification information is non-stimulus processing state identification information, obtaining the number of times non-stimulus processing state identification is obtained; calculating the duration of non-stimulus user speech based on the start time and end time information of the user speech information corresponding to the initial cognitive state identification information; and generating target cognitive state identification information based on the number of times non-stimulus processing state identification is obtained, the duration of non-stimulus user speech, and multiple preset target non-stimulus processing state division threshold information.When the initial cognitive state identification information is a pause state identification information, target cognitive state identification information is generated based on the multiple user speech interval time information and multiple preset target pause state division threshold information. When the initial cognitive state identification information is not a non-stimulus processing state identification information and is not a pause state identification information, target cognitive state identification information is generated based on the initial cognitive state identification information, multiple historical cognitive state identification information, multiple preset target cognitive state division threshold information, and preset user cognitive state switching order information.

2. The cognitive state recognition method as described in claim 1, characterized in that, The multiple user voice information includes first user voice information and second user voice information; the first user voice information includes first user voice time sequence identifier information, first user voice start time information, and first user voice end time information; the second user voice information includes second user voice time sequence identifier information, second user voice start time information, and second user voice end time information; wherein, the first user voice time sequence identifier information is less than the second user voice time sequence identifier information; the process involves generating multiple user voice text information and multiple user behavior text information based on the multiple user voice information, multiple user behavior information, a preset user voice-to-text conversion model, and a preset user behavior-to-text conversion model. The steps specifically include: calculating user voice interval information based on the end time information of the first user voice and the start time information of the second user voice; determining whether the user voice interval information is greater than a preset user voice interval threshold; if so, performing differentiation processing based on the first user voice information and the second user voice information to obtain multiple user voice information to be converted; if not, performing splicing processing based on the first user voice information and the second user voice information to generate user voice information to be converted; and generating multiple user voice text information and multiple user behavior text information based on the multiple user voice information to be converted, multiple user behavior information, a preset user voice text conversion model, and a preset user behavior text conversion model.

3. The cognitive state recognition method as described in claim 2, characterized in that, The user behavior information includes user behavior occurrence time information and user behavior action information; wherein, the user behavior occurrence time information and user behavior action information correspond one-to-one; the step of generating multiple user voice text information and multiple user behavior text information based on the multiple user voice information to be converted, the multiple user behavior information, the preset user voice-to-text conversion model, and the preset user behavior text conversion model specifically includes: extracting the start time information and end time information of the multiple user voice information to be converted to obtain multiple user voice start time information and multiple user voice end time information to be converted; filtering the multiple user behavior occurrence time information based on the multiple user voice start time information and multiple user voice end time information to obtain multiple user behavior occurrence time information to be converted; determining multiple user behavior action information based on the multiple user behavior occurrence time information to be converted; and generating multiple user voice text information and multiple user behavior text information based on the multiple user voice information to be converted, the multiple user behavior action information, the preset user voice-to-text conversion model, and the preset user behavior text conversion model.

4. The cognitive state recognition method as described in claim 1, characterized in that, After the step of generating target cognitive state identification information based on the initial cognitive state identification information, multiple historical cognitive state identification information, multiple preset target cognitive state division threshold information, and preset user cognitive state switching order information, the method further includes: calculating the cognitive state identification sequence length information based on the initial cognitive state identification information and multiple historical cognitive state identification information; when the cognitive state identification sequence length information is less than the preset cognitive state identification sequence length threshold information, generating multiple updated historical cognitive state identification information based on the initial cognitive state identification information and multiple historical cognitive state identification information; using the multiple updated historical cognitive state identification information as multiple historical cognitive state identification information, and returning to the step of obtaining multiple user voice information, multiple user behavior information, and multiple historical cognitive state identification information.

5. The cognitive state recognition method as described in claim 1, characterized in that, After the step of generating target cognitive state identification information based on the initial cognitive state identification information, multiple historical cognitive state identification information, multiple preset user cognitive state division threshold information, and preset user cognitive state switching order information, the method further includes: acquiring multiple user-designed stimulus information; and generating multiple user cognitive guidance information based on the target cognitive state identification information, multiple user-designed stimulus information, and multiple preset user cognitive guidance strategy information.

6. A cognitive state recognition device, characterized in that, include: The information acquisition module is used to acquire multiple user voice information, multiple user behavior information, and multiple historical cognitive state recognition information; The user voice text information and user behavior text information generation module is used to generate multiple user voice text information and multiple user behavior text information based on the multiple user voice information, multiple user behavior information, a preset user voice text conversion model and a preset user behavior text conversion model. The initial cognitive state recognition information generation module is used to generate initial cognitive state recognition information based on the multiple user voice text information, multiple user behavior text information and the preset initial cognitive state recognition model. The target cognitive state recognition information generation module is used to generate target cognitive state recognition information based on the initial cognitive state recognition information, multiple historical cognitive state recognition information, multiple preset user cognitive state division threshold information, and preset user cognitive state switching order information. The step of generating initial cognitive state recognition information based on the multiple user voice text information, multiple user behavior text information, and a preset initial cognitive state recognition model specifically includes: merging the multiple user voice text information and multiple user behavior text information to generate multiple voice behavior text information to be recognized; and then, based on the multiple voice behavior text information to be recognized and a preset user design task... The similarity information of the task information is calculated to obtain the similarity information between the voice behavior text and the design task; it is then determined whether the similarity information between the voice behavior text and the design task is less than a preset similarity threshold; if so, non-stimulus processing state recognition information is generated as initial cognitive state recognition information; if not, multiple user voice interval time information is calculated based on the start time and end time information of the multiple user voice information; when the multiple user voice interval time information is greater than a preset user voice interval time threshold, pause state recognition information is generated as initial cognitive state recognition information; when the multiple user voice interval time information is less than or equal to a preset threshold, pause state recognition information is generated as initial cognitive state recognition information; when the multiple user voice interval time information is less than or equal to a preset threshold, pause state recognition information is generated as initial cognitive state recognition information. When the user speech interval time threshold information is obtained, initial cognitive state recognition information is generated based on the multiple text information of the speech behaviors to be recognized and the preset initial cognitive state recognition model; the multiple preset user cognitive state division threshold information includes multiple preset target non-stimulus processing state division threshold information, multiple preset target pause state division threshold information, and multiple preset target cognitive state division threshold information; the step of generating target cognitive state recognition information based on the initial cognitive state recognition information, multiple historical cognitive state recognition information, multiple preset user cognitive state division threshold information, and preset user cognitive state switching order information specifically includes: when the initial cognitive state When the cognitive state recognition information is non-stimulus processing state recognition information, the number of times the non-stimulus processing state is recognized is obtained; the duration of the non-stimulus user speech is calculated based on the start and end time information of the user speech information corresponding to the initial cognitive state recognition information; the target cognitive state recognition information is generated based on the number of times the non-stimulus processing state is recognized, the duration of the non-stimulus user speech, and multiple preset target non-stimulus processing state division thresholds; when the initial cognitive state recognition information is pause state recognition information, the target cognitive state recognition information is generated based on the multiple user speech interval time information and multiple preset target pause state division thresholds.When the initial cognitive state identification information is not non-stimulus processing state identification information, and the initial cognitive state identification information is not pause state identification information, then target cognitive state identification information is generated based on the initial cognitive state identification information, multiple historical cognitive state identification information, multiple preset target cognitive state division threshold information, and preset user cognitive state switching order information.

7. A terminal device, characterized in that, The terminal device includes a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 5.

Images